Annual Report 1994 Workmg Document No 148 U, ''JI J !J! 1 1('1 A 1 l}~ U r L , ll' •e Ol'l TROPICAL LOWLANDS PROGRAM Apnl1995 CONTENTS Page EXECUTIVE SUMMARY VIl CHAPTER Pro¡ect TA 02 Dynam1cs of land use (Contnbut10n to Dynam1cs of land Use TL 01 TA 02) Joyotee Sm1th Manuel Wmograd Douglas Pach1co Gi/berto Gallopm w1th contnbut1ons from Peter Jones Raul Vera Sam Fupsaka V1cente Cadawd Antomo Brandáo and lgnez Lopes 1 TA 02 Charactenzat1on of Slash and Burn agnculture and deforestat1on 1n Pedro Pe1xoto Acre and Theobroma Rondélma Braz1l Sam Fupsaka 21 Complementary subpro¡ects UT 51 (UT 04) A d1agnost1c study of agncultural land use m the Southwest Braz1han amazon and T A 51 (T A 02) Alternat1ves to Slash and Burn Sam F1psaka 25 JI Pro¡ects TL 01 and TL 03 Prototype sustamable croppmg systems for the Llanos and dynam1cs of land use Edgar Amézqwta W¡//iam Be// Pe ter Jones José 1 Sanz Joyotee Sm1th Rau/ Vera Phanor Hoyos and D1ego L 111 Molma 49 Subproject TL 01 Weed populat1ons Ecology w1th1n prototype sustamable croppmg systems for the Colomb1an Llanos Albert F1scher José 1 Sanz and Denms Fnesen 59 Subpro¡ect TL 01 2 Adopt10n of ley farmmg systems m the Colomb1an Llanos José V1cente Cadawd and Joyotee Sm1th 65 Subpro¡ect TL 03 1 Vegetat1on and ecology of the serranla nat1ve pasturas of the eastern plams of Colombia (Llanos) Georges R1ppstem J K Broekhuljsen A G E Peters and G Escobar 76 SubprojeCt TL 03 2 Nat1ve pastura management a sequent1al burmng and rotat1onal grazmg of the nat1ve pasturas of the plam (Aitlllanura) of the Llanos Georges R1ppstem A G E Peters J K Broekhu1Jsen and G Escobar Pro¡ect TC 01 Prototype sustamable croppmg systems for the Braz1han Cerrados M1guel A Ayaaa ProJeCt TC 03 Dynam1cs of land use M1guel A Ayaaa 104 127 133 Page Subproject TC 51 Sotl tndtcators of sustatnable agropastoral systems M1guel A A yarza 1 36 IV Project TA 01 Prototype sustatnable cropptng systems for forest margtns M1chael Thung 137 V ProJects TL 02 and TC 02 Mechamsttc understandtng and models of sotl chemtcal phystcal and btologtcal processes tn agropastoral and sequenttal crop productton system long Term Expenment 1 Sustatnable crop rotatton and ley farmtng systems for the actd sotl savannas of Colombta (Culttcore Expenment) D Fnesen H Carmen H Delgado E Owen R Thomas M F1sher A Gl}sman 1 Rao 8 Volverás H F Alarcón G R1ppstem P Lave/le A Moreno and J J J1ménez 1 4 7 Long Term Expenment 2 Agropastoral systems for the cerrados of Braztl M Ayarza L Vlfela J C Mlfanda A Cardoso H Neufeldt R Westerhof W Zech and J Duxbury 159 Strategtc Research on Key Processes Recycltng of nutnents vta decomposttton of lttter and crop restdue R1chard Thomas N Asawaka and H F Alarcón 174 Phosphorus transformatton m tmproved pasturas A Oberson D K Fnesen H T1essen J Mo1r and G Barrero 1 82 Nutnent cycltng through the mtcrobtal btomass A J Gl}sman A Oberson D K Fnesen J 1 Sanz and R J Thomas 187 Evaluatton of some phystcal properttes of an Oxtsol after converston of nattve savanna tnto legume based or pure grass pastures A J Gl}sman and R J Thomas 191 A cauttonary note about the use of the CENTURY sotl orgamc matter model for ecosystems on troptcal low P satis A J Gl}sman A Oberson H T1essen and D K Fnesen 197 Dynamtcs and short term effects of earthworms tn natural and managed savannas of the Eastern Platns of Colombta M F1sher J J J1ménez T Decaens A G Moreno J P Ross1 P Lave/le and R Thomas 208 Crop nutnttonal requtrements and tnput use efftctency on llanos oxtsols under crop rotattons D K Fnesen J 1 Sanz D Molma and M R1vera 223 IV VI Page TI 02 lnterprogram ProJect w1th Forages Program N1trogen f1xat1on and mtrogen transfer m forage legumes Richard J Thomas 234 lnterprogram ProJ&Ct TI 01 Prototype systems for ecolog1cally sound mtens1f1cat1on of product1on m the H1lls1des E Amézqwta J Ashby R 8est O K Fnesen S Fupsaka A G1¡sman e /gles1as 8 Knapp J Kornegay e Lascano 8 Maass K Muller J 1 Sanz J Sm1th R Thomas e A Qwroz andA Meléndez lnterprogram Pro]ect m H1lls1des TI 01 subproJect Market opportumt1es hnked to eros1on control pract1ces a key to adopt1on? Jo y otee 237 Sm1th 244 PUBLICA TI ONS 1994 251 MEMBERS OF THE TROPICAL LOWLANDS PROGRAM 255 COLLABORATING INSTITUTIONS 257 V EXECUTIVE SUMMARY Strateg1es The research act1v1t1es of the TLP fall broadly tn two ma¡or categones (a) w1th1n agroecosystem area based research act1v1t1es lead1ng to the 1dent1f1cat1on development and test1ng of prototype susta1nable land use systems for representatiVa s1tes 1n the savannas and forest marg1ns based on an understandtng of the underlytng b1ophys1cal and soc1oeconom1c processes and mechan1sms and {b) across agroecosystem research led by or 1n cooperat1on w1th the Land Use SRG a1med at prov1dtng pohcy opt1ons and deftntng the prof1le of potent1al technolog1cal tnnovat1ons and based on an understandtng of past and present trends 1n land use and the analys1s of past and present pohc1es Area basad research tn the forest marg1ns begun tn the f1rst semester of 1994 wh1le that of the savannas was formally 1ntt1ated tn 1992 Exploratory research across agroecosystems traces back to the stud1es leadtng to the development of CIAT s Strateg1c Plan The above set of research act1V1t1es are orgamzed around three ma¡or types of pro¡ects ( 1 ) dynam1cs of land use {2) mechamst1c understand1ng of so1l chem1cal phys1cal and b1olog1cal processes tn agropastoral and sequent1al crop systems (3) development of prototype cropp1ng systems plus a smaller set of spec1al fundad more spec1f1c pro¡ects To vanous degrees all of these attempt to quant1fy the tradeoffs tnvolved tn tntervent1ons tn the two tropical lowlands agroecosystems researched namely the neotrop1cal savannas and the tntervened ratnforest margtns Funct1onal aspects of the two agroecosystems are ( 1) the prov1s1on of resources for agncultural product1on both pnmary and secondary (e g so1l water plants etc ) (2) the prov1s1on of vanous env¡ronmental serVIces tncludtng amemt1es and hfe support mechamsms (e g ecosystem stab1hzat1on chmate regulat1on ma1ntenance of genet1c d1vers1ty) and (3) ass1mlfatlon of waste products (e g C02 ) W1th1n th1s very large array of researchable 1ssues CIAT places particular but not exclusive emphas1s tn the use and ma1ntenance of the so1l resources 1n v1ew of the prevalence tn both agroecosystems of margtnal and low fert1hty so1ls and to the 1mpact of farmers dec1s1ons regard1ng land use on natural resources In concurrence w1th CIAT s strategy tn resource management the Program pos1t1ons 1tself as one part1c1pant tn the larga array of R&D mst1tut1ons 1nvolved tn RMR tn these agroecosystems Based on 1ts 1nternat1onal nature and neutrahty the Program attempts to play a faci11tattng role tn promottng the tnvolvement of other research mst1tutes tn both agroecosystems The h1ghhghts outhned 1n th1s report tntend to document how the Program 1s attempttng to carry out th1s dual role 1nvolv1ng 1ts own research most frequently developed tn cooperat1on w1th others and research by other lnstttuttons wh1ch have been prompted by our own act1v1t1es In th1s context spec1al ment1on needs to be made of sorne mst1tut1ons whose collaborat1on 1s essent1al for the Program s act1v1t1es In our host country the support and collaborat1on rece1ved from CORPOICA Canmagua 1s cruCial to the success of the Llanos work and the same comment apphes to the contnbut1ons of EMBRAPA CPAC tn the Braz11tan Cerrados and CPAF Acre and CPAF Rondoma tn the Forest Margtns area Notw1thstand1ng numerous other collaborators two tnternat1onal1nst1tutes have s1gmf1cantly contnbuted to the Program s research act1v1t1es These are IFDC USA who has ass1gned a So1l sc1ent1st to work w1th the TLP tn the area of so1l fert1hty management and CIRAD France wh1ch has ass1gned a Savanna ecolog1st to conduct research on plant dynam1cs 1n the Colomb1an Llanos Ma]or Program donors were the Colomb1an Government and the lnter Amencan Development Bank Research act1vlt1es Understand10g land use dvnam1cs Understand1ng and pred1ctmg development paths of both agroecosystems and the1r mteract10ns 1mplles the analys1s of past trends m land use the analys1s of past and current pollc1es and modelllng of future alternat1ve scenanos cond1t1oned by potent1al pollc1es and technolog1es Such an analys1s began 1n 1 994 based on llterature rev1ew and secondary data The root causes of deforestat10n 1n the Amazon are the opportumt1es 1t prov1des for land speculat1on to the nch and for mcreased labor returns to the poor The dnvmg forces behmd these opportun1t1es ong1nate pnmanly m government pollc1es both w1th1n and outs1de the agncultural sector In the savanna product1ve technolog1es have enabled areas Wlth good mfrastructure to make 1mportant contnbut1ons to nat1onal agncultural product10n at the expense of on and off s1te degradat1on Speculat1ve land demand has expanded the agncultural front1er mto marg1nal areas A modelllng exerc1se to dep1ct future land use scenanos up to the year 2020 and the1r 1mpllcat10ns m terms of product1v1ty labor absorpt1on and ecolog1cal sustamabd1ty was 1n1t1ated Jomtly by the TLP the Land Use SRG the lmpact Assessment umt and an NGO Prellm1nary results show that the combmat1on of favorable pollc1es and technolog1es for ecosystem management could make a ma]or contnbut1on towards s1multaneously reducmg deforestat1on wh1le enabhng nat10nal governments to obtam econom1c and soc1al benef1ts from the forest margm S1gmf1cant mcreases m agncultural output and agnbusmess could be sustamably ach1eved m the savanna and th1s could enable 1t to d1vert pressure from the Amazon by contnbut1ng to government s econom1c ob¡ect1ves d1vertmg venture capital and absorbmg labor 1n agnbus1ness Aspects of the favorable scenanos have started to occur 1n recent years Among them are changes m pollc1es emergmg mternat1onal markets 1n ecolog1cal serv1ces wh1ch could prov1de powerful mcent1ves for the adopt1on of sustamable systems and new ecolog1cal parad1gms wh1ch prov1de opportun1t1es for combmmg env1ronmental enhancement w1th econom1c gam Th1s constellat1on of events prov1des a umque opportun1ty for effect1ve ecosystem management Assessment of poss1ble future land use scenanos reqUire cont1nued understandmg of the factors dnvmg farmer s adopt10n of new and emerg1ng technolog1es An adopt1on study of pasture based technolog1es m the Colomb1an savanna showed that 98% of sample farmers had planted 1mproved pastures 17% of the surveyed area cons1sted of planted pastures w1th the are a mcreasmg 1n the last 1 5 years at the rate of 14% p a Only 18% of the 1mproved are a contamed legumes and 1 5% was planted to the nce pasture technology the latter be1ng a new technology at the t1me of the survey The predommant strategy of farmers was to supplement nat1ve savanna w1th a small area of planted grass pasture wh1ch was mcreased over t1me as cap1tal became ava1lable Capital and management reqUirements were 1dent1f1ed as key factors mh1b1t1ng area expans1on of planted grass pastures and adopt1on of grass legume pastures and ley farmmg systems As natural mcrease 1n herd s1ze occurs farmers appear to be acceptmg 1mproved grass pastures as an alternat1ve to over grazmg The study shows that reduct1on of the cap1tal and management mtens1veness of ley farmmg systems even at the expense of product1v1ty could make a ma]or contnbut1on to the prevent10n of VIII over grazmg by speedmg up the rate of adopt1on In terms of the savanna as a whole the study area hes towards the unfavorable extreme m terms of so1l fert1hty In terms of mfrastructure 1t 1s mov1ng rap1dly from the typ1cal towards a more favorable d1rect10n The ley farm1ng technology m 1ts current form should be targeted to the most favorable areas of the savanna Another external factor that may cond1t1on future land use m the forest margms and at least parts of the savannas 1s the emergmg development of mternat1onal markets for ecolog1cal serv1ces Technolog1cal mtervent1ons such as deep rooted germplasm that can effect1vely sequester Cm depth and poss1bly affect the exchange of other gasses as well (see below) may have an 1mportant role 1n the future and const1tutes an area wh1ch the Program w1ll modestly explore to assess 1mphcat1ons for further technolog1cal developments A great deal of effort and very s1gmf1cant resources were deployed m 1994 ter the contmued charactenzat1on and m depth analyses of land use dynam1cs m the two target agroecosystems under the leadersh1p of the Land Use SRG G1ven resource hm1tat1ons the Acre Rondoma area was pnont1zed followed by the Colomb1an Llanos Sorne etforts were made to obta1n add1t1onal and more deta1led secondary mformat1on to 1mprove the charactenzat1on of selected s1tes m the Cerrados of Braz11 wh1le charactenzat1on of the Venezuelan Savannas 1n cooperat1on w1th a number of nat10nalmst1tut1ons has been on hold System charactenzat10n research has been conducted m Acre and Rondoma m the Braz1han Amazon as part of the global 1n1t1at1ve Alternat1ves to Slash and Burn Agnculture (ASB) funded by GEF and coordmated globally by ICRAF and by CIA T m Latm Amenca and m partnersh1p w1th EMBRAPA (m Acre and m Rondoma) PESACRE ICRAF IFPRI CIFOR and TSBF The purpose of the ASB 1s to develop techmcal and pohcy alternatiVas wh1ch would help to decrease rates of tropical deforestat1on wh1le enhancmg the well bemg of forest resource users Trop1cal deforestat10n a contnbutor to global warmmg v1a release of atmosphenc C02 1s h1ghest m Latln Amenca comparad to Afnca and As1a Rates m the Amazon Basm of Braz1l mcreased from the early 1 960s through the m1d 1980s due to nat1onal pohc1es support1ng road bu1ldmg tax and cred1t mcent1ves to large corporat1ons and ranches and colon1zat1on proJects for the rural peor Changes m the same pohc1es seem to have contnbuted to observad dechnmg rates of deforestat10n for the Braz1han Amazon as a whole lnterv1ews conducted m the proJect area 1nd1cated that settler colomsts had parcels of a mean 88 ha 1n Pedro Pe1xoto (Acre) and 76 ha m Theobroma (Rondoma) These lands were approx1mately 60% m forest and 40% cleared for pastures and crops at the t1me of the mterv1ews Farmers cleared shghtly more than ene ha per famdy per year to produce f1rst nce for wh1ch y1elds were calculated to be approx1mately 1 5 t/h m the f1rst year but dropped drast1cally such that ma1ze and cassava were grown 1n second and 1n sorne cases a th1rd year of cult1vat1on Lands were then converted to pasture as farmers not only banked the1r savmgs m cattle (sorne 85% of the settlers had cattle w1th herd s1zes of about 25 30 head) but also and perhaps more 1mportantly sought to take advantage of substant1ally h1gher values for 1mproved lands 1 e cleared lands w1th pasture fencmg corrals and ponds Development of alternat1ves to slash and burn agnculture wh1ch would decrease rates of deforestat1on 1ncrease sustamabd1ty of resource use and enhance the well bemg of settlers would have to combme on farm and pohcy research IX A deta1led GIS coverage 1s bemg developed for the study area between Puerto Lopez and Puerto Ga1tán Meta m the Colomb1an Llanos Th1s w1ll help refme ex1stmg (Cochrane s et al ) class1f1cat1ons momtonng of land use changas and allow extrapolat1on of technolog1cal and pohcy opt10ns Maps of so1ls and s01ls charactenst1cs dramage topography vegetat10n etc have been d1g1t1zed Cnt1cal to the effort of the Program m assessmg externaht1es assoc1ated w1th present and potent~al land uses 1s the development of a d1g1tal elevat1on model wh1ch w1ll allow accurate est1mat10n of slopes and runoff for apply1ng watershed management and eros1on control models Th1s 1s an ongo1ng h1ghly t1me consummg act1v1ty Satelhte 1mages for prev1ous years were obtamed and are bemg sub1ected to unsuperv1sed class1f1cat10n we expect to obtam a new satelhte 1mage for 1995 wh1ch wlll be ground truthed to allow superv1sed class1f1cat10n of current land use To complement the above efforts farms located 1n contrastmg land systems (as class1f1ed ongmally by Cochrane et al ) and that d1ffer m the use of savanna resources have been momtored for over 18 months lnputs and outputs for spec1f1c f1elds and/or whole farms are quant1f1ed to the extent poss1ble and s01l vegetat1on and other resources are bemg charactenzed All of the data has been georeferenced but there are hngenng doubts regardmg the prec1s1on of the GPS mstruments used so far The recent acqu1s1t10n of prec1se GPS mstrumentat1on by the Land Use SRG w1ll allow venf1cat10n of the coordmates m 1995 followmg wh1ch the data can be 1ncorporated mto a GIS database and overlayed onto the maps and 1mages referred to above for use m land use s1mulat1on and other purposes S1mllarly extens1ve stud1es were conducted dunng 1 gg4 on the flonst1c compos1t1on of nat1ve savanna wh1ch const1tutes the mam ( > 70%) land use form of the reg10n Spec1es d1vers1ty was quant1f1ed dunng 1994 for a large cross sect1on of the undulatmg savannas or serramas a format10n that const1tute nearly 60% of the Colomb1an and poss1bly also a large percentage of the Venezuelan savannas As befare these data have been georeferenced and extens1ve s01l and plant t1ssue analyses were conducted lnterv1ews w1th ranch managers and md1genous commun1t1es were also undertaken to assess management pract1ces apphed to nat1ve savanna commumt1es that may affect plant dynam1cs On the whole 173 plant spec1es were 1dent1f1ed belongmg to 40 d1fferent fam1hes Only 89 of these spec1es are common to the levelled savanna stud1ed m the past m Can magua and surroundmg areas Large d1fferences 1n plant spec1es compos1t1on assoc1ated w1th topography so1ls and seasonal hydrology ex1st but add1t1onal and longer term research IS st1ll requ1red to understand and quant1fy the contnbut1on of these plant commun1t1es to the savanna ecosystem and the poss1ble tradeoffs mvolved 1f land use 1ntens1f1cat1on s1gmf1cantly affects some or all of these 1mportant commumt1es Th1s 1s an area that requ1res the contnbut1on of other d1sc1phnes part1cularly tropical ecology and for wh1ch the Program 1s attempt1ng to mvolve other research mst1tut10ns As md1cated prev1ously longer term stud1es of the dynam1cs of savanna vegetat1on m response to current and potent1al uses are reqUired A large ongo1ng expenment set up m Canmagua was des1gned to prov1de mformat1on on the effect of season and frequency of burmng and of grazmg management on vegetat10n dynam1cs above ground and underground b1omass product1on and seasonal changas 1n graz1ng preference Large d1fferences are apparent m terms of spec1es dynam1cs but 1t 1s st1ll too early to draw conclus10ns Complementary observat1ons are made on some of the farms bemg momtored as explamed above Plant dynam1cs and changes m so1l parameters are evaluated and w1ll be related to management pract1ces employed by farmers X Process oneoted b!ophys1ca! research lotenslf!cat•on of land use 10 the savannas may bnng about changes 10 many of 1ts !and resources mc!udmg above and under ground fauna and 10 hmnolog!ca! propert1es of savanna water streams The !atter two are oot areas 10 wh1ch the Program has expert1se An exp!oratory survey of above grouod fauna 1s bemg cooducted at Canmagua by a professor of the Nat1ona! Umvers1ty of Colombia Department of 81ology Bogotá made poss1ble by the fmaoc1al support prov1ded by CORPOICA Canmagua aod the TLP An ID!t!al survey of so1l fauna cooducted earher on by a French student 10 the Colomb1an Llanos led to the des1gn of a much more amb!t!ous oogo1ng project cooducted by a Spamsh student under the superv!S!OD of an ORSTOM (France) spec1ahst anda professor of the Umvers1dad Complutense de Madnd The study takes advantage of a large longterm expenmeot setup 10 Canmagua mvolvmg a w1de raoge of h1ghly contrast1ng prototype cropp1ng and pasture systems that allows quantlf!cat!oo of changes 10 so1l fauna 10 response to temporal changes 10 land use G1ven the low natural fert1hty of savanna ox1so!s the expenmeot •nvest1gates a combmat1on of crop and/or pastura componeots at two !evels of mtens•f•cat1oo based on hme to determme b!ophys1cal measures of system performance aod health Th1s as well as an eqU!va!eot expenmeot set up 10 the Cerrados of Braz1! are jO!ot!y ruo w1th the respect1ve nat10na! IOStltUt!ons CORPOICA and EMBRAPA CPAC respect1ve!y plus the Umvers1t1es of Bayreuth (Germany) and Cornell (USA) 10 the Braz1han case A !arge team of researchers !S 1nvo!ved 10 both cases The Canmagua cultlcore expenmeot was set up part!y 10 1993 aod the rest 10 1994 The expenment essent1ally contrasts three major prototypes that mod1fy so1! and vegetat1on to vanous degrees w1th grazed nat1ve savanna represent1ng the !owest degree of human !oterveot!oo and a ma1ze based set of treatmeots represeotmg the h1ghest degree of resource mod!f!cat1on Th1s !ast year w1tnessed the !rnp!emeotat!OD of the h1gh hme ma1ze based systems usmg cv S1kua01 commerc1ally re!eased by CORPOICA aod CIMMYT for the Colomb1an Llanos 10 1994 Legume green manures 1ncluded as part of the system rap1dly showed at least one of the tradeoffs 1nvolved lo effect gram y1e!ds s•gnlf!cantly mcreased followmg the green maoure but so d1d the concentrat1on of N03 N down the s01! prof1!e suggestmg that N leakage 10 these so1!s 1s a real poss1b1hty Ma1ze gra1n y1e!ds 10 the expenmeot were 2 5 t ha 1 or approx1mate!y 1 ton less than 10 adjacent satelhte expenments probab!y due to a combmat1on of poor establishment and problems of machmery at sow1ng Th1s 1s 111ustrat1ve of sorne of the problems 10 settmg up long term expenments w1th cootrastmg componeots when the1r agronomy 1s st!ll oot well developed Nevertheless the expenment 1s extreme!y successful overall and prov1des a common ground for a !arge lotennstltut!ooal and mult!d!sc!phnary team that 1s mvest1gatmg outnent cychog processes mput/output re!at•onsh1ps so1! and crop management 1ssues and so1! fauna dyoam1cs Results of th1s expenment are bemg used to cahbrate ava1!able crop growth models that w1ll allow the s!mulat!on of oumerous other crop and pastura sequences aod hopefully extrapolat1on to other so1!s as well The Braz!l¡ao long term expenment or crop pasture mtegrat1on expenment completad 1ts th1rd cropp1ng year w1th the best treatmeots y1e!dmg 10 excess of 7 t ha 1 of ma1ze Major y1eld effects so far are due to s01! fert1hty maoagement wh1ch 10 turo affects the weed populat1on Although so!l mechamca! 1mpedance has began to detenorate under contmuous croppmg and to a !esser extent under a grass !egume pastura as comparad to the nat1ve Cerrado no detnmeota! effects are appareot as yet Oo the other hand mycorrh1zal popu!at1ons momtored over the last three years have shown temporal vanat1ons 10 re!at1on to treatmeot IO!t!ally there was a 25 fo!d 1ncrease dunog the pasture establishment phase aod a 3 fold 1ocrease 10 soybeans re!at1ve to Cerrado but the d1fferences have decreased XI s1gmf1cantly over t1me Deta1fed stud1es on soli organ1c matter contents and d1stnbut10n and water and nutnent dynam1cs are bemg carned out The same expenment 1s bemg used by an EMBRAPA Cornell Umvers1ty team to momtor changas 1n carbon d1ox1de mtnc ox1des and methane Prehmmary results suggest that these so1fs may constltute a large methane smk w1th methane ox1dat1on bemg largest under corn and smallest under nat1ve Cerrado Nevertheless pare s1ze and so1f water balance as affected by t11fage appear to mod1fy sorne of th1s ab1hty to d1spose of methane Current trends 1n trop1cal so1f fert1hty management stress rehance on b1olog1cal processes 1nclud1ng the use of so1f adapted germplasm enhancement of s01l b1olog1cal act1v1ty and opt1m1zat1on of nutnent cychng to max1m1ze eff1c1ency of use of externa! mputs The above hsted longterm expenments together w1th a large array of short term f1eld and laboratory tnals are bemg used to quant1fy and model nutnent cychng processes Crops res1dues and crop and forages htter appear to play a key role 1n savanna ox1sols and the hgn1n N rat1o 1s the mam regulator of decompos1t1on w1th C N rat1os and the polyphenols contents playmg smaller roles The CENTURY model appears to be reasonably well SUited to model sorne of these N cyclmg processes 1f mod1f1ed to accommodate present expenmental results The transfer of N 1n forage legume htter to the assoc1ated grass was very slow whereas unne N was shown to cycle much faster Th1s combmat1on of the effects of orgamc add1t1ons v1a res1dues htter and/or ammal manure and unne were shown to 1mprove nutnent cond1t10ns for s01f flora and fauna and as consequence appear to benef1c1ally mod1fy orgamc P avalfab1hty measured us1ng more elaborate part1t1on1ng techmques than the trad1t1onal laboratory methods An excellent 1nd1cator of these changes 1n the blologlcal propert1es of savanna ox1sols appears to be so1f m1crob1al b1omass to the extent that nutnent cychng 1s very t1ghtly hnked to 1ts turnover rate In fact P flux through m1crob1al b1omass at least under grass legume pastures ( 12 34 kg ha 1 year 1 ) 1nd1cates that th1s could be a ma¡or pathway of P cychng 1n these s01fs These fmdmgs are supported by the detect1on of h1gh phosphatase act1v1ty under pastures wh1ch further 1nd1cates the 1mportance of blolog1cal processes 1n so1f P turnover lntroduced grass and grass legume pasturas were found to make a ma¡or contnbut1on to so1f orgamc matter to a depth of at least 100 cm 1n both on stat1on and on farm expenments As a m1n1mum 1t was est1mated that a three year old 8 d1ctyoneura pasture contnbutes 30 t C ha 1 1n 3 years and that the add1t1on of a legume s1gmf1cantly mcreases the amount of C sequestered These fmdmgs were pubhshed 1n Nature and do not requ1re further elaborat1on he re On farm testmg of prototypes Longterm controlled expenments are absolutely essent1al to develop a deta1fed understandmg of s01l plant ammal processes as shown above Nevertheless they have a certam ng1d1ty regardmg the crop and/or pasture components that can be tested To compensate for these hm1tat1ons and more 1mportantly to mcorporate 1nto the Program s research agenda farmer s perspect1ves and preferences regardmg resources use a hm1ted amount of on farm work 1s conducted 1n all of the area based act1v1t1es carned out by the TLP Th1s set of tnals have the longest trad1t1on 1n the savannas of both Colombia and Braz1f whereas they only begun to be Implementad 1n Acre Rondoma 1n m1d to late 1994 To vanous degrees and dependmg on the spec1f1c locat1on farmer part1c1pat1on 1n the des1gn and conduct of the tnals 1s mcluded Thus farmer s controls are always mcluded and farmers are mvolved 1n the d1scuss1on leadmg to select1on of some of the expenmental treatments select1on of crop or forage components etc whereas st1fl other treatments are purely researcher selected To the extent that farmer XII dec1s1on makmg processes are documentad 1n these on farm act1V1t1es th1s research 1s cons1dered to be as strateg1c as that of purely b1ophys1cal ongm Only a bnef ment1on 1s made here of one of these on farm tests whereas the rest are deferred to the mam body of the report The Matazu/ on farm tnals 1n the Llanos of Colombia were descnbed 1n the Savannas Annual Report 1992 1993 The oldest two tnals 1n th1s farm completad 6 years 1n 1994 One of them mvolved contmuous nce monocropp1ng usmg the best ava1lable knowledge regardmg management of soll phys1cal and chem1cal propert1es Over the penod 1989 1992 there was a linear decrease of 400 k9 ha 1 year 1 lmproved sml management pract1ces apphed subsequently were able to part1ally reverse th1s trend but contmued negat1ve changes 1n so1l compact1on weed bwldup and decreased m1crob1al b1omass contmue to consp1re agamst the longterm success of th1s alternat1ve Stud1es conducted on weed dynam1cs the1r nutnent uptake and overall compet1t1on w1th the sown crop are beg1nmng to 1dent1fy des1rable plant 1deotypes that could potent1ally be undersown to nce and compete Wlth undes1rable weeds In general then the monocrop prototype 1s prov1d1ng very valuable mformat1on on the extent rate and charactenst1cs of degradat1on processes 1n savanna ox1sols An equally old expenment whose detalls have been extens1vely reportad elsewhere the nce pasturas expenment entered 1nto the second phase of pasturas 1n the nce pasture rotat1on 1n 1994 followmg a renovat1on of the ongmal pasturas w1th nce 1n 1993 Th1s expenment 1s begmmng to 1llustrate sorne of the longterm tradeoffs mvolved w1th these relat1vely mtens1ve systems Pasturas estabhshed under nce support h1gh carrymg capac1ty and h1gher levels of ammal product1on than otherw1se By the sama token sorne aspects of the so1l phys1cal propert1es such as compact1on est1mated usmg trad1t10nal methods (cone penetrometer) begm to detenorate relat1vely early on dunng the pasture phase A crop phase namely nce temporanly reversas th1s trend Trends 1n so11 orgamc matter and m1crob1al b1omass contrast markedly w1th those of the monocrop referred to above In the Acre Rondoma area (Forest Margms) agronom1c work bagan 1n the second tnmester of 1994 1n1t1ally severely constramed by log1st1cal and transportat1on dlff1cult1es later overcome Ex1st1ng agronom1c knowledge on potent1al agros1lvopastoral components 1s scarce and make 1t unadv1sable to set up longterm complex expenments as yet Follow1ng the research agenda agreed upon w1th the other part1c1pants 1n the Alternat1ves to Slash and Burn project the Program bagan 1ts research act1v1t1es w1th the 1ntroduct1on of germplasm of the mam staple foods 1n the reg1on namely beans ma1ze nce and cassava to be followed later on by grass and legume forages mcludmg mult1purpose shrubby and trae legumes Fert1hzer 1nputs although mcluded as expenmental treatments are unfeas1ble under the current econom1c cond1t1ons so that b1olog1cal management of sml resources 1s even more cruc1al than 1n the other mandated reg1ons of the Program lmt1al st1ll h1ghly tentat1ve results suggest that 1t 1s poss1ble to use dead mulches that protect the soll wh1le at the same t1me reducmg the 1nc1dence of web bhght 1n beans based croppmg systems A number of mulches was testad both on stat1on and on farm mcludmg velvet bean Puerana phaseol01des ma1ze and nce Prehmmary results 1nd1cate that mulches such as those prov1ded by P phaseolotdes that decompose at a slower rata prov1de better protect1on agamst web bhght and presumably protect the so11 for a longar t1me Nevertheless 1t was obv1ous that the maJonty of the mtroduced hnes and vanet1es d1d not have an y1eld advantage ovar farmer s checks (mostly Caneca) XIII lnter program pro¡ects Bnef reference was made above to vanous research act1v1t1es mvolvmg collaborat1on of Program se~ent1sts w1th other researchers both ms1de and outs1de CIA T Severa! other ¡01nt pro¡ects w1ll be ment1oned here Severa! Program sc1ent1sts were mstrumental 1n des1gnmg and 1mplementmg a large mterprogram pro¡ect m the Cauca (Colombia) h1lls1des together w1th members of most other CIAT s programs A program spee~ahst 1s also respons1ble for the mamtenance and charactenzat1on of the forage legumes Rh1zob1um collect10n and for screemng and selectmg Rh1zob1um strams m cooperat1on w1th the Trop1cal Forages Program Severa! Program sc1ent1sts were heav1ly 1nvolved m des1gmng and 1mplementmg roa1ze based croppmg systems w1th CIMMYT and CORPOICA m the Colomb1an Llanos S1m1larly Program researchers act1vely contnbuted to evaluat1on of ma1ze nce and forages germplasm for alternat1ve product1on systems 1n the Cerrados of Braz1l and ma1ze nce cassava beans and forages germplasm m the Acre Rondon1a s1tes lnst1tUt1onal strengthemng and relat1ons In the area of mst1tut1onal bUIIdmg severa! 1mportant events took place dunng 1994 A C sequestrat10n workshop took place w1th the part1c1pat1on of severa! NARs and 1nternat10nalmst1tUtiOns to develop an mtermst1tut1onal pro¡ect on the sub¡ect for subm1sS10n to potent1al donors S1m1larly the Program was heav1ly mvolved 1n a land Use led workshop leadmg to development of a reg1onal pro¡ect for the savannas ent1tled Strateg1es for Sustamable Agncu/tural Land Use m the Low/and Savannas of South Amenca SSALLSSA The Program part1c1pated m the Global Steenng Comm1ttee meetmg of the ASB Pro¡ect and later on m a trammg workshop on charactenzat10n offered m Na1rob1 Kenya m the context of that same pro¡ect The th1rd Workshop on Agropastoral Systems was held 1n Venezuela wh1ch was supported by IOB and Venezuelan funds Jumor staff of the TLP were heav1ly mvolved m the development of an lnternat1onal Course on Agropastoral Systems for Ac1d S01ls offered 1n V1llav1cenc1o Colombia for extens10n agents and pnvate techmcal ass1stants lastly numerous students developed the1r BS and MS thes1s workmg m ex1st1ng research pro¡ects In a d1fferent context vanous Program sc1ent1sts were regularly 1nvolved m techmcal consultat1on meetmgs w1th the MAS consort1um PROCITROPICOS and 1ts savannas consort1um PROCIANDINOS FAO CORPOICA and vanous others S1m1larly through sorne of 1ts members the Program has been mvolved m two system w1de lnlt1at1ves namely CIAT s Ecoreg1onal pro¡ect and the S01l Water and Nutnent Management 1mt1at1ve Lastly 1t also prov1ded hm1ted mput through the Trop1cal Forages Program to the system w1de hvestock 1n1t1Bt1ve XIV CHAPTER 1 PROJECT TA 02 DYNAMICS OF LAND USE (Contnbutlon to Dynam1cs of Land Use TL01 TA021 Joyotee Sm1th (TLJ Manuel Wtnograd (GASEJ Doug/as Pach1co (/A} G1/berto Gallopm (LMJ w1th contflbutJons from Peter Jones (LMJ Rau/ Vera (TLJ Sam Fupsaka (LMJ V1cente CadaVJd (TLJ Antomo Brandlio and lgnez Lopes (FGVJ The purpose of th1s project 1s to develop technolog1cal opt1ons for arrestmg so1l nutnent deplet1on as well as pohcy opt1ons for promotmg stab1hzat10n of sh1ftmg cult1vat1on 1994 act1v1t1es focused on d1agnos1s The f1rst study reportad here g1ves a broad overv1ew of land use trends 1n the reg1on and 1mphcat1ons for pohcy and technology development The second study uses pnmary f1eld data for a deta1led mvest1gat1on of land use dynam1cs 1n two government colomzat1on projects 1n the western Amazon of Braz1l GIS work was also camed out 1n these areas and 1s reportad under project UT51 THE FOREST MARGINS ANO SAVANNAS A UNIQUE OPPORTUNITY FOR CONTRIBUTING TO NATURAL RESOURCE MANAGEMENT The object1ve of th1s study 1s to contnbute the soe~oeconom1c component towards developmg a strategy for the forest margms (FMI and savannas The work reportad here IS the f1rst step towards ach1evmg th1s obJectlve L1terature rev1ew and secondary data 1s used to synthes1ze an understandmg of past land use trends and new emergmg trends Prehmmary results of a land use model wh1ch s1mulates land use changes up to the year 2020 under alternat1ve pohcy and technology development strateg1es are presentad The results show that the con]unct1on of a number of nat1onal and mternat1onal developments makes th1s a h1ghly prop1t1ous t1me for mountmg a concertad effort for 1mprovmg the management of natural resources 1n the FM and savanna The v1s1on developed on the bas1s of th1s broad overv1ew w11i be progress1vely refmed 1n the next few years through 1n depth on farm stud1es 1n s1tes selected through GIS analys1s Savanna vs Hdls1de countnes The common percept1on about the three ecosystems 1n CIAT s ecoreg10nal mandate are qu1te d1stmct Agncultural expans1on 1n the FM 1s mternat1onally regarded as an env1ronmental catastrophe The H11is1des are cons1dered to be locked 1n a VICIOUS cycle m wh1ch poverty remforces env1ronmental degradat1on By contrast the savanna 1s regarded as the last agncultural front1er 1n the world w1th a very s1gmf1cant potent1al to be econom1cally explo1ted for agncultural and hvestock product1on We therefore broadly d1stmgU1sh between two groups of countnes The savanna countnes wh1ch have the poss1b1hty of developmg the savanna as a means of d1vertmg pressure on the FM and Hdls1des and the Hdls1de countnes wh1ch do not have a savanna ecosystem Macro data show that the savanna countnes have lower populat1on growth rates h1gher per cap1ta mcomes and h1gher rates of econom1c growth mcludmg agncultural output The mc1dence of poverty and food 1mports are also lower m savanna countnes However populat1ons are larger and the absoluta numbers of poor are h1gher Because of these d1fferences land use strateg1es are hkely to d1ffer among the two groups of countnes The work reported here focuses mamly on the savanna countnes Braz1l Colomb1a and Venezuela wh1ch conta1n over 90% of the total savanna area Land use dynam1cs H1stoncally the FM was occup1ed by md1genous sh1ft1ng cult1vators although boom and bust penods for extract1ve mdustnes have occurred sporad1cally dunng the last century Extens1ve cattle grazmg has been the trad1t1onal act1v1ty m the savanna smce the colomal penad The except1ons were the center south of the Braz1han savanna (C S) wh1ch was st1mulated by mfrastructure and the construct10n of Bras1ha m the 1950s mto becommg an 1mportant commerc1al center for cattle product1on and processmg and the western savanna m Venezuela where agncultural product1on was st1mulated by mcent1ves fmanced by 011 revenues Rap1d front1er expans1on occurred m both the savanna and FM from the 1970s In the Braz1han savanna (cerrado) the front1er expanded on an average by 2 1 m ha per annum between 1970 and 1985 w1th 70% of the cerrado mcorporated mto farms by 1985 In the Braz1han Amazon the rate of deforestat1on was 1 5 m to 2 m ha /year However there 1s st1ll considerable opportumty for 1nfluencmg the future pattern of land use as only 6% of the Braz1han Amazon has been cleared What 1s less well known 1s that the nsk of spec1es extmct1on 1s h1gher m the fore5ts of the cerrado than m the Amazon as a th1rd of the cerrado forests have already been cleared m many cases for charcoal product1on Pasture 15 the dom1nant form of land use m both ecosystems occupymg 60% to 82% of the cleared area m the cerrado and 60% to 70% 1n the Braz1han FM In the cerrado the 1mportance of pastura mcreases as agncultural development advances (Table 1 and 2) Concentrat1on m land ownersh1p 1s charactenst1c of both the savanna5 and FM In the cerrado 87% of the land 1s m farms rang1ng from 200 to > 1 0000 ha What 15 le5s well known 1s that there are substant1al numbers of relat1vely small farms More than 50% of farms are les5 than 50 ha m s1ze These small farms are mamly located lower down the topography often along the nver beds They play a cnt1cal role from the env1ronmental pomt of v1ew both because the nver beds are where the gallery forests are located and also because of the potent1al1mpact of the1r farmmg pract1ces on Slltat1on and pollut1on of nvers Over t1me land concentrat10n has remamed v1rtually stable though there has been a small decline m the proport1on of farm5 less than 50 ha and fragmentat1on of farms larger than 10000 ha In the FM land concentrat10n 1s lower than m the cerrado but st1ll h1gh w1th a Gm1 coeff1c1ent of O 79 m 1985 Small and large farms coex1st but there 1s a h1gh degree of soc1al confhct part1cularly between large scale absentee landlords and small scale squatters Sh1ftmg cult1vat1on by small holders 1s est1mated to account for 32% of trop1cal Amencan deforestat1on takmg place m 1980 There have been maror populat1on movements 1nto the FM part1cularly 1n the 1970s and substant1allabor absorpt1on m agnculture Movement of populat1on 1nto the savanna has been much less(Table 3) In advanced agncultural areas substant1al declines m rural populat1on have occurred s1multaneously w1th rap1d urban populat1on growth In the mumc1pahty of Uberlílnd1a m the C S of the cerrado rural populat1on m 1980 was 4 7% of the populat1on m 1950 However urban populat10n wh1ch compnsed 96% of the total populat1on m 1980 was more than 6 t1mes the 1950 level Presumably urban populat1on mcrease occurred because w1th 1mproved mfrastructure Uberland1a had developed mto a majar center for agnbusmes5 Whether or not the m1grants mto the urban areas carne from rural areas of the savanna IS however not known The d1vers1ty of cond1t1ons w1thm the savanna needs to be emphas1zed W1thm the cerrado for example the advanced agncultural area (C S) wh1le occupymg a th1rd of the are a contams 60% of the populat10n supphes 83% of the cerrado s soybean product10n and contams about a th1rd of 1ts cattle In the more remate areas extens1ve 2 cattle ranchmg IS the norm (Table 1) W1thm the FM a broad d1stmct1on 1s reqwred at m1mmum between areas such as Para (Eastern Brazll) where the government prov1ded mcent1ves for large scale ranchmg and areas such as Acre and RondOma (Western Braz1l) where government sponsored small sea le settlement too k place Small and large se ale farmers coex1st m both areas but the dynam1cs of land use have been d1fferent as Wlll be shown later In Colombia where most settlement was spontaneous the s1tuat1on 1s closer to areas of small scale settlement m Braz1l Technology development m the past has sought to mcrease land product1v1ty through the replacement of natural vegetat1on and trad1t1onal spec1es w1th new vanet1es chem1cal mputs mechamzat1on and 1rngat1on of dry areas A h1gh level of technology adopt1on has occurred m hm1ted areas of the savanna where h1gh land values good mfrastructure and prox1m1ty to markets made technolog1es of th1s natura worthwhlle In these areas agnculture 1s regarded as h1ghly successful The cerrado for example prov1des 2 5% of Braz1l s soybean output and contams a th1rd of the cattle populatlon The recent free trade pacts bemg negot1ated between latm Amanean countnes should st1mulate development m these areas as econom1c analys1s shows that the savanna has a comparat1ve advantage m hvestock and m soybean up to a d1stance of 2000 km from ports In the more remate areas of the savanna technology adopt1on has been very hm1ted and extens1ve cattle ranchmg 1s the norm l1ttle attent1on has also been g1ven to the development of technolog1es appropnate for small scale savanna farmers whose crops and resource endowments tend to be very d1fferent from those of larga scale farmers Tabla 1 land use m the Braz1han savanna (Cerrado)' Cerrado Modern Sub reg1on Rest of cerrado (C S)2 Total area (m ha) 155 50 105 1970 1985 1970 1985 1970 1985 Area w1thm farm 53 70 70 85 45 63 boundanes (% of total) Cleared area w1thm 23 47 28 61 20 37 farms (% farm area) % cleared area Crops (mcludmg plantad 16 19 20 19 12 19 forest) Plantad pasturas 45 61 69 75 31 46 Cleared but unused 39 20 20 6 57 35 1 Source Mueller et al 1992 2 Center South of Cerrado area around Uberland1a/G01ama See Mueller et al 1992 3 Tabla 2 Land use m the forest margms of Brazll Paré/Tocantms 1 Acre/Rondoma 1 1985 Agnculture 7 (annual/perenmal) Pastura 73 Secondary growth 4 6 years n/a 6 1 O years n/a > 10 years n/a Total 20 Forest n/a Source Mueller et al 1992 % mcorporated land 2 Source Moran et al 1994 4 6 years bush fallow 6 1 O years Woody spec1es > 1 O years el ose to mature forest n/a = not ava1lable 1985 % 29 57 n/a n/a n/a 14 n/a Altam1ra2 1985 1 1991 4 7 2 9 12 13 9 20 2 6 23 39 57 55 In the FM adopt1on of 1mproved pasturas has occurred but extens1ve management pract1ces remam the norm m large scale ranches Among small holders annual crops are grown w1th mm1mal externa! mputs In Braz1l y1elds of nce ma1ze and cassava have decreased sharply ovar the 1984 to 1990 penod m the FM m contrast to s1gmf1cant y1eld mcreases at the nat1onal level The env1ronmental effects of land use systems 1nclude fragmentat1on and reduct10n of natural hab1tats loss of ecosystems and spec1es changas m hydrolog1cal cycles s01l and water pollut1on compact10n and so1l eros1on assoc1ated w1th the 1ntens1f1cat1on of agnculture and s1gmf1cant emiSSions of green house gases due to deforestat10n Braz1l 1s now the fourth largest contnbutor to atmosphenc carbon m the world " 4 Table 3 Employment and demograph1c change m the forest margms and savanna of Brazll' Forest Margm Cerrado Paré/Tocant1ns 1 Acre/Rond6ma es 1 Cerrado (Total) Populat1on growth rata 1%1 19701980 84 11 4 34 27 1980 1991 49 63 25 2 Workers/1 00 ha of agr1cultural land 1970 85 85 7 6 1985 67 70 4 3 Source Mueller et al 1992 Expla1n1ng land use patterns The root cause of front1er expans10n has been 1dent1f1ed as the Latm Amencan style of development A large sect1on of the populat1on has been excluded from the benef1ts of growth and development Also macroeconomiC cond1t1ons have resulted m madequate and uncertam returns to fmanc1al assets (Table 41 Thus the front1er has been an escape valve for the d1scontent of the poor and the cap1tal of the nch lnequahty has 1ts roots m the h1gh levels of mechamzat1on m both agnculture and mdustry wh1ch hm1ted employment opportumt1es and led to large scale rural outm1grat1on from advanced agncultural areas The causes of th1s were h1stoncally h1gh levels of land concentrat10n government pohc1es such as negat1ve real mterest rates and subs1d1es for mechamzat1on and urban b1as m soc1al amemt1es wh1ch led to h1gh rates of urbamzat1on In add1t1on populat10n pressure m the hllls1des of the Andean countnes and Northeast Braz1lled to rural outm1grat1on wh1ch was remforced by government small scale settlement programs m partes of the Braz1han FM In Colombia people sought to flee from the poht1cal v1olence of the 1950s and 1960s These push factors caused the poor to look to the front1er as an escape valve (Figure 11 The cap1tal of the nch was drawn to the front1er because front1er land pnces contamed a speculat1ve component wh1ch pushed pnces well beyond actual product1on values Factors contnbutmg to th1s were the construct1on of penetrat10n roads government pohCies hnkmg land t1thng to product1ve land use access by landowners to cred1t at strongly negat1ve real mterest rates ( 25% to 35% between 1979 and 1986 m Braz1l) and opportumt1es for 1lhc1t cult1vat1on and processmg of coca (m Colombia) In add1t1on to th1s negat1ve real 1nterest rates h1gh unstable 1nflat1on (over 80% 1n Braz1l over 20% 1n Colombia and Venezuela m the 1980s) and protect10n of the bankmg sector reduced the return to fmanc1al assets and caused people to look to land as a hedge agamst mflat1on Thus the front1er was an escape valve for the cap1tal of the nch (Figure 2) 5 Table 4 Selected macro md1cators Brazli Colombia Venezuela Brazd Colombia Venezuela Per c6p1ta (constant 1980$) 1970 1112 897 4839 1980 2019 1125 4100 lnc1dence of poveny 1 1970 49 45 25 1980 39 39 22 llhteracy 1970 34 19 24 1980 26 12 15 lnfant Monahty (per 1 000) 1970 75 91 73 49 1980 85 71 41 39 Machanizat1on (No of ha/tractor) 1970 205 221 182 1980 90 183 99 Urbamzat1on (%) 1970 56 57 72 1980 68 64 83 1990 77 70 91 lnflat1on rate2 1971 12 18 4 1980 83 27 22 1990 2928 32 41 Households w1th mcome less than tw1ce the basket of bas1c foods 2 IMF lnternat10nal Fmanc1al Stat1st1cs 1981 Contralorla General de la Repubilca 1 994 Balance del cuatnemo Rev1sta Económica Colombiana (July August) Source CEPAL 1993 Anuano Estadlst1co de Aménca Lat1na y El Canbe Ed1c1ón 1992 Umted Nat1ons 6 Push Pull SubsJdiZed Roads mechn of ag/mdustry 1 ..... 1 Urban b1as 1 Front1er escape Demand for • valve for poor pasture land from ranchers 1 DegradatJon/pop pressure (NE) lncomes subst 1 h1gher than mm1mum wage F1gure 1 Causes of deforestabon Penetrat1on roads ~ lncreasmg land Front1er escape pnces valve for capital (1) 1' Cheap T1thng hnked cred1t to for land deforestat1on owners Hedge agamst mflallon F1gure 2 Causes of deforestatlon The front1er 1s charactenzed by abundance of land relat1ve to people h1gh transport costs wh1ch reduce the prof1tab11ity of econom1c act1v1ty mm1mal soc1al serv1ces and poorly enforced property nghts In v1ew of these charactenst1cs settlers at the extreme front1er are likely to have low levels of phys1cal and human cap1tal and therefore low opportumty costs G1ven the dlff1cult1es of operatmg a farm m the FM 1t 1s likely that these m1grants are from other rural areas as opposed to bemg the urban poor Small scale settlers m the extreme front1er of the FM clear a few hectares of land and grow nutnent demandmg annual crops such as nce and ma1ze for food and cash for a few years As fert11ity declines pasturas wh1ch are less nutnent demandmg are established and new land IS deforestad for annual crops As the front1er matures and transport and government serv1ces 1mprove people w1th h1gher levels of capital fmd 1t poss1ble to cover the1r opportun1ty cost G1ven the poor soc1al amemt1es these settlers are unl1kely to res1de at the front1er Extens1ve cattle ranchmg 1s the1r preferred form of land use as 1t reqUires low levels of mvestment and management Ranchers ach1eve savmgs m deforestat1on and pasture establishment costs are ach1eved by buy1ng out and consolidatmg the pasturas established by small scale holders Selling out to large scale ranchers also prov1des small holders w1th the cap1tal to acqu1re and deforest a new p1ece of land Data show a pos1t1ve relat1onsh1p between farm turnover and econom1c performance Thus m add1t1on to the push factors behmd m1grat1on of the poor to the front1er the pull factors mclude the demand for pastura land from large scale ranchers and the construct10n of penetrat1on roads wh1ch 1mprove access to front1er areas Recent data also show that mcomes of successful settlers are substant1ally above the m1mmum wage Thus land consolidat1on and trans1ence of land use and land users 1s a charactenst1c feature of the front1er m the FM and results m enormous env1ronmental costs both through deforestat1on and through degradat1on of pasturas on large scale ranches where land values are too low to 1ust1fy mtens1ve management The ava1lab11ity of new front1ers through the construct1on of new penetrat1on roads 1s a prereqUISite for th1s phenomena to operate Satelhte 1magery show that deforestat1on has been heav1est along these roads In areas where government mcent1ves for larga scale ranchmg ex1st land acqu1S1t1on and deforestat1on by people w1th h1gher opportun1ty costs occurs at the extreme front1er Absentee ownersh1p and poor enforcement of property nghts at the extreme front1er leads to squattmg by people w1th lower opportumty costs The result 1s mass1ve soc1al conflict and exacerbat1on of land degradat10n because of uncertamty of land tenure and because speculat10n and not product10n 1s the mam mot1ve for holdmg land The descnpt1on of the log1c behmd front1er expans1on shows that although pasture 1s the dommant form of land use m the FM much of the land under pasture was m1t1ally cleared for growmg annual crops lt also 1mplies that beef exports were not the dnvmg force behmd deforestat1on as has been allegad In fact data show that the Amazon barely produced enough beef for 1ts own needs In Central Amenca where exports were substant1al pasture expans1on and deforestat10n contmued after exports declinad m the 1980s The root causes of deforestat1on were opportumt1es at the front1er for land speculat1on and escape from poverty Once deforestat1on occurred pasture was the land use most m accordance w1th the log1c of the front1er The same front1er log1c leads to very d1fferent patterns of land use m the savanna The nature of the solls make 1t d1ff1cult for early settlers to make a livmg from small holdings Early settlers therefore go to the areas along the nver beds where the so1l 1s relat1vely more fert1le and grow crops such as cassava and beans wh1ch are more tolerant of low fert11ity Later settlers who arnve when the tront1er 1s more mature estabhsh extens1ve ranches on nat1ve savanna They do not b1d earlier settlers off the land as the topography near the nver beds 1s not conduc1ve to mechamzat1on and consolidat1on As the front1er matures m the savanna mtens1vely managed pasturas and f1eld crops replace extens1ve nat1ve pasturas Very large 9 farms are fragmentad to permtt more tntenstve management Small scale farmers wtth sorne capttal and entrepreneurtal abtltty become better tntegrated tnto markets and develop tnto medtum scale famtly farms buytng out margtnaltzed small holders who emtgrate because employment opportuntttes are mtntmal tn rural areas due to htgh levels of mechantzatton Employment opportuntttes tn agnbustness tn urban areas tn the savanna tncrease but whether or not rural mtgrants have the human capttal to take advantage of thts ts unknown Government tnterventton can alter these dynamtcs Substdtes for large scale ranchtng have led to better off mtgrants acqutrtng land at the extreme frontter for speculatton purposes As tn the FM the result ts squatttng soctal confltct and land degradat10n Substdtes such as untform fuel and output pnces have also led to expanston of tntenstve annual cropptng tn the extreme frontter These areas reverted back to extenstve ranchtng when substdtes were removed Thus we see that government poltcy tncludtng poltcy outstde the agncultural sector was a major determtnant of frontter expanston and land use patterns tn both the savanna and FM Emergtng trends In the last decade new trends tndtcate a reversa! of sorne of the most damagtng features of the past thus gtvtng nse toa untque opportuntty for effecttve resource management research Remete senstng data from vanous sources now tndtcate a major reductton tn deforestatton rates tn the Braztltan Amazon stnce the late 1980s A number of factors have reduced tncenttves for land speculatton among whtch are the removal of tncenttves whtch tncreased land pnces eltmtnatton of provtstons ltnktng land tttltng and deforestatton and tmprovement tn the returns to ftnanctal assets (Real tnterest rates tn the 1992 1994 pertod tn Braztl were as htgh as 23% to 43%) The constructton of penetrat10n roads has been reduced and government poltcy has changed from settlement to relattvely ttght control over deforestatton In Colombta acqutsttton of land tn the FM conttnued to provtde opportuntttes for money laundenng and culttvatton of tlltctt crops The push factors leadtng to smallholder mtgratton rematned unchanged However populatton growth rates tn the FM of both Braztl and Colombta decltned Thts occurred because of decltntng nattonal populatton growth rates and because of the development of ctttes whtch provtded an alternattve escape valve for the peor On the other hand there ts evtdence of an tncrease tn mmtng and loggtng acttvtttes whtch lead to penetratton roads whtch provtde access to later mtgrants Recent data on frontter expanston tn the savanna are not avatlable but the common perceptton ts that reduct10n of the speculattve mottve and the removal of untform fuel and output pnces has reduced frontter expanston tn the cerrado In Colombta by contrast otl dtscovertes tn the eastern savanna ts leadtng to major road constructton programs whtch are ltkely to lead to a major tncrease tn agncultural development The recent economtc cnsts tn Venezuela has destroyed conftdence although the currency devaluatton ts reportad to be tncreastng exports In the fteld of technology development new opportuntttes for tntegrated ecos,ystem management between ecologtsts and conventtonal technology developers such as sotl sctenttsts and germplasm spectaltsts are emergtng A ftndtng whtch ts parttcularly relevan! ts that ratn forests whtch ecologtsts had prevtously regarded as htghly fragtle and dtfftcult to rehabtlttate are now seen as restltent Data are now avatlable to tllustrate the tmportance of forest regeneratton after human tnterventton Satelltte tmagery show that tn Altamtra Braztl secondary growth ts the predomtnant form of land cover tn deforestad areas secondary growth ts tncreastng raptdly and 42% of new agncultural land was created by cleanng 10 secondary growth (Table 2 columns 3 and 4) Stud1es show that secondary growth can be managed to enable the regenerated forest to fulflll many of the ecolog1cal funct1ons of pnmary forests Th1s prov1des opportumt1es for combmmg econom1c gam w1th env1ronmental enhancement and may reqUJre majar changes m technology development strateg1es For example a majar focus of convent1onal technology development m the FM has been on preventmg secondary growth for example by stnvmg for pastura pers1stence These new fmdmgs show that secondary growth may have an 1mportant ecolog1cal funct1on and that an 1mportant focus of research m the FM should be the ennchment of secondary growth w1th faster grow1ng more valuable low volume spec1es wh1ch could make settlers mterested m regenerat1on and preservat1on of secondary forest as opposed to deforestat1on for crop and hvestock product1on The regenerat1ve capaCity of land use systems could be enhanced by for example mcreasmg carrymg capaCity w1th 1mproved pasturas wh1le s1multaneously lowenng the extmct1on threshold of spec1es through management pract1ces wh1ch protect seed banks from grazmg These concepts could lead for example to success1onal land use systems m FM w1th say annual croppmg followed by pasturas followed by forestry or sllvopastoral systems wh1ch whlle fulfllhng settlers objectlves would also fulflll ecolog1cal funct1ons Complementmg these new ecolog1cal parad1gms are advances m mformat1on technolog1es and opportumt1es tor knowledge mtens1ve as opposed to energy mtens1ve technolog1es through b1otechnology new matenals and new energy sources Another favorable development 1s that env1ronmental perspect1ves are becommg mcreasmgly 1mportant not only to the world commumty but also to nat1onal governments m latm Amanea as ev1denced by newly estabhshed m1mstnes of env1ronment and the mcreasmg 1mportance bemg g1ven to natural resource management m nat1onal research systems A number of mnovat1ve technolog1es wh1ch could be useful components m an mtegrated strategy for the management of these ecosystems have been developed Among them are ac1d tolerant nce and ma1ze a pers1stent and product1ve legume for grass legume pasturas and an mtegrated crop pastura system wh1ch m add1t1on to mcreasmg prof1tab1hty and 1mprovmg s01l phys1cal and chem1cal propert1es also appears to be actmg as a net smk for carbon and could thus have a majar role to play m m1mm1zmg the greenhouse effect An 1dea that 1s mcreasmgly bemg put torward 1s that developmg countnes may be able to benef1t trom a new comparat1ve advantage the prov1s1on of ecolog1cal serv1ces The 1dea 1s to use mternat1onal trade between countnes based on the1r comparat1ve advantage to ach1eve global ecolog1cal targets Thus a developed country government or pnvate sector company could purchase carbon storage serv1ces m developmg countnes m arder to meet reqUJred carbon em1ss1on standards 1f these serv1ces were cheaper than at home The Amazon ram forest prov1des Latm Amanean countnes w1th an asset capable of prov1dmg substant1al ecolog1cal serv1ces part1cularly the preservat1on of b1od1vers1ty and carbon storage Accordmg to an est1mate by the World Bank the value ot carbon storage m the Amazon ram forest ($976 to $7200/ha) 1s 2 to 30 t1mes the value of forest land even 1f the value of carbon 1n pasture 1s subtracted The ab1hty of crop pastura systems to act as a net carbon smk m the savanna prov1des another potent1al source of a comparat1ve advantage m ecolog1cal serv1ces There are s1gns that an mternat1onal carbon market m ay be developmg In v1ew of the agreement under the chmate changa convent1on to reduce carbon em1ss1ons to 1990 levels US ut1hty compames are mvest1gatmg opportumt1es for reducmg em1ss1ons through trappmg waste heat m lnd1a through reforestat1on m Mex1co and through modermzmg electnc1ty generat1on m Chma The Costa R1can government 1s developmg government guaranteed carbon storage cert1f1cates w1th the objectlve of tradmg them m 1nternat1onal markets Whlle part1C1pat1on 1n mternat1onal markets for ecolog1cal serv1ces could prov1de 1mportant mcent1ves for the adopt1on of ecolog1cally sound land use 11 systems problems ra1sed w1th th1s concept mclude Joss of nat1onal sovere1gnty and the h1gh transact1ons costs of ensurmg that ecolog1cal serv1ces are m fact prov1ded Many of these problems can be overcome by hav1ng markets m short term renta! contracts (as opposed to markets 1n Jand purchases) or by franch1smg contracts between local authont1es and the world commun1ty Such mechamsms enable nat10nal Jand owners to select the1r own Jand use strateg1es as long as they meet spec1f1ed env1ronmental standards As the contracts are short term and renewable they prov1de Jand owners w1th an mcent1ve to ensure that env1ronmental standards are met Research can make an 1mportant contnbut1on towards mcreas1ng comparat1ve advantage m ecolog1cal serv1ces Development of Jand use systems that mcrease mcomes wh1le s1multaneously prov1dmg ecolog1cal serv1ces (an example bemg the crop pasture technology) reduce the cost of prov1dmg ecolog1cal serv1ces and therefore mcrease comparat1ve advantage Conversely mcreasmg the prof1tab1hty of competmg Jand use systems wh1ch do not prov1de ecolog1cal serv1ces (say contmuous monocroppmg w1th mtens1ve t11lage) W11l mcrease the mcent1ve reqUired to mduce adopt1on of ecolog1cally sound Jand use systems and thus reduce comparat1ve advantage m the market for ecolog1cal serv1ces Thus research pnont1es need to be carefully coord1nated w1th the development of markets for ecolog1cal serv1ces The new developments descnbed above 1nteract to prov1de a un1que opportun1ty for effect1ve resource management (F1gure 3) Trans1ence of small settlers wh1ch 1s a ma¡or cause of deforestauon occurs because the cost of ga1mng access to a new plot 1s Jess than the mcome Jost because of y1eld dechne on the old plot Research can stab1hze ymld dechne on the old plot lncreasmg the cost of a new plot IS however d1ff1cult to ach1eve through research and 1f the cost of acqumng a new plot rema1ns low deforestatlon 1s hkely to contmue Changes m pohcy are however mcreasmg the cost of a new plot The construct10n of penetrat10n roads 1s bemg reduced New plots were f1nanced by selhng out old pasture plots to Jarge scale ranchers The demand for pasture Jand from Jarge scale ranchers IS however bemg reduced because the speculat1ve mot1ve for Jand acqU1s1t1on IS dechnmg as a result of pohcy changes At the same t1me reduct1on of mflat1on wh1ch benef1ts the poor should reduce the attract10ns of the front1er for small scale settlers The new ecolog1cal parad1gms also prov1de 1deas for technolog1es such as ennched secondary forests wh1ch could prov1de mcent1ves for long term settlement on the same plot Complementmg th1s 1s the emergmg market 1n ecologlcal serv1ces wh1ch 1f complemented w1th the nght technolog1es could prov1de 1mportant 1ncent1ves for the adopt1on of susta1nable land use pract1ces In add1t1on the savanna may be able to make an 1mportant contnbut1on towards reducmg deforestat1on As md1cated earl1er the savanna appears to have a comparat1ve advantage m certam crops and hvestock wh1ch could be further enhanced by recent technolog1cal advances lf technolog1es for susta1nable development of the savanna are adopted capital mvestment m savanna agnculture could d1vert cap1tal away from the FM and thus reduce the demand for large scale ranches 1n the FM In add1t10n 1mproved mfrastructure and mtens1f1cat1on of agnculture m the savanna could st1mulate the development of urban centers 1n the savanna wh1ch could prov1de employment opportun1t1es m agnbusmess and thus reduce the attract10ns of the front1er for the poor S1multaneously the savanna could contnbute to env~ronmental 1mprovement through carbon sequestrat1on by crop pasture Jand use systems whose adopt1on could be enhanced by the emergmg market m ecolog1cal serv1ces 12 Research lmprove rets to Remove subs1d1es reduce sust ag 1n 1nflat1on mcrease pnm/sec 1 cerrado edu New Ecol "' parad1gms/ tech rev Reduce demand for lncr opp cost of ranch1ng land m FM "' senlers Re use Sec "' growth 1 lncrease cost of Reduce gammg access to trans1ence new plot Reduce y1eld of small settlers decline on t old plot ~ Clase Front1er "' Reduce road building t + lnternat1onal trade 1n ecol Research Reduce loggmg serv1ces New ecol Sust land use parad1gms/ tech syst rev F1gure 3 A umque opportumty for resource management A vosoon for the year 2020 On the basos of our analysos of determmants of land use trends we have specofoed three scenanos for government pohcy and two technology development strategoes We then use a land use model te somulate land use changas up te the year 2020 for the FM and savanna m Brazol for dofferent combmat10ns of government pohcy and technology development strategoes Government Pohcy Scenano 1 Opemng up of the economy mcreases omports and puts pressure en mcreasmg agncultural exports Agncultural development of the cerrado os seen as a mechanosm for achoevong thos Expans10n of mfrastructure mto the cerrado contmues Ad hoc regulatory mechamsms are used to protect areas of the cerrado from agncultural expans10n but preve te be doffocult te enforce lnflatoon contonues te be unstable and hogh because of fallure to achoeve a sohd foscal posotoon Uncertamty about the return on fmancoal assets results m hogh speculatove demand for land The currency becomes overvalued and reduces the profotabohty of exports Urban boas contonues and leads te oncreased urbamzatoon resultmg m mcreased levels of urban cnme and unemployment Attempts are made to omprove oncome dostnbutoon by omposmg contractual obhgatoons en employers of agncultural and mdustnal labor But thos makes labor more costly and reduces employment Government becomes aware of the hnk between roads settlement and deforestatoon Constructo en of penetratoon roads onto the FM os slowed down subsodoes for large scale cattle ranchmg are removed Land tothng os no longer hnked te deforestatoon Mmmg and loggmg contmues because of lobboes and these actovotoes leave behmd access roads Government contmues to allow sorne road buoldmg m the Amazon because of mohtary reasons or because of local lobboes based on reasons of economoc development Scenaroo 2 Same as Scenano 1 except for foscal balance stable moderate mflatoon pohtocal and economoc stabohty whoch reduces the speculatove demand for land and benefots the poor through oncreasmg theor purchasmg power and creatong employment opportumt1es A stable real exchange rate reduces pressure on agncultural exports as a tool for balancmg the current account and mcreases the profotab1hty of exports Scenano 3 Same as scenaroo 2 plus soco al programs onented towards omprovmg the welfare of the poor such as greater access to promary and secondary educatoon omprovement of soc1al amemtoes m rural areas d1smanthng of contractual obhgatoons on employers of labor to encourage hogher levels of labor mtensoty These measures are expected to mcrease the opportumty cost te the poor of mogratmg to the extreme frontoer At mature frontoers there os better enforcement of property nghts and better provlsoon of socoal amen1toes to reduce absenteeosm squattlng and socoal confhct Constructoon of penetratoon roads os dommoshed te reduce the openmg up of new front1ers 1n both the FM and the cerrado but mtens1ve farm to market road networks are constructed m mature frontoer areas m both ecosystems wh1ch reduce the cost of 1nputs and the cost of marketing output Government servoces and 1nfrastructure are omproved m urban areas wothon mature frontoers m the FM and savanna Th1s stomulates local processmg and 14 d1stnbut1on of savanna and FM products These measures prov1de the pohcy env1ronment for the development of land use systems wh1ch are prof1table as well as ecolog¡cally sound lnternat1onal markets m ecolog1cal serv1ces develop and th1s makes the government aware of the concrete benef1ts of protectmg the env1ronment m both ecosystems Th1s creates the poht1cal w11l for mtegratmg env1ronmental cons1derat10ns w1th econom1c poht1cal and m1htary ob¡ect1ves for enactmg env1ronmental measures and for res1stmg mmmg and loggmg lobb1es Scenano 1 approx1mates the s1tuat1on that 1s hkely to preva1l 1n Braz¡J 1f the new stab1hzat1on plan fails Current developments however md1cate that the most hkely outcome 1s scenano 2 Scenano 3 dep1cts further pohcy changes wh1ch we would hke to see Colombia too 1s movmg towards scenano 2 The d1scovery of 011 m the eastern savanna 1s leadmg to ma¡or 1mprovements 1n mfrastructure and 1s hkely to lead to ma¡or advances m agncultural development m the savanna Th1s however could be senously ¡eopard1zed by rural secunty problems due to guernlla act1v1ty lf the government 1s unable to control drug traff1ckmg acqu1S1t1on of land for money laundenng and cult1vat1on and processmg of 1lhc1t drugs w1ll contmue and non product1ve mot1ves for land acqulsltlon w11l remam Another obstacle to the ach1evement of scenano 2 could be the appree1at10n of the currency m real terms wh1ch would reduce the compet1t1veness of exports and thus hmder the development of the savanna and 1mpede the creat1on of ¡obs 1n agnbusmess and m the nonagncultural sector Apprec1at1on of the currency m real terms 1s hkely 1f the government 1s unable to reduce mflat1on and repatnat1on of mcome from drug traff1ckmg H1gher mcomes from petroleum exports are hkely to contnbute to the upward pressure on the exchange rate The S1tuat1on m Venezuela 1s the least opt1m1stlc Unless ma¡or changes 1n pohcy are mtroduced the s1tuat1on 1n the short run 1s most hkely to approx1mate scenano 1 although the currency has deprec1ated sharply because of lack of conf1dence m the economy and 1s hkely to remam weak Th1s could ¡eopard1ze free trade agreements w1th other latm Amencan countnes and also adversely affect the exports of these countnes The economy 1s hkely to rema1n h1ghly regulated thus bu1ld1ng up senous structural ad¡ustment problems for the future These pohc1es are however unhkely to be sustamable and m the med1um term there could be s1gns of an emergmg scenano 2 Technology development stretegy Scenano A A component technology approach w1th research onented towards 1mprovmg the productiVIty of particular commod1t1es such as crops and hvestock Act1v1t1es mclude the development of agroecolog¡cally adapted vanet1es mcludmg vanet1es wh1ch reduce the reqUirement of externa! mputs management of b1ot1c and ab10t1c constra1nts to the mcreased product1v1ty of the mandate commod1t1es ident1f1cat1on of soc1oeconom1c constramts to technology adopt1on such as 1neff1c1ent markets Consequences of new technolog1es are mon1tored to avo1d the development of technolog1es wh1ch 1mpact negat1vely on non product1v1ty aspects such as eqUity and nutnt1on The mam focus 1s on technolog1cal solut1ons related to mandate commod1t1es w1th other components pohe~es and mst1tut1ons taken as g1ven Scenano B Th1s 1s the ecoreg1onal approach where the mandate 1s an agroecolog1cal zone w1th1n a geograph1cal reg10n The ob¡ect1ve 1s to 1dent1fy an appropnate role for agroecosystems w1thm the ecoreg10n takmg nat1onal and mternat1onal cons1derat1ons mto account and to st1mulate 15 the research requrred to achreve that vrsron The rnteractron between technology pohcres and rnstrtutrons rs exphcrtly recognrzed and recommendatrons for pohcy and rnstrtutronal reform are an rntegral part of the approach The approach therefore requrres close collaboratron between drfferent categones of research bodres rncludrng natronal and rnternatronal rnstrtutes and the pnvate sector as well as rmplementors of research frndrngs such as NGOs and polrcy makers In addrtron to productrvrty and equrty protectron of the envrronment rs an rmportant objectrve rncludrng the rdentrfrcatron of ecosystems where agncultural explortatron rs rnappropnate Forest Margrn In thrs ecosystem the overall strategy rs to reconcrle the rnternatronal communrty s objectrve of haltrng deforestatron wrth the reluctance of natronal governments to sacnfrce the economrc potentral of the FM for the sake of ecologrcal benefrts that accrue to the world communrty as a whole Technology development wrll be based on a land use plan for the FM whrch explorts the drversrty of envrronments wrthrn the FM rncludrng both cleared areas and surroundrng forest areas Thrs wrll be based on a sound understandrng of land use dynamrcs and wrll rnclude pohcy recommendatrons on rssues such as rnfrastructure protected areas reforestatron schemes and partrcrpatron rn global markets for ecologrcal servrces Technology development wrll rnclude a) lntensrfrcatron and stabrhzatron of agnculture on small holder plots partrcularly food crop plots to reduce transrence Research now suggests that contrnuous cropprng rs possrble wrth fertrlrzer use lmproved farm to market roads wrll however be requrre to make thrs economrcally vrable Penetratron road constructron wrll have to be halted to make these technologres adoptable b) The desrrabrhty of rntensrfyrng and stabrhzrng small holder hvestock productron wrll have to be carefully consrdered as aspects of these technologres may also be apphcable to extensrve large scale cattle ranches and m ay therefore encourage land consohdatron and promote deforestatron Technologres whrch rmprove mrlk productron such as genetrc rmprovement of darry cattle specres are less hkely to have thrs effect el Successronalland use systems on small holder plots wrth pasturas managed to encourage ennched secondary growth and the development of agroforestry or agrosrlvopastoral systems Thrs may anchor small holders to therr plots di Study of rndrgenous systems for rnsrghts rnto agncultural systems compatrble wrth forest regeneratron e) lncreased profrtabrlrty of sustarnable forestry systems non trmber forest products rubbertapprng sustarnable selectrve loggrng and the management of these systems to enhance therr capacrty to provrde ecologrcal servrces fl Regeneratron of degradad areas wrth reforestatron agroforestry or agrosrlvopastoral systems g) Management of secondary growth to reduce the cost of recultrvatron versus the cost of movrng to a new plot 16 hl Labor mtens1ve agnbusmess technolog1es 1n urban areas of the FM such as the processmg and d1stnbut1on of FM products such as non t1mber forest products and mllk Th1s 1s expected to mcrease the prof1tab1hty of these commod1t1es wh1le prov1d1ng alternat1ve employment opportumt1es for potent1al deforesters Savanna In th1s ecosystem the overall strategy 1s to prov1de an outlet for the government s econom1c ob]ect1ves thus rehev1ng pressure for explo1t1ng the FM for agnculture and extract1ve act1v1t1es Rural and urban development 1n the savanna 1s al so v1suahzed as d1vert1ng venture cap1tal from the FM wh1le agnbusmess development may absorb m1grants from rural areas who may otherw1se have m1grated to the FM The ob]ect1ve 1s to ach1eve th1s development whlle protectmg the env1ronment Th1s 1s part1cularly 1mportant as the savanna contams sem1dec1duous and gallery forest and 1s part of the watershed of three ma]Or nvers Thus savanna development could have env1ronmental1mpacts both w1th1n and outs1de the savanna As 1n the FM technology development w11t be based on a land use plan denved from a sound understandmg of land use dynam1cs Technology development w111 mclude al lncrease returns to cap1tal mvestment 1n the savanna by developmg a range of technolog1cal opt1ons w1th progress1vely mcreasmg levels of cap1tal and management 1ntens1ty targeted to a range of env1ronments varymg from moderate to h1gh levels of mfrastructure and land values The extens1ve end of th1s contmuum mcludes technolog1es for managmg nat1ve savanna grasses w1th or w1thout small are as of 1m preved pasture At the mtens1ve end 1t mcludes mtegrated ley farmmg systems w1th technolog1calmnovat10ns enhancmg the1r prof1tab1hty and carbon sequestrat1on capac1ty S1multaneously technology development on competmg land use systems wh1ch lead to soll degradat1on and do not prov1de ecolog1cal serv1ces (such as monocroppmg w1th contmuous heavy t1llage) w11t be de emphas1zed bl Development of technolog1es for susta1nable forestry 1n areas of ex1stmg sem1 dec1duous forests Agroforestry planted forest frUit and nut trees agros1lvopastoral systems part1cularly 1n areas where destruct1on of sem1 dec1duous forests has taken place Emphas1s w11t be on both prof1tab1hty and ecolog1cal funct1ons Th1s 1s expected to enhance adopt1on by mcreasmg comparat1ve advantage 1n markets for ecolog1cal serv1ces el Technolog1es for recuperat1on of degraded areas mcludmg ley farmmg systems 1n mtens1f1ed areas di lnvest1gat1on of the off s1te effects of mtens1f1cat1on such as s1ltat1on and pollut1on of nvers technolog1es to reduce the1r 1mpact and lnStltUtiOnal mechamsms for mternahzmg externaht1es el Labor mtens1ve technolog1es for processmg and d1stnbut1on of savanna products 1n urban areas of the savanna fl Technolog1es for commod1t1es 1n wh1ch the relat1ve resource endowments of small scale farmers have a comparat1ve advantage mclude vegetable product1on da1ry1ng and subs1stence product1on of cassava and beans Th1s IS expected to prevent outm1grat1on of small scale farmers Particular emphas1s w11t be placed on the env1ronmental effects of smalt scale farmmg because of the1r locat10n along nvers and gallery forests Technolog1es 17 for sustamable explo1tat1on of gallery forests and for mcreasmg the role of tree crops 1n systems 1n deforestad areas wlll be developed w1th emphas1s on the1r compat1b1hty w1th the resource endowments of small scale farmers gl Land use planmng w1ll 1dent1fy areas whose mam funct1on w1ll be the preservat1on of plant and an1mal b10d1vers1tY ldeally these wlll comc1de w1th remote areas and areas of low b1ophys1cal potent1al for agncultural expl01tat1on Technology development m these areas Wlll be hm1ted to recuperat1on of the ecolog1cal funct1ons of degradad areas and 1dentlf1cat1on and evaluat1on of the ecolog1cal serv1ces prov1ded by nat1ve grasses sem1 dec1duous and gallery forests Prevent1on of agncultural explo1tat1on of these areas w1ll reqUire the absence of penetrat1on roads Part1c1pat1on of these areas m markets for ecolog1cal serv1ces may contnbute to the1r preservat1on as may ecotounsm Technology development strategy A approx1mates past h1story The adopt1on of ecoreg10nal mandates by mternat1onal research centers m the reg1on the new emphas1s on resource management 1ssues m nat1onal research systems and NGOs m the reg1on md1cate however that the reahzat1on of strategy 8 1s h1ghly probable prov1ded sustamed donor comm1tment to act1V1t1es of th1s nature are avallable G1ven the long term nature of most of these act1v1t1es the 1dentlf1cat10n of concrete 1ntermed1ate outputs w11l be essent1al for obtammg sustamed fund1ng SIMULATING LAND USE CHANGE A s1mulat10n model d1v1des land m each ecosystem 1nto d1fferent ecolog1cal categones w1th d1fferent product1ve capac1t1es and d1fferent ecolog1cal funct10ns Movement of land between d1fferent categones 1s s1mulated the rate of change bemg defmed by the scenanos ldeally the model should contam a soc1oeconom1c sub model for s1mulatmg the rate of change m land use categones under each scenano Th1s 1s not ava1lable at the present t1me and rates of response to pohcy measures and technology development are denved from the JUdgement of the authors As deta1led f1eld work 1s carned out data w1ll be ava1lable for more accurate est1mates of rates of change The model 1s run up to the year 2020 for each combmat1on of pohcy and technology development scenanos A companson between scenanos 3A and 1 A shows what pohcy 1mprovements can ach1eve w1th a component technology approach The d1fference between 1 A and 1 8 shows what a sh1ft to an ecoreg1onal approach can ach1eve m the absence of pohcy 1mprovements 38 1s the 1deal scenano wh1ch shows what the ecoreg1onal approach can ach1eve 1n the presence of favorable pohcy The d1fference between 3A and 38 shows the Incremental contnbut1on of the ecoreg1onal approach over the contnbut1on of pohcy Prehmmary results are presentad In partiCular 1t should be p01nted out that results on only broad categones of land use change are avallable at th1s t1me Analys1s of the d1saggregat1on of these categones and the env1ronmental and soc1oeconom1c consequences are st1ll underway Thus wh1le results on changes 1n altered land are ava1lable analys1s of 1ts breakdown 1nto categones such as regenerated forest abandonad land and the consequences on the env1ronment employment etc are st1ll under way At th1s stage the model has been run separately for the cerrado and FM w1thout mteract1ons between the two although we hypothes1ze that these are 1mportant 18 The results ava1lable so far show the followmg Cerrado al Front1er expans1on 1s most rap1d under scenano 1 A because of speculat1ve mot1ves for holdmg land and because the pressure to mcrease exports leads to the construct1on of penetrat1on roads By 2020 only 6% of the land 1s m 1ts natural state Much of the expans1on occurs mto margmal areas unsUitable for agncultural explo1tat1on leadmg to resource degradat1on Front1er expans1on declines m 2A and 2B because the speculat1ve mot1ve declines and further m 3A and 3B (15% to 19% of the land areal when penetrat1on road construct1on declines because governments become aware of the value of env1ronmental serv1ces In 3B there 1s a qualitat1ve 1mprovement m the natural area because the land use plan 1dent1f1es the most appropnate areas from both the conservat1on and econom1c pomts of v1ew bl Gram product1on mcreases sharply m all scenanos The mcrease 1s least m 1 A (from 6 2 m t to 14 1 m ti because the speculat1ve mot1ve leads to land degradat1on Although recuperat1on technolog1es are developed m 1 B they are not adoptad because speculat10n and not product1on 1s the mam mot1ve for land acqu1S1t1on Product1on mcreases are h1gher m 2A and 2B and h1ghest m 3A and 3B (23 m t to 30 m ti because land values mcrease due to the closmg of the front1er and technolog1es for recuperatmg land are ava1lable In add1t1on m 3B the mcrease 1s ach1eved w1th a reduct1on 1n offs1te effects because of the ex1stence of technolog1es and mst1tut1onal mechamsms for mternalizmg externalit1es el Plantad pasturas mcrease from 26 m ha to about 50 m ha and sequester 150 to 750 m t C/ha The number of cattle per head mcreases most m scenanos 1 A and 1 B (from 32 to around 70 m headsl because of rap1d front1er expans10n In 3A and 3B cattle numbers are lower (about 60 m heads) but stockmg ratas and product1on per ammal are h1gher because of polic1es such as 1mproved farm to market roads and the development of appropnate technolog1es di By 2020 wood product1on 1s h1ghest m 3B (81 m ti because of the ex1stence of 18 m ha of agros1lvopastoral systems wh1ch store 85 t C/ha In the A scenanos very little adopt1on occurs because the component technology approach g1ves msuff1c1ent attent1on to understandmg the dynam1cs of land use systems and 1s therefore unable to 1dent1fy appropnate opportumt1es for technolog1cal mtervent1ons In 1 B and 2B although the ecoreg1onal approach leads to the development of appropnate technolog1cal mtervent1ons adopt1on does not occur because of unfavorable polic1es wh1ch lead to land speculat10n and front1er expans1on and therefore promete short term1sm Also 1n the A scenanos the wood product1on comes from deforestat1on whereas m 3B 1t 1s from sustamable systems Forest Marg1n al W1th favorable polic1es and an ecoreg1onal technology development strategy deforestat1on 1s 37% lower 1n 2020 than 1t would be w1th policy scenano 1 and a component technology development approach bl W1thout favorable polie~es the sh1ft to the ecoreg1onal approach ach1eves little change m land use 1n quant1tat1ve terms Under policy scenano 1 the area of natural ecosystem (275 19 m ha) IS unchanged when the ecoreg10nal approach 1s mtroduced because of the speculat1ve mot1ve and the lack of pohc1es to mcrease the opportumty cost of potent1al m1grants to the front1er e) In scenano 38 19 m ha of altered land are converted to a vanety of tree crop based systems Forest regenerat1on also mcreases sharply a1ded by the new ecolog1cal parad1gms The ecoreg1onal approach Jdent1f1es appropnate systems wh1ch meet ob1ect1ves of settlers wh1le at the same t1me prov1dmg ecolog1cal serv1ces Adopt1on occurs because farm to market roads mcrease profltabihty and because partlclpatlon m markets for ecolog1cal serv1ces prov1des a further mcent1ve for adopt1on Th1s enables nat1onal governments to obtam econom1c benef1ts from the FM wh1le at the same t1me contnbutmg to env1ronmental 1mprovement CONCLUSIONS A thorough understandmg of land use dynam1cs 1s essent1al for developmg a resource management strategy Analys1s of land use change 1n the FM and savanna of Latm Amenca based on secondary data and literatura rev1ew showed that many of the dnv1ng torces ongmate at the nat10nal pohcy level both w1thm and outs1de the agncultural sector as well at the global level An essent1al component of a resource management strategy for these ecosystems therefore should be recommendat1ons for pohcy and mstltutJonal reform The analys1s al so revealed that for effect1ve ecosystem management technology development will have to move away from the convent1onal approach of developmg product1ve components such as crops hvestock and tree spec1es to a broader approach that mcludes aspects such as forest regenerat1on management of secondary growth non t1mber forest products ldent¡f¡catlon of conservat10n areas development of agnbusmess technolog1es and enhancement of comparat1ve advantage m markets for ecolog1cal serv1ces A land use s¡mulat1on model revealed that even th1s broader approach to technology development 1s unhkely to be effect1ve m the absence of favorable econom1c soc1al and env1ronmental pohc1es Fortunately however many unfavorable pohc1es have been d1smantled deforestat1on has sharply decreased and speculat1ve front1er expans1on m the savanna has decreased part1cularly m Braz1l Re1nforcmg th1s 1s the nse of new ecolog1cal parad1gms wh1ch prov1de new opportumt1es for combmmg env1ronmental 1mprovement w1th product1ve land use Other favorable developments are the newly estabhshed resource management programs m nat1onal research systems and Mm1stnes of the Env1ronment m Latm Amenca A development of ma1or s¡gnlflcance 1s the emergmg mternat10nal trade m ecolog1cal serv1ces Both the savanna and FM of Lat1n Amenca appear to have a comparat1ve advantage m th1s f1eld and 1f supported by an appropnate technology development strategy th1s could be a key mechan1sm for mducmg adopt1on of env1ronmentally sound pract1ces While these developments go a long way towards prov1dmg a favorable background for effect1ve ecosystem management further pohcy and ¡nst1tut10nal changas are reqUJred such as 1mproved access to pnmary and secondary educat1on for the poor 1mproved farm to market roads control of drug traff1ckmg 1mprovement of peace and arder and the poht1cal w1ll to Wlthstand the pressures of m1nmg and loggmg lobb1es Our land use model md1cates that these pohcy and mst¡tutJonal changes and a bread based ecoreg1onal technology development strategy may make 1t poss1ble for nat1onal governments to obtam econom1c benef1t from the FM while contnbutmg globally to the protect10n of the env1ronment The strategy for the savanna should be developed m clase conjunctiOn w1th that for the FM 20 The analys1s shows that substant1al mcreases m crop hvestock and tree products are poss1ble from the savanna Thus the savanna could prov1de an outlet for the econom1c ob¡ect1ves of nat1onal governments and for venture cap1tal wh1le rehevmg pressure for explo1tmg the FM A carefulland use plan w11l be essent1al for ach1evmg th1s w1th accelerated rural and urban development m certam areas for attractmg venture capital combmed w1th labor absorpt1on m agnbusmess m urban areas of the savanna and the prevent1on of agncultural expans1on m remote margmal areas The strategy w11l have to be based on a sound understandmg of the ecologlcallmphcatlons of savanna mtens1f1cat1on as the savanna 1s part of the watershed of three ma¡or nvers and 1t 1s part1cularly notable that the pract1ces of small scale farmers may play a ma¡or role m these externaht1es Overall our analys1s md1cates that the present constellat1on of events relatmg to the savanna and FM of Latm Amenca 1s part1cularly favorable for ach1evmg ma¡or 1mprovements m the management of these ecosystems We therefore have a umque opportumty for successful and effect1ve resource management research TA02 CHARACTERIZATION OF SLASH ANO BURN AGRICULTURE ANO DEFORESTATION IN PEDRO PEIXOTO ACRE ANO THEOBROMA RONDONIA BRAZIL Sam Fupsaka Trop1cal deforestat1on due m part to slash and burn agnculture contnbutes to global warmmg v1a burnmg and release of C02 mto the atmosphere and Braz1l1s now the fourth atmosphenc carbon contnbutor after the US ex Sov1et Umon and Chma Deforestat1on 1s also leadmg to losses of genet1c and cultural d1vers1ty Decreasmg transp1rat1on and prec1p1tat1on w1thm and outs1de of areas cleared may also be a consequence of deforestat1on W1th fundmg from the Global Env1ronmental Fund (GEF) a pro¡ect Alternat1ves to Slash and Burn (ASB) was 101t1ated to address such problems and as th1s paper reports farmer settlers m the government colomzat10n pro¡ects of Pedro Pe1xoto m the Amazoman state of Acre and 1n Theobroma Rondoma were mterv1ewed as a part of act1v1t1es to charactenze local systems and system dynam1cs Fundmg for th1s stage of systems charactenzat10n was also prov1ded by the lnter Amencan Development Bank The pro¡ect currently 1ncludes s1tes m Cameroon and Indonesia and 1s a collaborat1ve effort among nat1onal research systems non government orgamzat1ons and at the 1nternat1onal level the lnternat1onal Centre for Research on Agroforestry (ICRAF) CIAT the lnternat10nal Food Pohcy Research lnstltute (IFPRI) and the Trop1cal S01l B1ology and Fert1hty Programme (TSBF) The pro¡ect seeks to develop techmcal and pohcy alternat1ves to slash and burn agnculture wh1ch would have effects of decreasmg global rates of deforestat1on whlle enhancmg the well bemg of forest users Settlement m colomzat1on pro¡ects such as Pedro Pe1xoto and Theobroma has been faci11tated by government pohc1es to populate front1er areas m the Amazon road construct1on and d1rect and md1rect subs1d1es Data presentad m compete form under Pro¡ect UT04 addresses the dynam1cs of deforestat1on at the farm level 1 e farmer dec1s1on makmg m hght of constramts and opportumt1es afforded by the local agroecosystem and w1thm the context of relevant local reg10nal nat1onal and mternat10nal pohe~es The conclus10ns of that research were 21 Deforestat10n at the farm level appears to contmue at a relat1vely steady pace averagmg somewhat more than one ha per year per fam1ly m the two colomes Settlers st1ll had more than half of the1r lands 1n forest and for the most part can and probably w1ll contmue to slash burn and cult1vate more pnmary forest as the1r currently most (econom1cally) v1able opt1on Two mam factors dnvmg land cleanng at the farm level were the need to produce food crops and mcent1ves to convert land mto pasture In terms of food product10n farmers consumed and sold nce and to a lesser extent beans ma1ze and cassava (or cassava meall R1ce cult1vat1on may dnve sorne deforestat1on m that although farmers usually planted a cleared f1eld for two or three years they could not for techmcal reasons sow nce other than m the f1rst year after cleanng As such a research pnoflty of the ASB proJect may be to determme 1f and under wh1ch management alternat1ves could nce product1on be made more sustamable (acknowledgmg the caveat that any 1mprovement m productiVIty may lead to greater econom1c attract1veness of the respective enterpnse and thereby mv1te more deforestat1on or other forms of resource over expl01tat10n) Farmers were clearly mot1vated to convert cleared lands mto pasture because of real or at least perce1ved resultmg mcreases m land values Farmers not only mamta1ned cattle as standmg bank accounts and obtamed cash from sales of ammals and m1lk but bUIIt savmgs L by 1nvest1ng t1me and resources m fencmg corrals ponds and other ranch1ng necess1t1es As observad throughout the two s1tes local ranchers and urban based speculators have purchased cont1nuous blocks of colon1sts parcels to form new ranches or to expand the s1ze of adjacent ranches and payments were reportedly much hlgher for cleared vs forestad port1ons of parcels Farmers pasture management pract1ces cons1sted of mtroducmg mamly Brach1ana spp annual pasture burnmg and rotat1on of ammals to d1fferent pasturas In sp1te of generally low stockmg rates there were substant1al areas of poor and degradad pasturas and pasture lands at both s1tes Although construct10n of new cheese processmg plants near Theobroma may result m mtens1f1cat1on at that s1te extens1ve and low mput cattle and pasture management appear to be the current norm for both s1tes Whether or not 1mproved pasture technolog1es are poss1ble and would be appropnate to farmers current cond1t1ons 1s another researchable ISSU8 Convers1on of cleared forest land to pasture has meant that at least m the colon1es stud1ed that there was relatlvely httle land placed m fallow w1th resultmg secondary regrowth and forest regenerat1on An 1mphcat10n 1s that research on 1mproved fallows m ay e1ther result m mak1ng fallows more attract1ve to farmers (as 1s hoped) because of more rap1d regenerat1on and better ma1ntenance of so1l orgamc matter or farmers may be unmterested 1n 1mproved fallows because convers1on to pasture 1s by far the preferred land use after cult1vat10n Theobroma farmers pract1ced agroforestry m the sense that they had s1gmf1cant areas of perenn1al crops mamly coffee and cacao Pedro Pe1xoto settlers have had less favorable expenences regardmg perenn1als c1tmg poor pnces and/or markets as majar constramts Sorne Pedro Pe1xoto farmers had been encouraged (prov1ded w1th cred1t) to plant urucu (81xa orellana u sed to produce red dye) When the trees started to produce pnces for the product fell to the pomt that farmers could not afford to harvest and lost the1r mvestments Currently settlers m Theobroma were be1ng prov1ded cred1t to grow acerola (Malp1gh1a pumc1foha) Although they were prom1sed a future market for all that they produce schemes to mtroduce or encourage perenmal crops or agroforestry remam nsky 22 A resource that was largely not ava1lable m the systems exammed 1s md1genous techmcal knowledge of the type usually assoc1ated w1th sh1ftmg cult1vators Farmers were asked about the1r s01l and land class1f1cat1on and correspondmg use systems Although they d1stmgU1shed so1ls by color and texture and called attent1on to lands e1ther havmg a sub surface compactad layer or low waterlogged areas they d1d not employ such d1stmct1ons m choosmg areas to clear and cult1vate Farmers s1mply cleared land m steady sequence from the areas closest to the roads and houses and towards the rear of the1r parcels Although a few named plant spec1es wh1ch 1nd1cated s01l 1mpovenshment and others wh1ch md1cated fallow regenerat1on most farmers appeared to have httle concept of use of fallows for b1omass (and subsequent soll fert1hty) regenerat10n or of (as ment10ned) the use of f1re to release nutnents for crops use Only a few Theobroma farmers ment1oned that there were med1cmal plants that could be harvested from the forest The heartemng news 1s that rates of deforestat1on m the Amazon Basm appear to be decreas1ng due to fewer mcent1ves to large corporat1ons to mvest m cattle ranchmg a v1rtual end to a penod of road bu1ldmg to open up the Amazon and to protect front1er areas and a decline m government ass1sted colomzat1on programs for the rural poor Secondary forests have shown relat1vely h1gh rates of regenerat10n both after slash and burn agnculture and after abandonment of degraded pastures and a s1gmhcant proport1on of deforestat1on for sh1ftmg cult1vat1on 1s now of secondary rather than pnmary forest On the other hand deforestat1on and/or consohdat10n of colomsts parcels to form or expand ranches has contmued w1th ranches now bemg formed less by front1er nsk takers and more by urban speculators m areas where land pnces have nsen and where government can protect such mvestment The ev1dence presentad suggests that deforestat1on on colomsts parcels contmues at an apparent steady pace dnven by food product10n needs and by mcent1ves s1m1lar to those now push1ng format1on of larger ranches To sorne extent the charactenzat10n data suggest the need for on farm research to develop alternat1ves to slash and burn agnculture Such research already underway or planned needs to examme the poss1bil1ty of makmg nce product10n more sustamable (and thereby hopefully reducmg demand for newly cleared forest) 1mprovmg (1 e 1ntens1fymg to the degree appropnate) pasture and cattle management mtroduc1ng or encouragmg more perenn1al croppmg and agroforestry and 1mprovmg fallows At the same t1me and equally or more 1mportantly th1s analys1s recogmzes the 1mportance to deforestat10n of nat1onal pohc1es regardmg road construct1on cred1t tax mcent1ves front1er settlement land tenure and the rural peor Development of alternat1ves w1ll need to mclude substant1al attent1on to pohc1es and future pohcy opt1ons 23 COMPLEMENTARY SUBPROJECTS UT51 (UT04l A OIAGNOSTIC STUOY OF AGRICUL TURAL LANO USE IN THE SOUTHWEST BRAZILIAN AMAZON ANO TA51 (TA021 ALTERNATIVES TO SLASH ANO BURN Sam Fupsaka (LMJ The purpose of CIAT s work 10 the forest margtns 1s to develop opttons for arrestmg sotl nutnent depletton as well as policy opttons for promottng stabtllzatton of shtfttng culttvatton The purpose of UT51 (a complementary subpro¡ect of UT04 Ecoregtonal Research and Exploratory Acttvtttes ) 1s to analyze the soctal cultural pollttcal and ecologtcal factors that shape the development of land use patterns m colontzed areas of the states of Acre and Rondonta Braztl The donar for thts pro¡ect ts IDB Research partners tnclude EMBRAPA (tn Acre and tn Rondonta) PESACRE ICRAF and IFPRI (Steve Vostt) Thts subpro¡ect was funded for 1994 1995 only The purpose of TA51 a complementary subpro¡ect of TA01 Prototype Sustatnable Croppmg Systems for the Forest margms ts to destgn sustamable agrostlvopastoral systems for deforestad areas of the Amazon Bastn Thts report covers research on 1nt1al system charactenzatton betng conducted at each of three Sttes tncluded 1n the Global tntttattve on Alternattves to Slash and Burn Agnculture (ASB) funded by GEF and coordmated globally by ICRAF and by CIA T tn Lattn Amenca The purpose of the ASB ts to develop techntcal and pollcy alternattves whtch would help to decrease rates of troptcal deforestatton whtle enhanctng the well betng of forest resource users Research partners tnclude EMBRAPA (tn Acre and tn Rondonta) PESACRE ICRAF IFPRI (Steve Vostt) CIFOR and TSBF Troptcal deforestatton due tn part to slash and burn agnculture contnbutes to global warmmg vta burnmg and ralease of C02 mto the atmosphere and Braztl 1s now the fourth atmosphenc carbon contnbutor after the US ex Sovtet Unten and Chtna Deforestat10n ts also leadtng to losses of genettc and cultural dtverstty Decreastng transptratton and prectpttatton wtthm and outstde of areas cleared m ay al so be a consequence of deforestatton Wtth fundtng from the Global Envtronmental Fund (GEF) a pro1ect Alternattves to Slash and Burn (ASB) was tntttated to address such problems and as thts paper reports farmer settlers 10 the government colontzatton pro¡ects of Pedro Petxoto In the Amazontan state of Acre and tn Theobroma Rondonta were tntervtewed as a part of acttvtttes to charactenze local systems and system dynamtcs The pro¡ect currently tncludes sttes tn Camaroon and lndonesta and ts a collaborattve effort among nattonal research systems non government organtzattons and at the tnternattonal level the lnternattonal Centre for Research on Agroforestry (ICRAF) CIAT the lnternattonal Food Pollcy Research lnstttute (IFPRI) and the Troptcal Sotl Btology and Ferttllty Programme (TSBF) The pro¡ect seeks to develop techntcal and pollcy alternattves to slash and burn agnculture whtch would have effects of decreasmg global rates of deforestatton whtle enhancmg the well betng of forest users Settlement tn colontzatton pro¡ects such as Pedro Petxoto and Theobroma has been factlltated by government pollctes to populate frontter areas 10 the Amazon road constructton and dtrect and tndtrect substdtes Data presented tn thts paper addresses the dynamtcs of deforestatton at the farm level 1 e farmer dectston maktng tn llght of constramts and opportuntttes afforded by the local agroecosystem and Wtthtn the context of relevant local regtonal nat10nal and mternattonal pollctes INTRODUCTION DEFORESTATION IN THE AMAZON Rates of Deforestat1on Compared to As1a and Afnca deforestat1on has been h1ghest m Latm Amenca 1n both absolute area (43 000 km 2/year) and percent of forest area cleared (0 64%/year) Of Lat1n Amencan forests standing 1n 1850 370 mlihon ha or 28% were converted by 1985 44% to pasture 25% to cropland 20% degraded and 10% to sh1ftmg cult1vat1on Recent analys1s of satelhte data 1nd1cates that some 230 000 km2 of the Braz1han Amazon 1s deforested and that 588 000 km' are affected 1f edge effects of one km mto ad1acent areas of forest are cons1dered Deforestat1on 1n the Braz1han Amazon has been h1ghest m Rondoma and Mato Grosso followed by western Maranhao Acre and northern Go1as Sh1ftmg cult1vat1on (slash and burn agnculture) was thought to account for 1 8 m1lhon ha or 32% of tropical Amencan deforestat1on takmg place 1n 1980 Causes of Amazon1an deforestat10n 1960 1985 lnterhnked factors contnbutmg to deforestation up to a decade ago mcluded road building land conversion to pasture and cattle ranchmg (and associated pohcies) demand for land by the rural poor cred1t and tenure pohcies which equated land improvements w1th cleanng and land speculat1on (espec1ally 1n a h1ghly lnflationary economy) Roads opened the Amazon to settlement and land to deforestat10n The Bras1ha Belem h1ghway was completed m 1960 The then m1htary government made the Amazon the focus of growth m 1964 BR364 between Porto Velho m Rondoma and Cu1aba m the south was completed 1n 1965 and 1mproved m 1969 The government dec1ded to bUIId the Trans Amazon h1ghway m 1970 and vanous sect1ons were completed between 1972 and 1976 Deforestat1on m the Amazon has been the heav1est along these roads O ver 1 O m1lhon ha of forest were converted to pasture from 1960 to 1990 The Agency for the Development of Amazonia (SUDAM) was estabhshed m 1966 largely to facli1tate pnvate mvestments m the Amazon through tax 1ncent1ves w1th most support granted to large ranches and corporat1ons Ranches now cover 8 4 m1lhon ha average 24 000 ha each but employ only one cowboy per every 300 ha and were prof1table only w1th full tax advantages and at considerable soc1al costs of O 5 ton of ramforest per quarter pound of hamburger Beef produced however was not for export to the US for 1ts fast food mdustry as Hecht proposed and as was the case 1n Central Amen ca Although beef suff1c1ent for the same quarter pound hamburger cost US $0 26 to produce 1n the Amazon w1th $ O 22 commg from subs1d1es the Amazon has barely produced enough beef for 1ts own needs Related pohc1es that contnbuted to deforestat1on were exempt1ons of agncultural mcomes from taxat1on rules determmmg secunty of land cla1ms wh1ch encouraged cleanng progress1ve land taxes encourag1ng convers1on to pasture cred1t schemes subs1d1s1ng corporate hvestock ranches and tax breaks for wood products mdustnes The rural poor were pushed to the front1er are as m search of land 4 5% of Braz1l s land owners hold 81% of the country s farmland wh1le 70% of rural households are landless In the penad 1963 73 Braz1l sh1fted from mward onented 1mport subst1tut10n pohc1es to modifled outward onented export substitUtiOn charactenzed by area expansion domestlcally produced mputs mechan¡zatlon through cred1t mcreased output and land concentrat1on and coffee eradicatlon The penad 1971 80 then saw a h1gh expans10n of agncultural exports and wealth concentrat1on a1ded by h1gh 1nflat1on combmed w1th low f1xed mterest rates for a few borrowers Mechamzat1on of soybean wheat and sugar cane product1on had the effect of dnvmg the rural poor out of Braz1l s populous and developed center south reg1on In that area soybean became a ma¡or crop and labor 1ntens1ve small scale agnculture was replaced by energy and machme mtens1ve cult1vat1on 26 Braz1l s Nat1onallnst1tute for Colomzat1on and Agranan Reform (INCRA) mst1tuted a Program for Nat1onal lntegrat1on (PIN) m 1970 to ass1st the rural poor H1gh rates of m1grat1on to Rondon1a mcludmg spontaneous m1grat1on and land 1nvas1ons resultad m f1ve m1lhon ha (21% of the state s land area) g1ven to settlers by 1977 Forest cleanng reached 17% of the state by 1987 and farmers Wlth rural cred1t cleared 25% more forest than those w1thout Problems such as lack of all weather access roads unsUitable seed early rams lack of plantmg matenals for perenmal crops and 1nsecure t1tles as well as the 011 cns1s m 1973 however led the government to sh1ft from supportmg small farmers v1a colomzat1on schemes to (aga1n) a1dmg large operators m 1974 Other factors contnbut1ng to Amazoman deforestat1on have mcluded d1splacement and resettlement of farmers from earher settled areas m the north as so1ls were depleted or as demands for land by new speculators mcreased loggmg and wood process1ng and mmmg and use of t1mber for fuel for smeltmg Deforestat1on smce 1985 Sorne of the ma1or mcent1ves prov1ded to large corporat1ons and cattle ranches were ellmmated m 1985 as Braz1l suffered from recess1on and hypennflat10n Although cause and effect relat1onsh1ps are d1ff1cult to estabhsh recent analys1s of remote sensmg data shows that rates of Amazoman deforestat1on have been decreasmg from 8 O m1lhon ha m 1987 to 1 8 m1lhon ha m 1989 1 4 m1lhon ha 1n 1990 and to 1 1 m1lhon ha 1n 1991 On the other hand two th1rds of deforestat10n due to ranchmg has occurred Wlthout SUDAM mcent1ves and explanatory factors appear to be a) that new demands for (cleared) front1er lands by urban based land speculators have appeared as front1er government has strengthened to the extent that 1t can protect property nghts (as opposed to ranchers havmg to fmance pnvate arm1es to mamtam the1r land cla1ms) and b) that cattle ranch1ng 1s econom1cally v1able under current land pnces The vanous factors contnbutmg to or tendmg to d1m1msh deforestat1on m the Amazon Basm over t1me are shown m F1gure 1 Opt1m1st1c v1ewpo1nts m the search for alternat1ves Opt1m1st1C fmd1ngs have emerged along w1th the observad decreasmg rates of deforestat10n F1rst forests may be more res1hant and m ay rebound from d1sturbances more qUickly and completely than prev1ously thought Second secondary forests globaliy are bemg formed at about mne m1lhon ha per year and now account for 40% of total forest area and th1rd much of currently observad deforestat1on 1s of secondary rather than pnmary growth As a result of such res1hance severa! researchers have proposed 1mproved management of secondary forests and encouragement of natural forest regenerat1on Other researchers have descnbed the d1vers1ty of env1ronments Wlthm the Amazon Basm and suggested land class1f1cat1on and zomng as startmg pomts for 1mproved land use and management Although there IS ev1dence that so1l fert11ity under cattle pasturas 1n the Amazon cannot be ma1ntamed research has also suggested that susta1nab11ity 1s poss1ble w1th 1mproved pasture spec1es Contrary to earher v1ews of ecolog1sts research also suggested that cont1nuous croppmg IS poss1ble w1th fert11izer use Agroforestry has al so been suggested as an appropnate land use and even steam sawm1lls and stewardsh1p rather than ownersh1p were suggested as appropnate and needed for the Amazon 27 "' CXI FIGURE 1 POL1CIES & OTHER FACTORS LEROING TO GRERTER RNO/OR LESSER OEFORESTRTION OVER TIME IN THE AMAZON 1960 1970 1990 1990 FAST REGENERATION INCRERSEO USE OF SECONDRRY ----- OF SECONORRY FORESTS FORESTS CURRENCY r DECLINE IN r 1 NOEXATI ON OF STRBILIZRTION RORO BUILDING THE ECONOMY f- & CONTROL OF & MRINTENRNCE INFLATION 1 ISOLRTION LOW TECHNICAL OIL CRISIS & 1985 7 GLOBAL FROH PROFITRBILTY FRILURES OF HYPER 1 NFLATI ON ELIMINATION PRESSURES ----- OEVELOPEO OF CRTTLE PUBLIC '-- LERO TO OF HRJOR ¡-. RGRINST RRERS OF RRNCHING COLON 1 ZRTI ON OECREASEO GOV INCENTIVES OEFORESTRTION BRRZIL SCHEHES SUBSIDIES FOR RRNCHES 1 l l l l l f- .... ¡..... People also salve the1r own problems W1th deforestat10n of the1r parcels settlers m the Bohv1an Amazon were observed to eventually reach a cnses stage due to s01l nutnent deplet1on weeds decreasmg y1elds and mcreasmg costs but then (some) successfully responded by turmng to mechamzat1on cattle or perenmal crops thereby escapmg to a consohdated stage w1th markedly h1gher mcomes THE RESEARCH SITES Acre and Rondoma The state of Acre covers 153 000 km 2 The area of Rondon1a 1s 239 000 km 2 Acre 1s dra1ned by the R1o Purus and R1o Jurua whlle Rondoma IS dramed by the R1o Made1ra Chmate 1s warm and hum1d tropical w1th a ramy season spannmg from July to October m Acre and from June to August 1n Rondoma Mean annual ramfall IS approx1mately 2000 mm and mean temperaturas are 22 26 C So1ls m Acre are Ox1sols (Latosolo Amarelo Vermelho D1strof1cos) and m Rondoma are Oxlsols (Latlsolos) Alf1sols (Terra Roxa) and Ult1sols (Podzohcos) The populat10n of Acre 1s 426 000 and that of Rondoma 1s 1 350 000 Reported cattle herd s1ze 1n 1990 was 40 000 m Acre and 1 700 000 1n Rondoma Reported crop product10n m Acre for 1993 was 32 000 t of nce 36 000 t of ma1ze 14 000 t of beans 20 000 t of cassava and 12 000 t of oranges Rondoma produced a reported 275 000 t of nce 390 000 t of ma1ze 93 000 t of beans 650 000 t of cassava 166 000 t of coffee and 26 000 t of banana Forest products for Acre were rubber ( 12 000 t) Braz1l nuts (Bertholletla excelsa locally castanha 18 000 t) and wood (300 000 m 3) Pedro Pe1xoto Acre and Theobroma Rondoma The Pedro Pe1xoto colon1zat10n s1te covers 370 000 ha d1v1ded mto 3700 parcels d1stnbuted to 3200 fam1hes Lots are located at d1stances of 50 to 1 00 km from the state cap1tal of R1o Branco Theobroma covers 300 000 ha d1v1ded mto 3000 parcels (reportedly) d1stnbuted to 3000 fam1hes The project area 1s located some 350 km from the state capital of Porto Velho (Marcus VmiCIUS personal commun1cat1on) METHODS Farmer settler Jnterv1ews A goal of the ASB pro]ect 1s to charactenze each s1te m a way that data can be synthes1zed and compared across s1tes Coordmated by ICRAF researchers met m Porto Velho 1n late 1 993 to develop and pre test a survey quest1onna1re that would be apphcable across s1tes Th1s prototype quest1onna1re was rev1sed for use and apphcat1on 1n Acre and Rondoma m m1d 1994 by researchers from CIA T and IFPRI 1n arder to ensure appropnateness to local cond1t1ons to ensure ease of use m the f1eld and to fac1htate data codmg Booklets to code data from each quest10nna1re were developed Open ended quest1ons were mcluded wh1ch were coded after the range of responses were rev1ewed Ouest1onnare and the data codmg booklets were fmahzed after da1ly mod1f1cat10ns and group d!scuss1ons m an m1t1al week of f1eldwork F1eldwork took place m late August and early September 1994 as settlers were burmng the1r f¡elds Partlcipants representad CIAT IFPRI EMBRAPA (CPAF Acre and CPAF Rondoma) w1th mterv1ewers from the Grupo de PesqUisa e Extensao em Sistemas Agrofloresta1s do Acre (PESACRE) a non government organ1zat10n contracted by EMBRAPA A team of some 15 1nterv1ewers was assembled and mterv1ewers worked mdlvldually (nat1ve speakers of Portuguesa) or m pa1rs (where the add1t1onal person was not a Portuguesa 29 nat1ve speaker) Pedro Pe1xoto and Theobroma were d1v1ded mto areas wh1ch are relat1vely access1ble or maccess1ble dunng the ramy season The group selected and worked m d1fferent areas 1n each s1te each day and farmers were selected at random m the f1eld by d1stnbutmg mterv1ewers at considerable d1stances from one another and then 1f appropnate by us1ng an every th1rd parcel approach E1ghty one farmers m Pedro Pe1xoto and 7 4 farmers m Theobroma were 1nterv1ewed w1th e1ther or both male and female heads of household mcluded Completad quest10nna1res were rev1ewed and data passed to codebooks each evemng There were considerable d1ff1cult1es m obtammg complete mterv1ews due to mexpenence of sorne of the contracted 1nterv1ewers The group moved to Theobroma after farmer mterv1ews m Pedro Pe1xoto and lastly worked together m Porto Velho to make sure that data cod1ng was as complete and read1ble as poss1ble Cop1es of completad quest1onna1res were left wtth EMBRAPA for the1r data analys1s and pubhcat1on CIAT al so retamed cop1es of the mterv1ews and has rev1ewed and cross checked data rev1sed the codmg system as needed set up data arch1ves and enterad and tabulated data Th1s paper presents and d1scusses descnpt1ve fmdmgs (s1mple frequenc1es and means) from th1s tablulated data GIS and Remate Sens1ng The CIA T tea m assembled ava1lable maps and secondary data from vanous sources and d1g1t1zed data m arder to present s1mple overlays of for example proJect roads and parcels by access 1n the wet season and roads and parcels by sotl type Satelhte 1mages of the s1te are bemg sought m arder to analyse rates of deforestat1on over t1me and as funct10ns of d1stances from roads (both pnmary and secondary) and t1me of road construct1on An overlay of the parcellary or cadastral maps on the 1mages 1s allowmg for analys1s of deforestat1on by parcel access parcel tenure and other vanables RESUL TS Farmer Survey Heads of household m earher settled Theobroma were born m the Northeast (4 7%) South (38%) and Center West (12%) ra1sed 1n the Southeast (48%) Northeast (22%) and South (20%) and pnor to amval m Theobroma hved m the Southeast (40%) South (31 %) and Northeast (16%) The data mdtcates h1gh geograph1cal mobi11ty a) away from the Northeast b) to and then from the Southeast e) and from the South Few Theobroma settlers were born ra1sed or hved prev1ously m the North Pedro Pe1xoto settlers had a d1fferent pattern of movement and settlement Almost half were born (45%) ra1sed (41 %) and last hved (50%) 1n the North Approx1mately a th1rd were born and ra1sed m Acre 1tself and a th1rd although not born or ra1sed there hved m Rondoma p~ror to settlement m Pedro Pe1xoto Other ma1n pomts of ongm for Pedro Pe1xoto settlers were the South and the Northeast (Table 1) Settlers left prev1ous places of res1dence largely m search of land (52% m Pedro Pe1xoto and 58% 1n Theobroma) and a better hfe (Table 2) 30 Table 1 Ongms of heads of household (% of respondents) Pedro Pe1xoto Acre (n = 81) & Theobroma Rondoma (n = 74) Braz1l 1994 Pedro Pe1xoto Theobroma Born Ra1sed Last Born Ra1sed Last North (Amazoma) 45 41 50 3 4 1 Rondon1a o 3 34 3 4 1 Acre 35 35 10 o o o Other 10 3 6 o o o Northeast 25 12 14 47 22 16 Southeast o 20 12 o 48 40 South 27 22 20 38 20 31 Center West 3 4 4 12 5 12 Table 2 Reported reasons (% of respondents) for leavmg prev1ous res1dence Pedro Pe1xoto (n = 60) and Theobroma (n = 67) Reasons for leavmg Pedro Pe1xoto Theobroma No land 52 58 Search for better hfe 18 25 Work on rubber plantat1on 10 Others 20 17 Although Theobroma was 1n1t1ated as a colomzat1on pro¡ect earher than Pedro Pe1xoto settlers had hved m the former a mean s1x years and m the latter a mean e1ght years poss1bly 1nd1catmg e1ther arnval of a second generat1on of settlers m Theobroma or mclus1on m Pedro Pe1xoto of early spontaneous settlers who remamed 1n the pro¡ect area For Pedro Pe1xoto almost a th1rd arnved pnor to 1975 and two th1rds arnved m the 1980s For Theobroma arnval was mamly 1n the 1970s (36%) and 1980s (58% Table 3) Mean s1zes of respondents parcels were 88 ha m Pedro Pe1xoto and 76 ha m Theobroma For Pedro Pe1xoto and for 1993 94 a mean 69% was st1ll forested 20% was 1n pasture 6% 1n fallow and 4 A> m annual crops Theobroma was charactenzed by more cleared land (46%) more pasture (26%) and more perenn1al crops (5% Table 4) Because mterv1ews were conducted after f1eld burmng for 1994 95 changes m land use from 1993 94 to 1994 95 could be calculated add1t1onal forest cleanng m late 1994 meant that forested port1ons of the settlers parcels decreased from 69% to 66% m Pedro Pe1xoto and from 54% to 50% m Theobroma Overall sorne 60% of the land 1n the two colon1es remams forested w1th more 31 than half of the cleared area converted to pasture Only 7% of settlers lands were m fallow (Table 4) Table 3 Year of arnval (% of respondents) m Pedro Pe1xoto (n = 81) and Theobroma (n = 74) Year Pedro Pe1xoto Theobroma 1990 94 3 3 1985 89 40 24 1980 84 23 34 1975 79 5 21 1970 74 13 15 < 1970 16 3 Total 100 100 Mean years at s1te 8 6 Table 4 Land use (mean areas) Pedro Pe1xoto Acre (n = 81) & Theobroma Rondoma (n = 74) 1993/94 & 1994/95 Pedro Pe1xoto Theobroma 93/94 94/95 D1f 93/94 94/95 D1f ha % % % ha % % % Total Forest 61 69 66 5 4 54 50 7 61 1 Cleared 27 31 34 + 11 3 46 50 +8 39 5 Pasture 17 20 25 +30 2 26 29 +10 23 Fallow 5 6 2 60 o 8 4 50 7 Annual crops 4 4 7 +50 6 7 9 +40 6 Perenmals 1 1 o 5 5 8 +50 3 4 TOTAL 88 100 100 7 100 100 100 6 Settlers cleared a mean of shghtly more than one ha of pnmary forest per year at both 32 s1tes Two th1rds of Pedro Pe1xoto farmers cleared pnmary forest every other year wh1le most Theobroma farmers cleared every other year (46%} or every three years (25% Table 51 Farmers m Pedro Pe1xoto needed 23 days/ha to clear pnmary forest and 16 days/ha to clear fallowed land Theobroma farmers spent less t1me cleanng land but needed the same amount of labor to clear forest and fallow (14 days/ha Table 6} Reasons for d1fferences between s1tes were not apparent Farmers used chamsaws axes and machetes to clear pnmary forest and axes and machetes to clear secondary forest Labor was mamly fam1ly followed by h1red labor often for chamsaw operat10n (Table 6} Table 5 Respondents reported frequency of forest cleanng (% of respondents} & mean area of forest cleared per year (ha} Pedro Pe1xoto and Theobroma Pedro Pe1xoto Theobroma Does not clear forest 4 2 Frecuency of cleanng 9 13 Every year 66 46 Every two years 17 25 Every three years 4 14 > 3 years Total 100 100 Mean area cleared/year (ha} 1 2 1 1 Table 6 Respondents reported labor use (type and days/hal for cuttmg & slashmg forest and fallow pnor to burmng (days/ha) Pedro Pe1xoto and Theobroma Pedro Pe1xoto Theobroma Forest Fallow Forest Fallow Labor use (days/hal 22 6 16 4 13 5 13 6 Labor type (% respondents} Fam1ly 72 81 92 85 H1red 30 12 56 25 Exchange 8 2 10 4 Use chamsaw 90 1 89 6 33 Burmng of slash followed cuttmg and a penod of drymg Most farmers thought that the funct10n of burn1ng was to make space for the crops and only a few at each s1te ment1oned that ash 1mproved so1l fert11ity or that f1re decreased the mc1dence of weeds and other pests (Table 7) Th1s result h1ghhghts d1fferences between such settlers and trad1t1onal slash and burn agnculturahsts who umversally perce1ve burmng m terms of nutnent management and pest control Farmers (92% 1n Pedro Pe1xoto and 70% m Theobroma) planted nce m the f1rst year of cult1vat1on of what was pnmary forest and cult1vated ma1ze cassava (m Pedro Pe1xoto) and pasture 1n the second year R1ce the most 1mportant crop for both consumpt1on and sales of surpluses was not grown m the second or subsequent years of plot use (Table 8) Table 7 Respondents reponed reasons (% of respondents) for burmng slash Pedro Pe1xoto and Theobroma Clear land/make space Ash/f1re 1mproves so1l fert1hty Remove weeds/pests Table 8 Respondents reportad land use Pedro Pe1xoto 81 11 11 f1rst & second years cleanng Pedro Pe1xoto (n = 70) & Theobroma (n = 67) Pedro Pe1xoto 1st 2nd R1ce/nce + ma1ze( beans) 92 o Ma1ze o 42 Cassava o 26 Pasture 1 17 Fallow o 8 Other 7 7 Total 100 100 Theobroma 79 10 6 (% respondents) after Theobroma 1st 2nd 70 o o 44 o o 13 25 o 8 17 23 100 100 Farmers reported the1r 1993 and the1r normal low and h1gh y1elds for nce ma1ze and beans m terms of rat1os of y1eld to seed sown Y1elds m 1993 were somewhat below average w1th normal y1elds at the two s1tes bemg about 70 75 1 for nce 95 1 for ma1ze and 27 1 for beans (Table 9) Mean reported y1elds at both s1tes d1d not d1ffer Farmers reportad sowmg from seven to 60 kg of nce seed per ha at both s1tes w1th a mean of about 21 kg/ha Farmers normal nce y1elds would then be 1n the 1 5 to 1 6 t/ha range wh1ch 34 would appear to be a reasonable est1mate g1ven local cond1t1ons and farmers management pract1ces Most farmers at the two s1tes reported problems (defmed as factors wh1ch led to decreased y1elds) w1th nce and beans Stmk bug stemborer and b1rds 1n Pedro Pe1xoto and b1rds and storage weev1ls m Theobroma were problems of nce (Table 10) Web bhght (Thanathephorus cucumens) and beetles (D~abrottca spp) were the ma¡or problems of beans at both s1tes Few farmers reported problems w1th ma1ze although problems ment1oned mcluded b1rds w1ld ammals and sp1ttle bug (Deots flavoptcta) Table 9 Respondents reportad mean normal low and hlgh crop y1elds (ratiO of product1on to seed sown) Pedro Pe1xoto and Theobroma Pedro Pe1xoto Theobroma Crop 1993 Normal Low H1gh 1993 Normal Low H1gh R1ce 63 70 34 103 69 75 42 96 Ma1ze 88 94 56 118 76 96 59 154 Beans 19 28 4 52 10 26 10 41 Farmers cult1vated lands cleared from pnmary forest for a mean penad of two (Pedro Pe1xoto) to 2 5 years (Theobroma) S1xty percent 1n Pedro Pe1xoto cult1vated such plots for two years wh1le Theobroma farmers used the1r newly cleared lands for from ene to more than three years 1n somewhat equal proport1ons (Table 11) Agam reasons for d1fferences between s1tes were not apparent and bear further mvest1gat1on Farmers reportad that d1scont1nuat10n of annual croppmg on lands cleared from forest was dueto the not mutually exclus1ve reasons of lower product1v1ty weeds espec1ally lmperata sp (locally sape) and msects and d1seases (Table 12) After food crops two th1rds of Pedro Pe1xoto farmers and nearly half of Theobroma farmers converted the1r lands to pastura About a th1rd m both areas left sorne land m fallow (although much of such land may also be used as ummproved pastura) Theobroma (20%) but not Pedro Pe1xoto farmers also converted sorne land from annual to perenn1al crop use (Table 13) Farmers at both s1tes normally left any f1elds wh1ch they fallowed for a mean 2 5 years although they thought that 3 O 3 5 years of fallow would be 1deal R1ce followed by ma1ze and beans were the ma1n crops plantad 1n the re opened fallows (Table 14) 35 Table 1 O Reported crop problems (% of respondentsl Pedro Pe1xoto (n = 71 l & Theobroma (n = 72) Problem Has problem St1nk bug B1rds Stemborer D1seases Storage weev1ls M1scellaneous Has problem B1rds W1ld ammals Sp1ttle bug Storage weev1ls M1scellaneous Has Problem Web bhght Chrysomehdes M1scellaneous Local Sc1ent1flc names RICE Perceve¡o T1braca spp Oebalus spp Pasanm/Grauna Pamcu/a branca D1atrea spp Symptoms referred to as quema Gorgu/ho MAIZE Pasan m C1garmha Deo1s flavop1cta Gorgu/ho BEANS Mela Thanathephorus cucumer1s Vaqumha OJabrot1ca spp Pedro Pe1xoto 78 45 14 17 7 6 31 41 5 o 5 14 23 78 58 30 29 Theobroma 65 5 34 5 2 15 28 28 8 5 2 5 18 96 58 40 28 T able 11 Respondents reported normal years of cult1vat1on of lands cleared from forest (% of respondents) Pedro Pe1xoto & Theobroma Years cult1vated Pedro Pe1xoto Theobroma 1 13 26 2 60 24 3 16 31 >3 11 19 Total 100 100 Mean years 2 1 25 36 Table 1 2 Respondents reportad reasons for d1scontmuat1on of annual croppmg m what were newly cleared lands (% of respondentsl Pedro Pe1xoto & Theobroma Reason Pedro Pe1xoto Theobroma Lower product1v1ty 36 68 Weeds 51 34 Weeds sapé 43 5 01seases & pests 25 12 Other 31 40 Table 13 Respondents reportad general or normal use of lands after annual croppmg (% of respondents) Pedro Pe1xoto (n = 70) & Theobroma (n = 67) Land use Pedro Pe1xoto Theobroma Pastura 64 44 Fallow 36 36 Perenmals o 20 Total 100 100 Table 14 Respondents reportad use of fallow (years & % of respondents) Pedro Pe1xoto and Theobroma Pedro Pe1xoto Theobroma Years of last fallow 24 27 Years 1deally fallowed 34 3 1 Years normally fallowed 24 24 Crop plantad after fallow ( %) R1ce 62 72 Ma1ze 23 13 Beans 12 8 Cassava 3 o Pastura o 2 Other o 5 37 For perenmal crops 35% of Pedro Pe1xoto farmers reportad problems of low pnces or poor markets 29% reportad no problems and 40% had not grown perenmals In Theobroma only 20% had not grown perenmals 43% reportad product1on problems (1 e msects and d1seasesl and 45% reportad havmg had no problems (Table 15) Most settlers (91 % m Pedro Pe1xoto and 81 % m Theobroma) had cattle Herd s1ze was a mean 18 head (w1th 6 g1vmg mil k) m Pedro Pe1xoto and 26 (4 g1vmg m1lkl m Theobroma (Table 16) M1fk product1on ranged from a mean two to three hters per head m the dry season and tour to f1ve hters 1n the wet season Settlers reportad stockmg 2 2 head per ha m Pedro Pe1xoto and 1 6 head per ha m Theobroma w1th roughly equal numbers of respondents at both s1tes report1ng stockmg al less than one b) one to two and e) more than two head per ha (Table 17) Table 15 Respondents reportad problems (% of respondents) w1th perenmal crops Pedro Pe1xoto (n = 1 7) and Theobroma (n = 49) Problem Product1on problems Low pnces/poor markets Lack of transport Too much labor reqUired F~re dunng burmng season No problem Has not grown perenmals Pedro Pe1xoto 6 35 18 3 12 29 40 Theobroma 43 10 o 1 2 45 20 Table 16 Respondents reportad hvestock holdmgs (head) Pedro Pe1xoto and Theobroma Pedro Pe1xoto Theobroma Percent of settlers w1th cattle 91 81 Number of cattle Max1mum number 1n last year * 18 26 Number sold m past year 4 2 Current number 23 30 Mil k Mean number g1vmg m1lk (head) 6 4 Mean hters/head/day dry season 2 3 Mean hters/head/day wet season 4 5 D1fference 1n max1mum number last year and current number reflects cattle born deaths and purchases 38 Table 17 Current number of ammals per ha(% respondents) Pedro Pe1xoto and Theobroma Ammals par ha < 1 1 2 > 2 Pedro Pa1xoto 33 38 29 Theobroma 37 37 26 The predommant pasture spec1as usad at both s1tes were 8rachtana bnzanta 8 decumbens and 8 humtdtcola Sorne Pedro Pe1xoto farmers also had Puerarta phaseolotdes or m1xes of P phaseolotdes and brach1ana Theobroma settlers also had pasturas of Pamcum maxtmum (Table 18) In terms of pastura management a mean 70 75% of farmers burned f1elds yearly and 70% of farmers at both s1tes rotated camle to d1fferent pasturas at a mean of every 2 O 2 5 months (Table 19) Table 18 Respondents reportad types and are as (% & ha) of pastura Pedro Pe1xoto & Theobroma Pedro Pe1xoto Theobroma % of mean % % of mean % responde are a of responde are a of nts (ha) are a nts (ha) are a 8rachtana brtzanta 80 12 57 78 15 51 8rachtana decumbens 60 7 24 36 9 13 8rachtana humtdtcola 11 3 2 32 5 5 Puerarta phaseolotdes 7 4 1 Pamcum maxtmum 12 7 3 P phaseolotdes +8 7 22 10 decumbens Other m1xtures 4 7 2 8 20 7 Nat1ve pastura 7 8 3 Other 12 2 1 37 11 21 39 Table 19 Pastura management pract1ces (% of respondents and frecuency over t1me) Pedro Pe1xoto & Theobroma Use f1re for regenerat10n (%) Frequency of f1re use (years) Rotate cattle to d1fferent pasturas (% 1 Frequency of rotat10n (months) Pedro Pe1xoto 75 1 3 70 26 Theobroma 71 1 1 70 2 1 Table 20 Respondents (% l use of forest products Pedro Pe1xoto and Theobroma Forest Product Braz1l nut Wood Huntmg Palm1to F1sh Rubber Med1cmal plants Acat (another palm) Jatoba (a tree legume) Other Pedro Pe1xoto Theobroma 90 36 33 50 18 29 5 17 6 13 11 o 5 6 o 6 o 6 11 4 Settlers expl01ted the1r forest lands for Braz1l nut (although substant1ally more m Pedro Pe1xoto) wood huntmg palm hearts (more m Theobroma) f1sh and rubber (Pedro Pe1xoto Table 20) Although an attempt was made quant1t1es of these products were very d1ff1cult to ellc1t or calculate Farmers cash mcome sources (Tabla 21) m Pedro Pe1xto were from sales of labor or pens1ons (63%) and sales of nce (50% solda mean 2 3 t/year) ma1ze (47% 1 8 tonsl beans (41% 1 3 tons) Braz1l nut (44% 1 O tons) and cattle (26%) More farmers (57%) sold more nce (3 1 tons) 1n Theobroma m add1t10n to havmg mcomes from labor and pens1ons (53%) coffee (36% 3 3 tons) m1lk (30%) cacao (25% 1 6 tons) and cattle (22%1 40 Table 21 Respondents (%) sources of cash 1ncome and approx1mate quant1t1es sold (tons/year) Pedro Pe1xoto (n = 68) • and Theobroma (n = 631 • Source Pedro Quant1ty Theobroma Ouant1ty Pe1xoto Crops R1ce 50 23 57 3 1 Ma1ze 47 1 8 16 27 Beans 41 1 3 9 02 Cassava 4 1 5 o Fannha 7 4 1 o Cotton 1 06 9 04 Coffee 3 1 3 36 33 Cacao o 25 1 6 Cattle (head) 26 23 22 5 M1lk 7 30 Poultry/p1gs 9 o Rubber 7 1 6 o 137 Castana 44 1 o 6 Wood o 9 Other 9 8 Labor and pens1ons 63 53 Agnculture male 11 26 Agnculture female 3 o Non agr males 15 8 Non agr female 16 1 Pens1ons 19 16 Fewer respondents prov1ded quant1t1es of each product Farmers may be earmng more from apprec1at1on of land values 93% of Pedro Pe1xoto and 97% of Theobroma settlers perce1ved the1r land values as hav~ng nsen (values were d1scussed 1n terms of eqwvalent numbers of cattle) at annual rates of 74% 1n the former and 1 57% 1n the latter Farmers reportad total 1ncreases s1nce occupy1ng the1r parcels 1n value of about 800% 1n Pedro Pe1xoto and 950% 1n Theobroma w1th reasons for 1ncreases attnbuted to add1t1on of pastura or cleared areas fenc~ng ponds or other water sources corrals houses perenmal crops and 1mproved access (Table 22) Negat1ve factors assoc1ated w1th the colomes were poor roads or access malana and lack of health posts schools and potable water (Table 23) 41 Table 22 Respondents evaluat1on of and reasons for (% of respondentsl changmg land values Pedro Pe1xoto (n = 691 and Theobroma (n = 701 Pedro Pe1xoto Theobroma Reported mcreased value (%) 93 97 Total mean mcrement m cattle (%) 778 952 Mean annual 1ncrement m cattle (%) 74 157 Reasons for mcrease More pasture 60 50 Fencmg 56 36 Pond/water 30 13 More cleared area 12 26 Corral 12 16 House 26 16 Access/roads 25 27 Perenmal crops 12 35 School 10 5 T1mber 8 1 T1tle 7 3 Good SOIIS 4 5 Table 23 Respondents (% of respondents) reported negat1ve factors assoc1ated w1th llvmg 1n Pedro Pe1xoto or Theobroma Pedro Pe1xoto Theobroma Peor roads/access 51 27 Malana 25 35 Lack of transport 24 9 health post 22 30 schools 14 22 potable water 10 4 electnc1ty 4 8 cred1t 10 1 Other 23 21 Non e 4 11 42 GIS and Remota Sens1ng Results Overlaymg the Theobroma parcel map on a Braz1han so1ls map shows about 1400 parcels (37% of parcels) and 114 thousand ha (45% of area) of more favorable Alf1sols (Podzohco Vermelho Amarelo Eutrof1cos) 2200 parcels (57%) and 124 thousand ha (49%) of med1um quahty Ox1sols and Ult1sols and 250 parcels (6%) and 14 thousand ha (6%) of poor D1stropepts (l1t1cos d1strof1cos Tabla 24 Th1s mapped data w1ll perm1t a follow up study to determme 1f there are d1fferences m land use and m product1v1ty relat1ve to so1l quahty The same mappmg of cadastral data and so1ls m Pedro Pe1xoto showed that solls throughout the s1te are relat1vely s1m1lar m terms of agncultural potent1al (Bell and Fu¡1saka forthcommg) Parcels m Pedro Pe1xoto w1th t1tles and w1th t1tles pendmg were mapped for Pedro Pe1xoto (Table 25) and wlll be overlayed on satelhte 1mages m order to determme the role of tenure on rates of deforestat1on Fmally a satelhte 1mage from 1991 wh1ch mcludes Pedro Pe1xoto f1ve haciendas and other land uses and covers 35 7 thousand ha of wh1ch 19% of the total 1s deforestad 20% of the colomsts 270 thousand ha 29% of the haCiendas 22 thousand ha and 10% of the 64 thousand ha of other were deforestad (Tabla 26) The off1c1ally reportad area of the Pedro Pe1xoto pro¡ect of 370 thousand ha appears to mclude what are now haciendas and other land uses bes1des settlers parcels The currently farmer reportad amounts of forest cleanng appear to agree w1th the 1mage analys1s g1ven the1r reportad rates and the amounts shown m the 1991 1mage and g1ven that analys1s was unable to d1st1ngUish between relat1vely estabhshed secondary growth and pnmary forest As ment10ned add1t1onal analys1s 1s attemptmg to analyse deforestat1on as a funct1on of d1stance from roads and add1t1onal 1mages are bemg obtamed to analyse ratas m Pedro Pe1xoto and Theobroma over t1me Tabla 24 Number of parcels and area (ha) by s01l quahty Theobroma Parcels Area (ha) Number % ha % So1l Type Good 1392 37 113 524 45 Alf1sols Med1um 2172 57 123 689 49 0XISOIS UltiSOIS Poor 241 6 14 120 6 D1stropepts 43 Table 25 Number and area (ha) of lots w1th t1tles and w1th pendmg t1tles Pedro Pe1xoto W1th t1tles T1tle pendmg Parcels Number 1103 379 % 74 26 Area (ha) ha 75 253 26 021 Table 26 Satelhte 1mage analys1s of Pedro Pe1xoto ( 1991) Forest Deforestad Ha % Ha % Total (ha x 000) 290 3 81 66 4 19 Colomsts lots 217 5 80 53 5 20 Haciendas (5) 15 5 71 64 29 Other 57 4 90 65 10 DISCUSSION % 74 26 Total Ha 356 7 271 o 21 8 63 8 Deforestat1on at the farm level appears to contmue at a relat1vely steady pace averagmg somewhat more than one ha per year per fam1ly m the two colomes Settlers st1ll had more than half of the1r lands m forest and for the most part can and probably w1ll contmue to slash burn and cult1vate more pnmary forest as the1r currently most (econom1cally) v1able opt1on Two mam factors dnvmg land cleanng at the farm level were the need to produce food crops and mcent1ves to convert land mto pastura In terms of food product1on farmers consumad and sold nce and to a lesser extent beans ma1ze and cassava (or cassava meal) R1ce cult1vat1on may dnve sorne deforestat10n m that although farmers usually planted a cleared f1eld for two or three years they could not for techmcal reasons sow nce other than m the f1rst year after cleanng As such a research pnonty of the ASB proJect may be to determme 1f and under wh1ch management alternat1ves could nce product1on be made more sustamable (acknowledgmg the caveat that any 1mprovement m product1v1ty may lead to greater econom1c attract1veness of the respective enterpnse and thereby 1nv1te more deforestat1on or other forms of resource over explo1tat1onl Farmers were clearly mot1vated to convert cleared lands mto pastura because of real or at least perce1ved resultmg mcreases m land values Farmers not only mamtamed cattle as 44 standmg bank accounts and obta1ned cash from sales of ammals and m1lk but bUIIt savmgs by mvest1ng t1me and resources m fencmg corrals ponds and other ranchmg necess1t1es As observed throughout the two s1tes local ranchers and urban based specultors have purchased contmuous blocks of colomsts parcels to form new ranches or to expand the s1ze of adjacent ranches and payments were reportely much h1gher for cleared vs forested port10ns of parcels Farmers pasture management pract1ces cons1sted of mtroducmg mamly Brachlana spp annual pasture burmng and rotat1on of anmmals to d1fferent pastures In sp1te of generally low stockmg rates there were substant1al areas of poor and degraded pastures and pasture lands at both s1tes Although construct1on of new cheese process1ng plants near Theobroma may result m mtens1f1cat10n at that s1te extens1ve and low mput cattle and pasture management appear to be the current norm for both s1tes Whether or not 1mproved pasture technolog1es are poss1ble and would be appropnate to farmers current cond1t1ons IS another researchable ISSUe Convers1on of cleared forest land to pasture has meant that at least m the colomes stud1ed that there was relat1vely httle land placed 1n fallow w1th result1ng secondary regrowth and forest regenerat1on An 1mphcat10n 1s that research on 1mproved fallows may e1ther result 1n makmg fallows more attract1ve to farmers (as 1s hoped) because of more rap1d regenerat1on and better ma1natenance of soll orgamc matter or farmers may be unmterested m 1mproved fallows because convers10n to pasture 1s by far the preferred land use alter cult1vat1on lnteractmg factors leadmg to deforestat10n at the on farm level m Pedro Pe1xoto and Theobroma are presentad m a quahtat1ve model (F1gure 2) Theobroma farmers pract1ced agroforestry 1n the sense that they had S1gmf1cant areas of perenmal crops mamly coffee and cacao Pedro Pe1xoto settlers have had less favorable expenences regardmg perenn1als c1t1ng poor pnces and/or markets as major constramts Sorne Pedro Pe1xoto farmers had been encouraged (prov1ded w1th cred1t) to plant urucu (81xa ore/lana used to produce red dye) When the trees started to produce pnces for the product fell to the po1nt that farmers could not afford to harvest and lost the1r mvestements Currently settlers m Theobroma were bemg prov1ded cred1t to grow acerola (Malp1gh1a pumclfofla) Although they were prom1sed a future market for all that they produce schemes to Introduce or encourage perenmal crops or agroforestry remam nsky A resource that was largely not ava1lable m the systems exam1ned 1s md1genous techmcal knowledge of the type usually assoc1ated w1th sh1ftmg cult1vators Farmers were asked about the1r so1l and land class1f1catlon and correspondmg use systems Although they d1stmgU1shed so1ls by color and texture and called attent1on to lands e1ther havmg a sub surface compactad layer or low waterlogged areas they d1d not employ such d1stmct1ons m choosmg areas to clear and cult1vate Farmers s1mply cleared land m steady sequence from the areas closest to the roads and houses and towards the rear of the1r parcels Although a few named plant spec1es wh1ch md1cated so1l 1mpovenshment and others wh1ch md1cated fallow regenerat1on most farmers appeared to have httle concept of use of fallows for b1omass (and subesequent so1l fert1hty) regenerat10n or of (as ment10ned) the use of f1re to release nutnents for crops use Only a few Theobroma farmers ment1oned that there were med1c1nal plants that could be harvested from the forest 45 .¡:. "' FIGURE 2 DEFDRESTRTION CAUSES & FRRMER RESPONSES TO RESERRCH TO OECREASE ARTES CLEARING CATTLE BRNK PERENNIRL CREDIT FOR IMPROVEMENTS RCCOUNT BUT CROP PRICE RGROFORESTRY FRCIL!TRTING AECUIRE - FLUCTUATI ONS PRID IN RICE CREOJT TITL!NG PASTURE FENCES OISCOUAAGE MORE & SALES & PONDS RGROFORESTRY CLEARING RICE MAIN CASH NEEDED POLI CY FEW VIABLE & CONSUMPTIDN INCOMPATIBLE ECONOMIC CROP ONLY ON DEFORESTATION 11 FARMERS & AL TERNA TI VES NEWLY CLEARED RRNCHERS C!NCLUOING LRNOS NEEDS NTFP•l SOIL NUTRIENT RRP ID PRSTURE COST CA ADOPTABLE C?l LDSS DISERSES DEGRRDRTION INEFFECTIVENESS INTEGRATED & INVRSION OF FOLLOWED BY ~ OF TECHNICRL PEST NUTRIENT lEEOS & FOREST MORE PRSTURE AL TERNA TI VES FRLLOW SPECIES FORMRTION BLOCKS ADOPTION MANAGEMENT LRNDS CLERRED EXTENSIVE UNKNOWN CLERR PROBLEMS EVERY ONE CA YERRLY EFFECTS ON OF LRCK OF CROP TWO VERAS FOR PRSTURE - OEFORESTATI ON & PASTURE FOOO CROPS BURN!NG OF IMPROVING SUSTRINABILITY PASTURE --' '- -- ~ CONCLUSIONS The heartemng news 1s that rates of deforestat10n m the Amazon Basm appear to be decreasmg dueto fewer mcent1ves to large corportat1ons to mvest m cattle ranchmg a v1rtual end to a penad of road bUIIdmg to open up the Amazon and to protect front1er areas and a decline m government ass1sted colomzat1on programs for the rural poor Secondary forests have shown relat1vely h1gh rates of regenerat1on both after slash and burn agnculture and after abandonment of degradad pasturas and a s1gmf1cant proport1on of deforestat1on for sh1ftmg cult1vat1on 1s now of secondary rather than pnmary forest On the other hand deforestat1on and/or consohdat1on of colomsts parcels to form or expand ranches has contmued w1th ranches now bemg formed less by front1er nsk takers and more by urban speculators m areas where land pnces have nsen and where govenment can protect such mvestment The ev1dence presentad suggests that deforestat1on on colomsts parcels contmues at an apparent steady pace dnven by food product1on needs and by mcent1ves s1m1lar to those now pushmg format1on of largar ranches To sorne extent the charactenzat10n data suggest the need for on farm research to develop alternat1ves to slash and burn agnculture Such research already underway or planned needs to examme the poss1b1hty of makmg nce product1on more sustamable (and thereby hopefully reducmg demand for newly cleared forestl 1mprovmg (1 e mtens1fymg to the degree appropnatel pasture and cattle management 1ntroducmg or encouragmg more perenmal croppmg and agroforestry and 1mprovmg fallows At the same t1me and equally or more 1mportantly th1s analys1s recogmzes the 1mportance to deforestat1on of nat1onal pohc1es regardmg road construct10n cred1t tax mcent1ves front1er settlement land tenure and the rural poor Development of alternat1ves w11l need to mclude susbstant1al attent1on to pohc1es and future pohcy opt10ns 47 CHAPTER 11 PROJECTS TL 01 and TL-03 PROTOTYPE SUSTAINABLE CROPPING SYSTEMS FOR THE LLANOS ANO DYNAMICS OF LAND USE Edgar Amézqwta (TLJ Wlfllam Be// (LM SRGJ Peter Jones (LM SRGJ Jose 1 Sanz(TL Pro¡ect Off1cer) Joyotee Sm1th (TLJ Raul Vera (TLJ Phanor Hoyos (TLJ D1ego L Molma (TLJ A INTRODUCTION JUSTIFICA TION The purpose of th1s pro]ect 1s to generate technolog1es that allow lastmg mcreases 1n the eff1c1ency of resource use and that control s01l and water degradat10n The Colomb1an Llanos have trad1t1onally been ded1cated to extens1ve cattle ranchmg but th1s 1s changmg through the mtroduct1on of annual crops and new cattle based systems Remote sensmg and ground truthmg stud1es md1cate mcreasmg s1gns of env1ronmental degradat10n such as eros1on and loss of nat1ve spec1es as a result of the more mtense use of nat1ve vegetat1on by cattle as well as other non su1table forms of land preparat1on and management Th1s pro]ect a1ms to reconc1le more mtens1ve agncultural product1on w1th conservat1on and enhancement of the natural resources for the d1verse cond1t10ns found throughout the Colomb1an Llanos The proJect mcludes amongst other aspects long term tnals a1mmg to contrast ex1stmg and new technolog1es generated by the Tropical Lowlands Program (TLPl for the South Amencan savannas B PROBLEM SOLVING STRATEGY ANO EXPECTED OUTPUTS The followmg stud1es are bemg carned out and 1llustrate the proJect s strategy Trad1t1onally nat1ve vegetat1on for grazmg m the savannas 1s burnt morder to have fresh and tender new regrowth to feed cattle W1th the poss1b1hty of growmg crops 1t 1s also necessary to d1spose of e1ther nat1ve vegetat1on and/or crops/weeds res1dues Burmng 1s a cheap way for domg 1t and our work has shown pos1t1ve effects of burmng res1dues on crop y1elds Unfortunately burmng 1s not ecolog1cally sound and for obv1ous reasons the Program has started research smce early 1994 1n order to look for econom1cally feas1ble and env1ronmentally fnendly pract1ces to d1spose of such res1dues w1thout burmng or at least to reduce the frequency of 1ts use In the last few years several 1mproved pastura spec1es grasses and legumes adaptad to ac1d so1ls have been successful m systems w1th nce e1ther for the1r establishment or the1r recuperat1on New crops have been makmg the1r way mto the ac1d savannas One of the successful crops has been ac1d s01l tolerant ma1ze developed by CIMMYT and CORPOICA Before the ralease of the f1rst vanety for savannas and through collaborat10n between CIMMYT and the TLP 1t was dec1ded to evaluate from early 1994 ma1ze pasturas systems also to have another alternat1ve to the not so much des1rable monocrop The mew ma1ze vanety cv S1kuam 1 1s adaptad to so1ls w1th up to 60% Al saturat1on Th1s means hmmg the ongmal s01l w1th 1 1 5 t/ha to lower Al saturat10n whereas fert1hzat1on reqUirements are s1m1lar to those of ramfed nce Th1s amount of hme represents a majar quahtat1ve change m croppmg systems and for the Llanos at least 1t IS representativa of relat1vely h1gh mput systems m wh1ch the so1l 1s part1ally mod1f1ed to su1t a more demandmg crop lt 1s further hypothes1zed that the residual value of hme and fert1hzers allow the 1ntroduct10n of relat1vely more demand1ng forage germplasm such as Pamcum maxtmum cv Vencedor selected m Braz1l for the better endowed Cerrado solls In arder to cope w1th s1te spec1f1C1ty and lookmg for both extrapolat10n of results from tnals and collectlon of data m representat1ve areas the followmg cross sect1onal stud1es are bemg carned out Farms located 1n contrastmg land systems m the Colomb1an Llanos d1ffer slgmf1cantly m the use of resources Sorne of these d1fferences are assoc1ated w1th vanat1on m landscape relat1ve d1stance from markets and resource endowment lt has been hypothes1zed that at least sorne of the d1fferences 1n land use between farms m1m1c changas that w1ll take place over t1me so that a cross sect1onal study can represent also the1r temporal evolut1on lf th1s were 1ndeed the case relat1vely short term stud1es would be an adm1ttedly 1mperfect proxy for much longer term long1tudmal stud1es A set of seven contrastmg farms m terms of s1ze locat1on and product1on systems are bemg momtored m the two mam land systems of the Colomb1an Llanos #201 w1th 424 000 ha of flat land Vlrtually 1dent1cal to #202 w1th 1 447 800 ha and #203 w1th 1 200 000 ha of rolhng slope topography Whole farm mputs and outputs are bemg momtored from the beg1nmng of th1s year W1thm each farm f1xed pomts located m d1fferent areas of the landscape and subject to d1fferent uses have been 1dentlf1ed As a mm1mum and cond1t1onal to relat1ve mtens1ty of use so1l samples are obtamed tw1ce a year for detallad phys1c chem1cal charactenzat10n w1th the obJeCtlve of relatmg present use to s01l parameters An attempt at charactenzmg s01l b10log•cal act1v1ty 1s bemg made m cooperat1on w1th the Un1vers1dad Tecnológica de los Llanos (UT) S1m1larly m those areas under nat1ve vegetat10n or mtroduced pasturas surveys are be1ng made for botamcal compos1t10n and stockmg dens1ty and grazmg also w1th the cooperat1on of Umvers1dad Tecnológica de los Llanos A deta1led GIS coverage 1s bemg developed m the Colomb1an Llanos study area (between Puerto Lopez and Puerto Ga1tan Department of Meta) Th1s 1s to help the Lowland Trop1cs team to class1fy the1r expenmental s1tes to extrapolate from them and to ass1st m momtonng the change m land use and the adopt1on of new technolog1es as they are developed The Augustme Codazz1 1 100 000 scale map of s01ls for the area has been d1g1t1zed A user mterface has been developed for th1s GIS coverage lt mcludes data on s01l charactenst1cs dramage topography etc m 11 d1fferent data coverages The 1 25 000 topograph1c maps are 1n the process of bemg d1g1t1zed The topography has been completad D1g1t1zmg of the dramage lmes 1s at present m process the1r d1rect10n and the start and endpomts of each of the dramage ares bemg noted Th1s w1ll be used 1n con]unct10n w1th a Laplac1an spline algonthm to calculate the d1g1tal elevat1on model for the whole of the study area The model 1s needed to accurately calculate slopes and runoff for applymg catchment management and eros1on control models D1g1t1zmg topography at th1s scale 1s very labor mtens1ve In fact th1s proJect has taken up over 3 man years of work to date 50 Satelhte 1mages for vanous prev1ous years have been obtamed and at present the 1989 Landsat TM 1mage IS bemg analyzed The analys1s bemg conducted 1s kpown as an unsuperv1sed class1f1cat1on Smce ground truth for 1989 1s unava1lable the class1flcat10n algonthms of the satelhte 1mage analys1s software are allowed to produce the best poss1ble resolut1on of class1f1cat1on of the 1mage w1thout the type of ground cover 1t 1s denotmg m 1ts class1f1cat1on actually bemg known Th1s 1s now bemg usad to produce a strat1f1ed samphng wh1ch w1ll allow f1eld work and samphng of actual ground truth (1 e recordmg the actual land use on the ground 1n 1nd1v1dual f1elds regeoreferenced to the satelhte 1magel Th1s w1ll be at the end of January 1995 A Landsat TM 1mage has been comm1sS1oned wh1ch w1ll then be d1rectly relatad to the ground truth of the area Th1s w1ll allow construct1on of a superv1sed class1f1cat1on m wh1ch the actual land use 1s known W1th the superv1sed class1f1cat1on and the topograph1cal 1mages from the d1g1tal elevat10n model that class1f1cat1on can then be transposed back m t1me to the prev1ous 1mages and trends m land use obtamed Senes of a1r photographs datmg from the 1940s at roughly 5 year mtervals are also avallable Havmg obtamed the superv1sed land use class1f1cat1on from the most recent satelhte 1mage certam study areas wlll be selected for mterpretat1on from the a1r photographs and a senes of land use change 1mages runmng back over the last 40 years w!ll be constructed These w1ll be used as base data for producmg models of land use change 1n the area Other sources of data that w1ll be gradually mcorporated mto the respective database refer to descnpt1ons of typ1cal so1l prof1les results of prev1ous farm surveys conducted by CIA T vegetat1on surveys and others Th1s dynam1c database w1ll serve among other purposes to better dehm1t domams of extrapolat1on for technolog1cal components and prototype croppmg and farmmg systems Used together w1th GIS based s1mulat1on models of land use 1t w1ll allow eventually the spat1al and temporal s1mulat1on of alternat1ve land uses and to evaluate the1r 1mpact on the reg1on For an analys1s of land use dynam1cs m the Llanos see report for ProJect T A 02 See also Subprojects T 01 2 Adopt1on of ley farmmg systems m the Colomb1an Llanos and T 03 1/2 Vegetat1on and ecology of the serranfa nat1ve pasturas of the Eastern Plams of Colomb1a Llanos and Nat1ve pasture management For the purpose of th1s report and m arder to 1llustrate the strategy two on gomg long term contrastmg tnals started m 1989 wlll be d1scussed here These two tnals represent contrastmg systems for wh1ch CIA T m collaborat1on w1th other 1nternat1onal and nat1onal mst1tut1ons has been developmg both the germplasm (crop and pasturesl and the technology for the1r management One of them contmuous upland nce monocroppmg was estabhshed s1nce early 1989 w1th the purpose of followmg closely what could be the most obv1ous type of adopt1on of the then prom1smg germplasm now released The fear then was that of further degradat10n of the already very ac1d mfert1le and eros10n/compact1on prone Ox1sols 1f ut1hzed under contmuous monocroppmg The 1mt1al hypothes1s was that y1eld would declme ovar t1me and the so1l resource would degrade after a few years of contmuous croppmg The tnal has been closely momtored and attempts were made to opt1m1ze so!l chem1cal 51 phys1cal and mlcrobiologlcal changas as well as weed mc1dence w1th the a1m of mamtam1ng y1elds w1thout decline 1 e trymg to make 1t sustamable 1f at all poss1ble / In contrast an expenment m wh1ch nce and grass legume pasturas are sown together at the same t1me and shanng the same f1eld was also established early 1989 Th1s was one of CIAT s technolog1cal alternativas to the potent1al nsk of monocroppmg and also an alternat1ve to trad1t1onal ranchmg The hypothes1s was that 1t would be poss1ble to produce nce econom1cally while establishmg 1mproved pasturas wh1ch m turn would benef1t from the res1dual fert11ity left by the nce crop W1th th1s monocroppmg would be av01ded and v1gorous pasturas would cover the soil for a few years unt1l a pastura recuperat1on was needed through another nce crop Ammal product1v1ty was expected to be enhanced as well as preservat1on of the resource base 8oth tnals above have been followed w1th penod1cal nutnent balance sheets These mclude the nutnent part1t1omng to the d1fferent parts of the system 1 e m the so1l (befare and after each croppmg season) added fertilizar nutnents uptake by plants (crops grasses legumes weedsl as well as percentage taken out w1th produce (eg nce gram) nutnents f1xed m roots or lost/not found AdditiOnally soil phys1cal measurements m the second tnal above are also bemg useful to momtor compact10n of grazed vs non grazed pasturas These pasturas also y1eld valuable mformat1on on rates of forage b1omass product1on Records on the use of machmery are ava1lable Th1s will allow us to draw balances on the use of foss1l energy C RESUL TS ANO DISCUSSION a) Cont1nuous upland nce monocroppmg In the contmuous nce monocrop expenment there was a contmuous linear y1eld decline for Oryz1ca Sabana 6 at a rate of nearly 400 kg/ha/yr (r2 =O 96) over the penad 1989 to 1992 Desp1te fert1hzer mputs bemg ad¡usted yearly based on so1l tests to mamtam appropnate nutrrent levels and balances y1elds declinad from 3 8 to 2 6 t/ha In 1993 w1th 1mproved agronom1c pract1ces y1elds p1cked up by about 700 kg/ha Smce weeds mc1dence was one of the most clear obstacles to have a stable product1on the use of herb1c1des was mcluded m 1994 Unfortunately the1r eff1c1ency m the savanna Ox1sols 1s not nearly as good as 1t 1s m other nce growmg areas mcludmg the P1edmont of the Llanos Th1s w11l probably requ1re s1te and crop spec1f1c research on weed management wh1ch should be carned out by local NARS and others See SubproJect T 01 1 Weed populat1on ecology w1thm prototype sustamable croppmg systems for the Colomb1an Llanos Added to th1s and from the so1l phys1cs pomt of v1ew severa! measurements have been carned along w1th the tnal to document changas mduced w1th the pract1ce of contmuous nce croppmg Plant roots elongate and enlarge m d1ameter 1f the pressure ms1de new cells 1s suff1c1ent to overcome externa! res1stance causad by surroundmg so1l matnx Sorne researchers have demonstrated that the max1mum Jong1tudmal root growth pressures of severa! crops range from 9 to 15 bars (1 bar"' 1 kg/cm 2 ) and that roots (Raphamts sattvus L l cease to enlarge 52 (dlameter) when sub1ected to rad1al constramts of about 8 5 bar Roosell and Goss (1974) showed that O 2 bar pressure apphed to a 9lass bead system reduced the elongat1on of barley root by 50% and that O 5 bar reduced the rate by 80% lf a s01l 1s eas1ly deformable roots penetrate 1t and grow unt1l sorne factor d1fferent from mechamcal1mpedance stops elongat1on rate In most cases roots grow partly through ex1stmg pore spaces and partly by mov1ng as1de so1l part1cles Most of the penetrab1hty values found m Matazul (Figure 1) are h1gher than those generally reportad as hm1tmg values (9 15 bar) speCially at depths larger than 7 5 cm no matter the t1llage (disk harrow mould board ploughmg) treatments lt IS therefore hkely that mechamcal 1mpedance 1s negat1vely 1nfluencmg nce y1elds F1gure 1 also shows that the savanna average res1stance values are closer together or have a lower range than those obtamed by t1llage treatments They vary from 12 bar (2 5 cm depth) to 22 bar (32 5 cm depth) approx1mately whlle m the d1sk harrowmg treatment they vary from 7 (2 5 cm) to 24 bar (32 5 cm) and m mould board ploughmg treatment from 8 (2 5 cm) to 25 bar (32 5 cm) These results suggest that the control (savanna treatment) exh1b1ts better pore contmUity Another aspect shown m F1gure 1 1s that all treatments mcludmg the control present s1m1lar values at 22 5 cm depth showmg that there 1s presumably a compactad layer or a change from so1l to subso11 wh1ch could affect so1l processes and be determmant to y1elds lt 1s clear from F1gure 1 that t1llage loases the so1l from O to 22 5 cm depth but compacts 1t from 22 5 downwards and that mould board ploughmg mcreases compact1on m depth The stat1st1cal analys1s of the data (Table 1) shows that there are s1gn1f1cant (P< O 05) d1fferences m the values of penetrab1hty when treatments are comparad at 2 5 7 5 cm depth Nat1ve savanna presents the h1ghest values wh1ch means that 1s more compactad In the 7 5 1 5 O cm and 15 O 22 5 cm depth the analys1s shows more Similar values and less d1fference between treatments but st1ll savanna has S1gn1f1cantly h1gher values at 7 5 15 O cm At 22 5 32 5 cm depth mould board ploughmg treatment shows S1gn1f1cantly h1gher values than savanna hence causmg deep soll compact1on Table 1 Average so11 penetrab1hty values m the s1xth year of contmuous upland nce monocropp1ng m Matazul Llanos of Colombia 1994 Soll depth (cm) 25751 751501 15 o 22 5 1 22 5 32 5 Treatment Penetrab1hty kg/cm 2 Savanna 13 9 a11 16 2 a 17 7 a 19 7 b Mould board plough/93 10 8 b 14 7 ab 17 2 a 22 7 a D1sk harrow/93 93b 14 5 b 17 5 a 21 1 b 1/ Data m the same column followed by the sama letter do not d1ffer S1gn1f1cantly (P.. ~ o o 8o 500

O 051 between groups 1 and 2 and between groups 3 and 4 (data not shownl and by the short projeCtiOn of the vectors of most spec1es (Figure 4 1 Regardmg the fert1hty levels C pubescens and Paspalum mult1caule were 1nd1cators of h1gher fert1hty as shown by the1r vectors wh1le A purpusu and D1g1tana sangumafls (a known ruderall relatad to lower fert1hty (Figure 4 1 The relevance of weeds 1n rendenng nce monocroppmg unsustamable 1s clear from the far greater y1eld response to weedmg than to mcreased fert1hty (Tabla 6 1 Sorne of these weeds could have been mtroduced w1th agnculture mto the Llanos By fert1hzmg a weedy nce crop y1elds were reduced by the compet1t10n from weed commumt1es that had sh1fted towards spec1es who stnve 1n such 1mproved cond1t1ons as shown by the COA analys1s lt had already been observad 1n Similar env1ronments that weeds could take up more Ca and K than nce (San 19941 W1th the current weed pressure externa! weed controlmputs for nce monocroppmg w1ll be needed Also to deal w1th the present flora combmat10ns of herb1c1des for a broader spectrum of control may be reqwred Th1s w1ll have an 1mpact on the farm1ng economy the env1ronment and on people s health COA can help 1dent1fy the ecolog1cal role of spec1es w1th1n plant commumt1es lf the mam ecolog1cal roles of relevant weeds and pasturas are understood a pastura spec1es could be chosen such that 1t would successfully se1ze the mches opened by the new agncultural mputs and substituta for sorne of the weeds A 63 legume appeared strongly assoc1ated to h1gher levels of P and hme (F1gure 4) th1s would also su1t other forage legumes as well Therefore the COA techmque appears prom1smg as a tool for plant populat1on analys1s and to 1dent1fy spec1es assoc1ated w1th certam patterns of resource ava1lab1hty Further stud1es w1ll relate relat1ve spec1es abundance to other s01l and agronom1c parameters Table 8 References Part1al results of the canon1cal d1scnmmat1on analys1s for two levels of fert1hzat1on Wlth data from an expenment followmg f1ve consecut1ve years of nce monocroppmg Matazul 1994 Group p K L1me 1 25 50 300 2 43 83 500 Prob > Mahalanob1s d1stance group 1 to 2 = o 0073 H1ghest total sample standard1zed coeff1c1ents for canomcal vanable Axonopus purpusu 1 76 Centrosema pubescens 3 55 Ayarza M A 1994 Sustamable Agropastoral Systems for the Cerrados Pages 67 85&n CIAT Savannas Program B1enmal Report 1992 1993 Workmg Document No 134 Bonham C D 1989 Measurements of terrestnal Vegetat1on J W1ley & Sons New York Derksen D A G P Lafond A Gordon Thomas and C Swanton 1993 lmpact of agronom1c pract1ces on weed commumt1es T1llage systems Weed Sc1 41 409 417 Derksen D A A Gordon Thomas G P Lafond H Loeppky and C Swanton 1994 lmpact of agronom1c pract1ces on weed commumt1es Fallow w1th1n tlllage systems Weed Sct 42 184 194 Fnesen D K 1994 Nutnent Cychng Pages 1 13 in CIAT Savannas Program B1enn1al Report 1992 1993 Workmg Document No 134 Johnson R A 1992 Apphed Multtvanate Stat1St1cal Analys1s Th1rd Ed1t1on Prent1ce Hall Englewood Chffs New Jersey S A S lnst1tute lnc 1989 User s Gu1de Vers1on 6 Fourth Ed1t1on Volume 1 Cary NC S A S lnst1tute lnc San J 1 and R R Vera 1994 Prototype Croppmg and Farmmg Systems Pages 51 65 1n CIAT 64 Savannas Program B1enmal Report 1992 1993 Workmg Document No 134 Walter H 1983 Weed sampling m the f1eld and mterpretat1on of sampling data Pages 65 80 m H Walter (ed ) PLITS Weed management m the Ph11ippmes Report of Semmars Stuttgart Federal Republic of Germany SUBPROJECT TL 01 2 ADOPTION OF LEY FARMING SVSTEMS IN THE COLOMBIAN LLANOS Jose Vtcente Cadavtd and Joyotee Smtth (TLJ Product1on systems m the most mtens1f1ed areas of the savanna (mamly m the center south of the cerrado and m the western savanna of Venezuela) are charactenzed by contmuous monocroppmg and contmuous t1llage w1th heavy machmery Whlle these systems are prof1table they result m so1l eros1on compact1on reduced m1crob1olog1cal act1v1ty declimng quality of orgamc matter and loss of s01l phys1cal propert1es Pest bUIId up 1s also a charactenst1c of these systems Sem1 mtens1ve cattle ranches are also found m these areas where about 54% of the pastura area cons1sts of plantad pastura w1th stockmg rates of about 1 24 heads per hectare of plantad pastura The rest of the savanna 1s mamly charactenzed by extens1ve ranches cons1stmg mamly of nat1ve savanna grasses and w1th low stockmg rates often less than O 2 heads per ha Pastura degradat1on m both these systems 1s common w1th bald patches leadmg to eros1on and weed mfestat1on The crop pastura technology was developed m response to these problems An annual crop and a grass legume pastura IS sown The annual crop 1s fert11ized The pastura benef1ts from the res1dual fert11ity leadmg to qUicker establishment and better pastura quality Th1s makes 1t poss1ble to graze the pasture 1mmed1ately after the nce harvest After bemg grazed for about three years the pasture 1s renovated by plantmg the fert11ized annual crop and the legume aga1n The concept of the crop pastura technology was based on a v1s1on of mtens1f1ed and sustamable agncultural development throughout the savanna Savanna mtens1f1cat10n was expected to mcrease the returns to capital mvestment m the savanna and d1vert venture capital from the Amazon ram forest lntens1f1cat1on was also expected to reduce expans10n of the agncultural front1er m the savanna and thus conserve plant and ammal spec1es W1thm th1s v1s1on the ob¡ect1ve of the crop pastura technology was to prov1de a susta1nable alternat1ve to contmuous monocroppmg m mtens1f1ed areas mduce mtens1f1cat1on m extens1ve areas Wlthout the resource degradat1on that charactenzes contmuous monocroppmg and to contnbute to resource preservat1on by renovatmg degradad pasturas and mamtammg pastura quality The ley farmmg system was expected to ach1eve th1s by elimmatmg t1llage dunng the pastura phase of the rotat1on and by mamtammg pastura quality through the res1dual effect of the fert11izer applied to the annual crop The avallab11ity of ac1d tolerant nce and ma1ze meant that remed1al lime could be substant1ally reduced Grass legume pasturas were cons1dered to be more sustamable than pure grass pasturas as farmers were cons1dered to be unw1lling to fert11ize pasturas m v1ew of the long lead t1me mvolved 1n obtammg a return In add1t1on to f1xmg mtrogen legumes also st1mulate so1l b1olog1cal act1v1ty 1mprove nutnent cycling and the nutnt1ve value of forage and the system s carbon sequestrat1on ab11ity The system has also been shown to 1mprove s01l phys1cal and chem1cal propert1es m companson w1th annual crops and nat1ve savanna grasses At the same t1me cash flow versus pasture 65 only systems was expected to be better because an annual crop would be ava1lable for sale a few months after the cost of pasture establishment was mcurred and the pasture would be ready for grazmg 1mmed1ately after the crop harvest Pasture quallty was expected to be mamtamed because the mcent1ve to obtam cash revenue from the annual crop was expected to lead to a crop bemg sown every tour years or so thus prov1d1ng mamtenance fert1llzer for the pasture As charactenzat1on of the savanna and analys1s of land use dynam1cs proceeds sorne of the assumpt1ons on wh1ch th1s strategy 1s based are commg under scrutmy The study reportad here prov1des 1ns1ghts mto one of these assumpt1ons that technology can mduce mtens1f1cat1on 1n extens1ve areas lf th1s assumpt1on 1s correct technology development should be focused on mtens1ve technolog1es lf 1t 1s not correct a very d1fferent portfollo of technolog1es m ay be 1nd1cated wh1ch reconc1le farmer demand for technology w1th strateg1es for ecosystem management Econom1c analys1s of the crop pasture technology w1th nce as the annual crop was carned out 1n 1990 based on tnals under controlled expenmental cond1t1ons The results gave attract1ve mternal rates of return for beef product10n on med1um and large scale farms but not on small scale farms md1catmg the ex1stence of econom1es of scale R1ce y1elds of 2 to 2 5 t/ha were suff1c1ently h1gh to offset nce product1on costs and make a substant1al contnbut1on towards the cost of pastura establishment The nce pasture system also prov1ded net pos1t1ve mcome streams from the th1rd year as opposed to f1ve years for pasturas establlshed w1thout nce Wh1le the technology offers these advantages 1t 1s also considerad more cap1tal and management mtens1ve than e1ther the trad1t1onal system based on nat1ve savanna grasses or planted pasturas Sorne of the requ1rements of movmg from the trad1t1onal system to 1mproved pure grass pasturas 1nclude land preparat10n fencmg and the cost of grass seed Add1t10nal ammals too m ay be reqUired to obtam adequate returns from the 1mproved pasturas A ma¡or add1t1onal 1nvestment m management 1s also reqUired The trad1t1onal system 1s capable of bemg managed by absentee landlords who make occas1onal v1s1ts to the farm and leave the day to day runmng of the farm to a relat1vely untra1ned h1red hand The planted grass pasturas reqUire controlled grazmg and rotat10n of ammals to prevent over grazmg on the one hand and to prevent the grass growmg overly tall on the other Thus proper management by absentee landlords 1s hkely to mvolve the expense of h1nng a res1dent tra1ned admm1strator The move to grass legume pasturas 1s st1ll more management mtens1ve as 1t the correct balance between grasses and legumes have to be mamtamed Th1s reqUires more control over grazmg the establishment of more paddocks for rotat1on of ammals and therefore more fencmg Legume seeds are more expens1ve than grass seeds and more d1ff1cult to produce on farm lnvestment m expandmg the herd 1s also necessary to recuperate the extra cap1tal and management costs A sh1ft to the ley farmmg system reqUires a combme harvester seed dnll fert1llzer and msect1c1de for the annual crop Most 1mportantly though the crop component reqUires constant momtonng and therefore a change m attltude for those accustomed to pasturas wh1ch can often be left unmomtored for several weeks w1thout senous 1mphcat10ns lt appears therefore that nat1ve savanna grasses pure grass pasturas grass legume pasturas and ley farmmg systems he along a contmuum of mcreasmg 1nvestment m capital and management When land values are low mcreases m product10n can be obtamed more cheaply through area expans1on than by 1ntens1fymg mputs of capital and management per ex1stmg land area There 1s httle 1ncent1ve m max1m1zmg returns per land area because land 1s an abundant resource Farmers prefer technolog1es wh1ch reqUire m1mmum mvestment per land area Th1s explams the prevalence of extens1ve grazmg on nat1ve savanna grasses m areas of low land values Product1on 1s mcreased as the herd 66 expands through natural mcrease over t1me As land 1s abundant the expanded herd 1s mamtamed by bnngmg more nat1ve savanna grasses under grazmg by burmng and by ut1hzmg more low lymg areas dunng the dry season As the herd expands beyond the hm1ts of these pract1ces over grazmg occurs Plantmg a small area of 1mproved pastura 1s an alternat1ve to over grazmg and 1s also a means of 1mprovmg ammal nutnt1on and therefore reproduct1ve rates Adopt1on of small areas of 1mproved pasturas may be mduced therefore through naturalmcreases 1n herd s1ze As land values mcrease 1t 1s only worthwh1le holdmg on to land 1f h1gh returns per land area can be obtamed Thus mvestment 1n capital and management becomes more attract1ve When lncreasmg land values are mduced by 1mproved mfrastructure marketmg margms are reduced and output pnces mcrease relat1ve to 1nput pnces thus further mducmg mcreased levels of mput use Progress1on along the technology contmuum 1s therefore hkely to reqUire mcreased land values and 1mproved mfrastructure In the absence of these precond1t1ons technology 1s unhkely to be able to mduce mtens1f1cat1on Conversely mtens1f1cat1on 1s unhkely 1f the pre cond1t1ons ex1st but technology wh1ch makes prof1table mtens1f1cat1on poss1ble 1s absent The study reportad here mvest1gates adopt1on of 1mproved pasturas 1n the mumc1paht1es of Puerto López and Puerto Ga1tán 1n the department of Meta 1n the savanna of Colombia wh1ch 1s part of the OnnoqUia reg1on (Figure 51 73 randomly selected farms were surveyed 1n 1989 and 1992 Farms smaller than 40 ha were excluded from the sample The Colomb1an savanna 1s charactenzed by h1ghly aC1d1c s01ls (pH of 3 8 51 w1th h1gh levels of alum1mum ( > 80%1 Annual ramfall 1s 1700 2500 mm w1th a dry season of 3 4 months Extens1ve grazmg 1s the ma1n form of land use w1th the greatest pert of the vegetat1on cons1stmg of nat1ve savanna grasses Woody areas are found w1thm the savanna and gallery forests are located along the nver beds wh1ch are also the areas where small holders tend to be located Nomad1c 1nd1genous populat1ons of hunters and gatherers are al so found Most are undergomg a process of sedentanzat1on and the1r numbers are dechmng fast The department of Meta has grown faster than the rest of the country w1th GDP growth rates of 6 82% between 1970 and 1990 and 7 05% between 1980 and 1990 comparad 3 5% to 4% for the rest of the country Th1s 1s dueto the recent evolut1on of commerc1al agnculture and petroleum relatad mvestment Meta contnbutes 3% of nat1onal agncultural product and 2 7% of the mmeral output lt should be p01nted out however that the advanced agncultural area P1edmont 1s not part of the savanna ecosystem V1llav1cenc1o the mam urban center 1n Meta has expenenced rap1d growth 1n the recent past as a growmg d1stnbut10n center for consumar goods to 011 mstallat1ons 1n the east Populat1on dens1ty 1n Meta was 5 5 hab1tants per km sq 1n 1985 and 1s projected to mcrease to around 8 mhab1tants by 1 995 60% of the populat1on hves 1n urban areas Thus rural populat1on dens1ty 1s extremely low Transportat1on 1s a majar constramt In the reg1on of OnnoqUia only 1 6% of roads are paved The towns of Puerto López and Puerto Ga1tán 1n the study are a are located 94 and 207 km respect1vely from Vlllav1cenc1o The V1llav1cenc10 Puerto López road was paved m the early 1 980s but road connect1ons to farms away from the mam road or hamlets are frequently nonex1stent or only seasonally usable Recent majar 011 and natural gas d1scovenes m the eastern savanna are expected to lead to majar 1mprovements 1n mam roads but farm to market roads are hkely to remam poor 1n the short run Also s1gmf1cant 1mprovements to the road connectmg V1llav1cenc1o to the cap1tal Bogota sorne 90 km away are now bemg carned out The study area can therefore be charactenzed as a trad1t1onal extens1ve cattle ranchmg area w1th unfert1le s01ls very low rural populat1on dens1ty and 1n 67 the process of movmg from poor to mtermed1ate mfrastructure The area can be cons1dered fa1rly representat1ve of the Lat1n Amencan savanna except that 1t can be cons1dered to be towards the unfavorable extreme 1n terms of so1l fert1hty and 1n terms of mfrastructure movmg from the typ1cal towards a more favorable d1rect1on The rate of change bemg expenenced 1n the area 1s faster than 1n other areas such as the Braz1han savanna where front1er expans1on IS slowmg down Survey data conf1rm that rap1d changes are takmg place 1n the study area Between 1979 and 1992 average farm s1ze dechned S1gmf1cantly from around 5000 hato 1551 ha Th1s 1s apparently the result of new farmers movmg mto the area 52% of sample farms were purchased after 1981 and of these around 22% were acqUJred after 1990 Th1s has changed the character of the farmmg populat1on 1n the area Pnor to th1s t1me the bulk of farmers had owned the1r land for many years Sorne generally those Wlth lower levels of educat1on and therefore w1th lower opportumty cost res1ded on farm w1th subs1stence as the1r mam motive Others were absentee owners who carne from trad1t10nal ranchmg famJ11es and had acqUired the1r land through 1nhentance Both categones regarded land as a cheap abundant resource because 1t had been acqUJred w1thout cap1tal expend1ture on the1r part The1r pnonty therefore was a system reqUJnng httle management or capital mvestment per umt area By contrast 1n 1992 31 % of farmers employed techmcally quallf1ed res1dent adm1mstrators 32% pa1d for techmcal ass1stance 53% had the1r own tractor (Table 9) Educat1onal levels were h1gh w1th 41% havmg umvers1ty educat1on Only 21% llved on the farm 64% llved 1n Bogota The number of land purchases made for non product1on related mot1ves are unknown However the data show that product1ve mvestments 1n land were occurnng Adopt1on of 1mproved pasturas 1n the study are a mcreased at an annual rate of 14% between the 1978 1992 penod In 1992 173 000 ha or 17% of the sample farm area was planted to 1mproved pasturas Only 18% of the 1mproved area cons1sted of grass legume pasturas and 1 5% was planted to the crop pastura technology 82% cons1sted of pure grass pasturas (Tabla 1 0) The 1mproved ac1d tolerant nce vanety was only released 1n 1991 Therefore the low leve! of adopt1on of ley farmmg 1s not surpnsmg The most w1dely grown grass 8 decumbens was never off1c1ally released 1n Colombia though 1t was released 1n a number of other Latm Amencan countnes startmg w1th Cuba 1n 1987 The next most popular grass B hum¡d¡cola was released 1n Colombia 1n 1992 and 1n other countnes startmg m 1985 In the case of both these grasses spontaneous adopt1on occurred w1th farmers obtammg matenals from other countnes and unoff1c1ally from expenment stat1ons m Colombia and plantmg 125 000 ha by 1992 By contrast although the most w1dely grown legume S cap1tata was released 1n Colombia 1n 1983 adopt1on 1n 1992 was llm1ted to 28 000 ha lt 1s also notable that although only 17% of total farm area was plantad to any type of 1mproved pasture over 98% of farmers were adopters Apparently farmers started w1th a small area of 1mproved pastura and expanded gradually as cap1tal became ava1lable More than 90% of farmers mcreased the1r 1mproved pastura area over t1me w1th 1mproved pasture mcreasmg relat1vely slowly (9 6% p a ) between 1978 and 1981 and demonstratmg a steeper mcrease ( 1 5 5% p a ) after that penod Survey data reveal major d1fferences between recommendat1ons and farmers pract1ces 31 % of plots were not fert1hzed at establishment 96% rece1ved no fert1llzer for mamtenance purposes and 75% were nevar renovated 76% of farmers who plantad 8 decumbens cla1med that 1ts product1v1ty had dechned comparad to the f1rst two years after establishment w1th 43% cla1mmg product1v1ty declines of 50% or more Casual observat1on 1nd1cates that overgrazmg of 1mproved pasture plots dunng the dry season may have contnbuted to pasture degradat1on 68 "' U) 8 4--. ""' ' z < u o -41 1 -8 -8 GROUP 4 a 431qfha P 83 K 500 UMEAND RICE+LEGUME GROUP4 / u 1 1 , ~ ' a en~· GROUP3 • 43 K 500 UM lq/ha P 83 IN MONOC~::~.:CE -4 GROUP 1 1 GROUP 1 ,. 300 UME ~:11&/h• P 50 K, ' 1 MONOCULTURE RICE ··- 1 \~e GROUP :z K 300 UM•E :Z51q/ha p 50 RICE+LEGU~D 1 GROUP :z o 4 8 12 CAN! Figure 4 Canonical discnmmant analysis onhnatJOD diagram or treatment clusters ror two levels o combmed WJtb two croppmg modes, and b1plot scahng orweed spec1es vectors lndicatmg treatment as Symbols correspond to observatlon s1tes and tbe1r posJtJon JS located based on relatJve spec1es comp SJgnlficance or ddTerences between treatments are dJscussed ID tbe text SpecJeS assocJatJon WJtb S mrerred rrom tbe dJredJOD or tbe vectors Vector lengtb mdJcates the relatJve strength or assocJ • j • ~ ~i ¡.. ¡.. ¡j !; j ~~ 'S ! E§ ii ~ ~ !; ~ ~~ ~ ~ ~ i ~ ~ ~ ...l "' ~ t:l r: 70 Tabla 9 Charactensttcs of study area 73 Farmers tn Alttllanura of Colombta 1992 Vana bies Su m Average Std Maxtm Mtntm Dev Pastura area (hal 18756 256 93 473 42 2900 1 Farm SIZO (hal 113199 1550 67 2276 75 16000 40 Area w1th forest (hal 4820 66 03 99 49 500 Ares w1thout forest (hal 108379 1484 64 2220 95 15600 32 Mechan1zable ares (hal 66954 917 18 1170 07 7100 Tractor ownersh1p (dummyl 39 o 5342 o 5022 1 o Labor use (wages/ha yearl 159 29 2 18 2 50 9 12 010 Loan (dummy) 18 o 25 1 o 01stance from Pueno López (Kml 6642 9098 69 51 275 9 Farm admtmstrator (dummy) 23 o 31 o Res1dence at farm (days/yearl 104 25 126 84 365 Res1dent of Santafé de Bogotá 47 064 o 4821 1 o (dummyl Res1dent of V¡lfavlcenclo (dummyl 11 015 o 3602 1 o Farm res1dent (dummyl 15 o 21 1 o Untverstty educatton tdummy) 30 o 41 1 o Secondary educatton (dummy) 18 o 22 1 o Pnmary educatton (dummy} 27 o 37 1 o Age (yearsl 52 01 12 23 78 30 Years of farm ownershtp 12 09 10 73 44 Land t1tle (dummyl 56 077 o 4255 o Labor use/year 69286 945 10 875 97 6020 292 Famdy labor use/year 7139 89 27 186 70 730 o Other labor use/year 62147 172 58 441 03 3245 o Other actiVIIIes (dummyl 62 086 o 36 1 o Farm purchase 1981 1992 (dummyl 38 o 62 o 60 o Expenence w1th crops (dummy) 6 007 1 o Techmcal asststance {dummy) 23 o 32 04677 1 o 71 Table 10 lmproved pastura area 11 on farms21 1n the Alt1llanura of Colombia Are a est1mate 11 (ha) % Farm area Pasturas Grass and Legumes Andropogon gayanus Brach!ana decumbens Brach1afla hum1d!co/a Brach1afla d1ctyoneura Other Grass Total grass S cap1tata + 8 decumbens S cap1tata + 8 d1ctyoneura S cap1tata + 8 humldlcola S cap1tata + A gayanus e acutlfollum + A gayanus Other assoc1at1ons Total grass + laguma pasturas Legumes31 Crop + Pastura Total 1mprova pasturas Area nature savanna grass Total farm area 1989 4453 45 559 26 479 1 371 3 598 81 604 1 490 847 o 2480 339 1 794 6950 323 88 877 871 459 960 336 1992 7 129 67 580 57 476 5 584 4 013 141 782 15 137 5 558 6 765 621 o 2 699 30 780 277 2 737 176 676 850 408 1 026 984 1/ Error of est1mat1on 20% Leval of conf1dence 80% Survey 86 and 82 farms 1n 1989 and 1992 respect1vely 1989 1992 50 4 1 51 3 38 5 29 8 32 7 1 5 32 42 23 91 8 80 8 1 7 86 1 o 32 00 39 28 04 04 00 20 1 5 78 17 3 04 02 1 5 100 o 100 o 90 7 82 9 2/ Farms greater that 40 ha 728 localizad m Puerto López and Puerto Ga1tán (Meta Colomb1a) 3/ Legumes (Seed bed) S capltata A pmto/ e BCUtlfO/IUm p facilitieS Around 75% of producers had cow and calf operat1ons The rest also carned out sorne fattemng Although the number of producers who fattened mcreased from 21% to 28% between 1979 and 1989 the d1fference was not statiStlcally s1Qnlf1cant 72 T1me senes regress1on analys1s over the 1978 1992 penod revealed that the construct1on of the paved road between V11lav1cenc1o and Puerto López had a stat1St1cally s1gmf1cant effect on the net accumulated area plantad to 1mproved pasturas as d1d mcreases m real terms m agncultural GDP and mcreases m the supply of meat (Table 11) The paved road would be expected to st1mulate the use of mput mtens1ve technolog1es by mcreasmg land values and reducmg mput/output pnce rat1os wh1le the other vanables represent generally favorable market cond1t1ons for agnculture Table 11 T1me senes mult1ple regress1on est1mates of factors determmmg cumulat1ve net area m 1mproved pasturas 1978 1992 Exogenous Slgns and t Level of Rz J DW F Vanables Coeff1c1ents stat S1gmfic Constant 39 81 2 76 o 020 o 93 1 44 51 02 Agncultural GDP" (Ln) +2 59 3 27 0008 Cattle Supply (Ln) +1 07 1 85 o 093 Paved Road (dummy) +0 65 3 94 o 003 Number of released +0 08 1 55 O 15 ns forages B/C mdex + ns 1/ Constant 1978 pesos R2 1 Coeff1c1ent of mult1ple correlat1on ad]usted D W Coeff1c1ent of autocorrelat1on Durbm Watson F F stat1st1c n s Not s1gn1f1cant at the O 1 O level Ln Natural loganthm Cross sect1onal regress1on analys1s showed that adopters were those who mvested m the1r farms tractor owners those who had farm admm1strators those who pa1d for techmcal ass1stance Other factors md1cated that adopt1on mcreased w1th land values The area w1th 1mproved pasturas was h1gher on farms w1th legalland tltles and close to Puerto López Also there was a negat1ve relat10nsh1p between 1mproved pastura area and forestad farm area Extens1ve forestad area md1cated land abundance wh1ch made the adopt1on of mtens1ve technolog1es unnecessary (Table 12) Contrary to w1dely held v1ews res1dence on farm was negat1vely relatad to adopt1on Th1s m ay be because those who res1ded 1n urban centers had better access to cap1tal and were therefore more hkely to adopt technolog1es wh1ch reqUired considerable cap1tal mvestment These more cap1tahzed farmers were able to adopt a management mtens1ve technology 1n sp1te of hvmg off farm by h1nng tramed res1dent adm1mstrators Unexpectedly years of educat1on were negat1vely relatad to adopt1on When asked about the1r reqUirements farmers sa1d they wanted nutnt1ous pers1stent pasturas adaptad to the1r s01ls and res1stant to pests Most 1mportantly farmers wanted pasturas that met these cond1t1ons under the1r level of management Pers1stence of legumes under farmer 73 management was seen as a particular problem Farmers therefore appear to be unwllhng to change the1r level of 1nvestment and management 1n arder to take advantage of new technolog1es Tabla 12 Cross sect1on mult1ple regress1on est1mates of factors determ.n.ng area 1n 1mproved pasturas 73 farms 1992 S1gns and t Exogenous Vanables coeff1c1ent Stat1st est1mates Constant + 133 835 o 799 Area w1th forest (ha) o 995 2 474 Area w1thout forest (ha) o 049 1 873 Mechamzable area (ha) +0 378 8 058 Tractor ownersh1p (dummy) +123 985 1 720 D1stance from Puerto López (km) 1 103 1 838 Techmcal ass1stance (dummy) +126671 1 563 Farm adm.n1strator (dummy) +135 128 1 729 Res1dence at farm (days) o 588 1 955 Educat1on (years) 96 109 2 058 Land t1tle (dummy) + 175 608 2 027 R2 = o 6695 F = 15 58 DW = 2 32 1 R2 1 Coeff1c1ent of Mult1ple Correlat1on Ad¡usted Coeff1c1ent of Autocorrelat1on Durb.n Watson F stat1st1c DW F level of Coeff1c1ent SIQnlf of Correlat1on o 427 o 016 o 11 o 066 o 42 <0 001 o 69 o 090 o 37 o 071 008 o 123 o 37 o 089 o 42 o 055 016 0044 o 15 o 047 o 17 The results of th1s study are cons1stent w1th the theoret1cal pred1ct1on that the pre cond1t1ons of 1ncreas.ng land values and 1mprov.ng .nfrastructure have to be met befare technolog1es of .ncreas.ng cap1tal and management .ntens1ty are adoptad Grass legume pasturas and ley farm.ng systems appear to requ1re a h1gher level of .nvestment 1n capital and management than farmers 1n the study area are wllhng to .ncur at th1s t1me Even the recommendat1ons for pure grass pasturas appear to be too .ntens1ve at current land values and at current levels of mfrastructure In fact the ma.n dnv.ng force beh.nd adopt1on may have been natural .ncreases 1n herd s1ze There are .nd1cat1ons however that ma¡or 1mprovements 1n .ntrastructure are hkely 1n the study area and th1s 1s hkely to lead to the adopt10n of more .ntens1f1ed technolog1es ft woufd be very 1mportant therefore to make a deta1led study of 74 farmers ob]ect1ves resources and pract1ces to gUide the mod1flcat1on of the technology to better meet farmers needs The pay off may be part1cularly h1gh m v1ew of the technology s apparent abi11ty to sequester carbon at a t1me when mternat1onal markets m carbon sequestrat1on serv1ces appear to be emergmg lt should be p01nted out however that th1s prognosis may be undermmed 1f guernlla act1v1ty and the drug trade contmues unchecked A larga part of the savanna m Latm Amanea IS hkely to remam w1th poorer mfrastructure than the study area m the med1um term Charactenzat1on of the heterogene1ty w1thm the savanna IS reqUired to more accurately target ex1stmg technolog1es and develop a prof1le for appropnate technolog1cal alternat1ves An adopt1on study m the most mtens1f1ed areas of the savanna such as the center south of the Braz1han savanna wh1ch on an ex ante bas1s appears to be the most appropnate target area for the technology IS clearly md1cated A land use plan for the savanna based on both b1o phys1cal and soc1oeconomlc cons1derat1ons IS needed to 1dent1fy whether agncultural development 1s des1rable m remete areas of the savanna or whether these areas should be set as1de for conservat1on purposes In particular the 1mportance of preservmg nat1ve savanna for b1od1vers1ty purposes needs to be emphas1zed The fact that the most prom1smg legume currently ava1lable A pmtot 1s a nat1ve spec1es from the Braz1han savanna 1s test1mony to the 1mportance of preservmg b10d1vers1ty A pmtot s pers1stence may also 1mply that nat1ve spec1es may be more robust under farmer management For those less 1ntens1f1ed areas where agncultural development 1s part of the strategy technolog1es wh1ch save labor cap1tal and management but use land m abundance may be more appropnate In particular legumes wh1ch were developed w1th areas of poor mfrastructure m mmd where access to fert1hzer may be d1ff1cult appear to reqUire ma1or decreases m the1r capital and management reqUirements 1f they are to fulflll the role for wh1ch they were ongmally mtended lt may also be worthwh1le mvest1gatmg whether m areas of peor mfrastructure where stockmg ratas are hkely to be low sustamable product1on can be ach1eved through the development of technolog1es wh1ch combme the management of nat1ve grasses w1th pure grass plantad pasturas Th1s would bUIId on farmers current pract1ce thus mcreasmg the chances of acceptab1hty Survey results 1mply that natural mcreases m herd s1ze may be propelhng adopt1on of pure grass pasturas to some extent Adaptat1on of the technology to the reqUirements of these unmtens1f1ed farms 1s reqUired to prevent them from takmg the over grazmg route In th1s context 1t should be notad that the current technology for recuperat1on of degradad lands the crop pastura technology 1s too mtens1ve for these areas An alternativa technology wh1ch bullds on the farmers strategy of ut1hzmg low lymg areas and plantmg small areas of pure grass pasturas may be more adoptable A detallad understandmg of current pract1ces and the objectlves of the d1fferent categones of farmers m the savanna mcludmg small scale farmers 1s clearly md1cated lt should also be p01nted out that analys1s of the econom1cs of the new technolog1es and evaluat1on of the1r ecolog1cal 1mpact has been based so far on data from controlled expenments Vahdat1on of these results under farmers pract1ces 1s reqUired Also very httle attent1on so far has been g1ven to offs1te 1mpacts such as the effect of 1ncreasmg demand for fencmg matenals on the preservat1on of gallery forests L1ve fencmg usmg forage traes may be an opt1on worth mvest1gatmg 75 Subpro¡act TL 03 1 vagatation and acology of tha sarrama nat1va pasturas of tha aastam plams of Colombia (Llanos) G R1ppstem (CIA T/CIRAD EMVTJ J K Broekhw¡sen (UWJ A G E Peters (UW) G Escobar (TL) 1 INTRODUCTION The Eastern Pla10s savanna of Colombia (Llanos Onantalas) occup1es almost 17% of Colombia 1s surface area (Figure 6 Tables 13 and 141 Naarly 80% of thts area conststs of a grass vegetatton where extenstva cattle ra1s10g 1s betng pract1ced The rest of the area 1s covered by forest or 10 use for agnculture Around 40% of the Llanos of Colombta are covered by Serranfa ( Serranfa and Altlllanura ondulada 1 The people of the Llanos have gtven th1s name to the htlly savanna landscape whtch 1s s1tuated 1n the South of the Llanos Although 1ts surface area 1s very extenstve the Serranfa 1s not mtens1vely usad because of 1ts d1ff1cult accesstbthty and the very poor quahty of the maJonty of 1ts solls Table 13 Colombtan Eastern Plam savannas (Llanos) WELL DRAINED SAVANNA Flat Alt1llanura Ondulated Alt11lanura and Serranfa (sloppy) Fluvtal banks Food htlls BAD DRAINED SAVANNA Old alluv1um TOTAL LLANOS TOTAL COLOMBIA Source Vera and Seré 1985 1 000 hectares 1 % 3 438 21 2 6 385 39 3 1 245 77 925 57 4 934 30 4 16 927 100 o 103 900 Table 14 Rehef repart1t1on 10 Serranfa and Alttllanura ondulated (7 farms) 1 Total Serranfa Ondulated savanna Hectares 25200 15600 9300 1 % 100 62 37 The Llanos Onentales 1s sparsely populated The most tmportant acttv1t1es wtthm the area are cattle ranch10g and subs1stence agnculture Through the development of an mfrastructure commerCJal agnculture w11l be more and more practtced 10 the Llanos Onentales Thts development 1s occumng espec1ally m the flat savannas Altlllanura plana whtch are more access1ble and have a h1gher soll fert1hty Cattle ranch10g well be torced to move to areas whtch are less su1table for agnculture such as the Serranfa 76 ...¡ ...¡ • o "" Lllnd systsms 201 202 203 204 205 206 271 1 1 1 1 1 1 BOGOTA 1 o Flllt hlgh p/11/n Flllt hlgh p1sln ~ ~- o Undulatmg h1ghp/llln ----------~~ -; _-:: --9==----~ u~~ """"' ......... ..,..,.,.%..,_ ),,~~~~~''"""' ~ Sarr11niB HíDs <-, ;;._,¿~~- - ~ ---::.:.~ =:r.:..=.," ~;~ -~~t¿ -~~--- ~-"L / --- -- _,_..__/"'-~~~~ /' -------- c.-v / - -------==--~- =----~-~ :::=-2 O> NUMBER OF SPECIES 36¡---------------, 30 26 20 ,6 ,o 6 [j Top ot tho hlll Slopo D y lovvlo.nd Hurnld lowla.ncl Flooclplaln TOPOGRAPHY • Numb spec1es [8J Numb poacea O Numb cyperacea ISJ Numb Jegummocea 8 Numb otb monocotyl 8 Numb o lb dJcotyl Figure 7 Vegetahon of serranta Topography and number of spec1es '00 BO " Cll) ea 111 ... 1:1 " t ~ 40 O> 20 ..... o AERIAL COVER ( % ) Top of thc hlll S lo pe Dry lowta.nd t--iumld lowla.,.-,d Floodploln TOPOGRAPHY 1% Poacea El% cyperacea 0% Legum1nocea ~% Other monocotyt!fh Other d1cotyl F1gure 8 Vegetahon of serrama Topography and relahve aenal cover by herbaceous plants • j O> O> 7000 esoo- eooo- 6600 - soco- 4600- 4000- 3600- 3000- 2600- 2000- sao- AERIAL PHYTOMASS ( Kg DM/Ha) CCJ GIRE EN CIJ DEAD -17''70 r- Q807" r- .;a•o• r- ........ rr ....... r ---7 .. =- ... ~ ... r- .a .. o"P ....-- o07"r0 ao:a• 11 e:aj. lí 1 000 ---{ ...... 600 o 1 e 30 so a7 go s4 ea ea 76 ao 270 aoo aoo 3es ==- 3ee- 3es Top DL HL Top FP DL DL &L HL DL HL Top DL Top FP DL DAYS OF REGROWTH SITE F1gure 9 Aer1al phytomass from d1fferent nahve pasture after burnmg 00 "' AERIAL PRODUCTION ( Kg DM/Ha/DAY ) 30 215 - ..... .. .. .. o .. 20- ry_s >x ... ~ .... ~ S ~ )< f'<)< X IX oe ~ ... t< )< v ~ r-',., )< X: ¡:;,; )< ?1 ¡;< S 2 ";.,; )< ~ !X" ["); '0- )< x' (':1 X X ?),_ < >< 1>< :x1 I<.AJJ 15 - 10- 15~ 2 S rx .. ~ S < .. ... 9 ao so ar so 164 se 1ae 176 eo 2ro aoo aoo ase Top DL HL TOP FP DL DL SL HL DL HL TOP TOP FP DAVS OF REGROWTH SITE Fagure 10 Aeraal productaon from dafferent nahve pasture after burnmg "' o % OF THE DRY MATTER o e e ... 2 o 1 1 :>, > :>, :>, 21 1 > > > > >f L > > > > ;t l> > > > > 1 D >, > > > 1 1 TC>P SL D L t-i L FL PL Ftgure ll Protem content m the pasture related to the dtfl'erent topographtc sttes o 14 012 o 1 o oe o oe o 04 "' - o 02 o % OF TBE DRY MATTER w. t-itLL TC>P S LC>F" E C> L 1-t L FL PL fZJ Senes 1 D Senes 2 } Fagure 12 P and Ca content m the pasture related to the ddTerent topographtc sates 3 9 Soll textura F1gure 13 shows that the sandy clay loarn structure dornmates at the top of the h1lls and slopes and m the dry lowlands In the hurn1d lowlands and the floodplam a clay loarn or a loarn structure dorn'lhates 3 1 O Chern1cal soll analysJs The percentage orgamc rnatenal (Figure 141 1s the lowest at the top of the h1lls (2 0%1 and the h1ghest m the hurn1d lowlands (12 2%) lt 1ncreases shghtly frorn the top of the h1lls to the dry lowlands lt shows a strong mcrease frorn the dry lowlands to the hurn1d lowlands and a shght decrease to the floodplam In F1gure 15 the C/N ratio frorn the d1fferent s1tes 1s g1ven lt shows that the C/N rat1os m the so1l are very h1gh even on the top and the slopes They fluctuate between 18 and 24 The Ca content 1s lowest at the top of the hllls (0 11 rnllheqUJvalent/1 00 gr s011) and h1ghest at the slopes (0 14 rnllheqUJvalent/1 00 gr soll) The concentrat1on of Mg and K follow the sarne pattern They are lowest at the top of the h1lls (0 04 rnllheqUJvalent/1 00 gr so1l and O 05 rnllheqUJvalent K/1 00 gr s011) and h1ghest m the hurn1d lowlands (0 00 rn1lhequ1valent/1 00 gr so1l and O 1 O rnllheqUJvalent K/1 00 gr soll) (F1gure 16) The Al concentrat10n 1s lowest at the top of the h1lls O 62 rnllheqUJvalent/1 00 gr so1l and IS mcreasmg to 3 42 rnllheqUJvalent/1 00 gr so1l for the hurn1d lowlands Afterwards 1t 1s decreasmg to 2 22 rnllheqUJvalent/1 00 gr so1l for the floodplams (Figure 171 4 CONCLUSIONS lt 1s poss1ble to conclude that there are d1fferences m the cornpos1t1on of the vegetat10n relatad to d1fferences m rehef soll type and hydrology to sorne extent Four d1fferent plant cornrnun1t1es were 1dent1f1ed (at th1s stage of the work) wh1ch show a relat1on w1th the topography Two cornrnumty types are 1dent1f1ed for the top of the h1lls and slopes For the f1rst cornrnumty f1ve charactenst1c spec1es could be 1dent1f1ed of wh1ch the grasses Trachypogon vestltus and Paspalum pectmatum are the rnost typ1cal In the second cornrnumty f1ve charactenst1c spec1es are also 1dent1f1ed of wh1ch the grass Trachypogon p/umosus 1s the rnost typ1cal Two cornrnumty types are 1dent1f1ed for the hurn1d lowlands and floodplams There are s1x charactenst1c spec1es 1dent1f1ed for th1s cornrnumty of wh1ch the grass Andropogon sel/oanus can be class1f1ed as rnost typ1cal For the Jast cornrnumty 1 3 charactenst1c spec1es could be 1dent1f1ed of wh1ch the grasses Andropogon b1corms and Sch1zachyrwm brewfolwm could be class1f1ed as rnost typ1cal D1fferences 1n aenal cover relat1ve aenal cover and average nurnber of spec1es between the topograph1c umts are to sorne extent due to d1fferences 1n so1l propert1es and s01l textura lt 1s shown that the nutnent supphes m the hurn1d lowlands and floodplams are sornewhat better than at the top of the h1lls slopes or m the dry lowlands lt seerns that the nutnent supphes are best 1n the hurn1d lowlands except for Ca and P Th1s last elernent rnay be f1xed by alurnm1urn whose percentage 1s h1ghest m the hurn1d lowlands The great nurnber of d1fferent spec1es m the hurn1d lowlands can to sorne extent be explamed by a h1gh rno1sture 92 percentage a h1gher percentage of orgamc matenal and a h1gher nutnent supply The data of th1s research show a great d1fference 1n aenal b1omass among the topograph1c s1tes A very low percentage of green matter after 30 days of regrowth was observad The aenal product1on 1s much h1gher m the lowland for short regrowth penods For longar penods 1t IS d1ff1cult to conclude because of more mtens1ve grazmg 1n the lowlands F1gures 9 and 1 O not only show us a very h1gh product1on for the hum1d lowland but also for the top of sorne hllls The d1fferences 1n forage value among the topograph1c umts are very small The data of the chem1cal plant analyses show a low nutnt1ve value for the Serranla pasturas The average percentage P IS even lower than the cnt1cal value necessary for good ammal performance The f1gures for mtrogen content are not much better When the nutnt1ve value of the d1fferent topograph1c s1tes are comparad 1t 1s shown that the s1tuat1on 1s shghtly better m the floodplams However 1t IS also 1mposs1ble to draw conclus1ons from these data because d1fferences m age and stage of the plant matenal were neglected when the chem1cal analyses were made To come to a further d1fferent1at1on and class1f1cat1on of the Serranla vegetat1on and ecology 1t IS necessary to cont1nue research More samples should be taken and a system has to be made m th1s sample recordmg 93 75% TOP S LO PE D L H L FLOOD ll!l88i SAND Oil] LOAM 821 CLAV F1gure 13 So11 texture related to d1fferent topograph1c s1tes 94 1 o " N N " ' ' 11 ft ~ ft ~ ~)< ~)< 1><)< J Q. J ~ J I J D IU Q. o J U! J J I IU I ~ 11. o ~ Q. ~ o O ID ID 't N O ID ID 't N O~ N r r r r r 95 - - = P e:> F t--I 1 L L SL = L 1--i L FL PL Oca 0Mg DK F1gure 16 Concentratlon of Ca, Mq and K m tbe sod related to tbe d1fferent topograpb1c s1tes ( Meq 1 lOOg soll ) .. 3 S 3 2S 2 CD a> , S , os o ALUMINIUM ( MEQ 1 100 SOIL ) • "x t---tiLL TC:>P SL C> L 1--i L FL PL TOPOGRAPHY Fagure 17 Alummaum concentrataon 10 the soal related to the dafferent topographac sates ( Meq 1 100 soal ) ANNEX 1 PLANT FAMILIES ANO SPECIES FROM THE SERRANIA NATIVE PASTURES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 FAMILIES ACANTHACEAE AMARANTHACEAE AMARYLLIDACEAE ANNONACEAE ARACEAE ARISTOLOCHIACEAE CAESALPINIACEAE COMPOSITAE CONVOLVULACEAE CYPERACEAE 36 DILLENIACEAE 37 ERISTOLACEAE 38 EUPHORBIACEAE 39 40 41 42 43 FA8ACEAE 44 45 46 47 48 49 50 SPECIES Rue/1 a gem n flora Ahe n nthera sp Curcul go scorzoneraefo!Ja Xylopa sp Calad um macrotl/es Ansto/och a nummulanfo/¡a Chamaecnsta sp Chamaecr sta cultr tola Chamaecnsta desvaux11 Chamaecnsts d sphylla Chamaecnsta h1sp dula Chamaec sra kunth ana Chamaec Sta rotund folla Catea colomb ana Compos1te sp Eupstonum sp Eupstor um amygdal num lchthyothere term nal1s St lpnopappus p ttJen lpomea f stulosa Merrem1a aterens1s 7 Bulbosryl s paradoxa Cyperus bre fol us Cyperus flavecens Cyperus flavus Dchomen clara Rhynchospora ba bata Rhynchospora cont ms Rhynchospo a corymbosa Rhynchospo a globosa Rhynchospora podeosperma Rhynchospora pubera Rhyn hospors subplumosa Scler a d stans Curatella amencana Eryth oxylum sp Croton tr mtat s Euphorbta commun s Eupho b a htrta Phy/lanthus sp Phyllanthus n u Centrosema angust foltum C nt osema enosum Cl to sp Clttona gUJanensts Crotalan p e o tala n t dula Desmod um barbstum Enosema sp 99 51 62 63 64 65 66 57 68 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 FLACVORTIACEAE GENTIANACEAE GUTTIFERAE HAEMODORACEAE IRIDACEAE LABIATAE LYTHRACEAE MALPIGHIACEAE MALVACEAE Enosema cr n tum Enosema rufum Enosems s mpl!ctfol um Gafact1a glaucescens Gsfsctta JUSs aeana fndtgofe a lespedez o des Macropt 1 um monophyllum Styfosanthes gUJanens s Vtgna f near s Zona sp Caseana syf estns Cssear a ulmtfola lrlbach a afata subsp alata Schuftesta sp V sm a des/bata Schtekta o nocens s Cypel/a 1 neans HyptJs sp HyptJs atrorubens HyptJs brach ata HyptJs capttata HyptJs conferta HyptJs dtlatata HyptJs vertletllata Cuphea calophylla Byrson ma sp By son ma crass fofa Byrson ma ve base falta Ps ona sp Peftaea spec osa 81 S da acuta 82 S da 1 mfolla 83 MELASTOMA T ACEAE C/ dem a ubra 84 Desmoce/lts llosa 85 Mtcon a sp 86 Mtcon a rubtgmoss 87 Mtcon a rufescens 88 Pterogastra mayor 89 Tococa gUJanens s 90 MENISPERMEACEAE C ss mpe/os ovalfol a 91 MIMOSACEAE 92 MYRTHACEAE 93 94 95 96 OCHNACEAE 97 ONAGRACEAE 98 PALMAE 99 PASSIFLORACEAE 100 POACEAE 101 POACEAE 102 103 Mmosapudca Myrc a gwanen s Pstdum sp Ps d um ma bense Ps dum salut Sauv ges a e ecta Ludw g a decu ens Maunta m no Pass flora foet da 7 And opogon sp And opogon b co n s Andropogon gayanus 104 Andropogon hypogynus 141 Pssps/um psrVIflorum 105 Andropogon leucostachyus 142 Ps&fJB/um pectmstum 106 Andropogon sel/osnus 143 Psspslum p/¡catulum 107 Andropogon vtrgstus 144 Pas/)11/um stellatum 108 Ans&ds cspJI/sces 145 SBCCIOiepiS myuros 109 AnstJds npans 146 SchJzschynum sp 110 Ansuds tern1pes 147 Schtzschynum brev¡fo/Jum 148 SchJzschynum htrtlflorum 111 AnstJda tJncts 149 Setsns gemculsts 112 Axonopus sp 150 Thrasys petrosa 113 Axonopus affints 114 Axonopus snceps 151 Trachypogon sp 115 Axonopus sureus 152 Trachypogon plumosus 116 Axonopus chrysob/Bp/IBns 153 Trschypogon vesfltus 117 Axonopus purpus11 154 POLYGALACEAE Poly¡¡s/a psmcu/ata 118 ChusquSB pmtlobs 155 RUBIACEAE Barrene cspitsts 119 Coelorsch1s rsmosa 156 Decbeux1s frutJcoss 120 Cten/UITI plamfohum 157 Ssbltflll co/ombJsns 158 Sablees vtlloss 121 Dlgltllns n66Sisns 159 S1panes prstenSJs 122 Elyonurus csnd1dus 160 Spermscoce sp 123 Ersgrostis msypurensis 124 Gymnopogon fastigJstus 161 Spermscocs cspttatB 125 Gymnopogon fo/¡osus 162 SCROPHULARIACEAE Buchnera pustlls 126 Hypogyn/um VJrgstum 163 l..mdemJB dlffuss 127 Leptocoryph1um /snstum 164 SOLANACEAE Solsnum sp 128 Mesosstum Ptttlen 165 STERCUUACEAE &ttnena mo/11s 129 Otschynum vers coiDI' 166 MeJoch1a colombiana 130 Psmcum sp 167 Meloclua af pyranvdats 168 Meloch1a VI/losa 131 Psmcum ps!VIfo/Jum 169 TURNERACEAE Tumers af punvlea 132 Psntcum stanodes 170 Tumers u/m¡fol a 133 Pssps/um sp 134 Psspslum BCUtnlnatum 171 XIRIDACEAE Xyns sp 135 Pasps/um csrinatum 172 Xyns caro/ mana 136 Paspalum c/svuhferum 173 Xyns caro/imana var mayor 137 Paspslum contrsctu 138 Pasps/um convexum 139 Paspalum mmus 140 Pssps/um multJcsuls 100 ANNEX2 SYNOPTIC TABLE ( BRAUN - BLANOUET METHOD) No SAMPLE (RELEVE) 1 1 1 1 1 3 1 2 2 2 2 2 2 2 1 2 1 1 3 2 1 2 1 o 2 1 5 7 4 o 4 o 2 5 5 3 1 7 9 8 7 9 4 3 6 1 6 8 ~ 6 3 e 9 TOPOGRAPHIC SITE S t t t d S d ddd S d t S t t t S d h f h d h f h h d h f f 1 D D D 1 1 1 1 1 1 1 ID 1 D D D 1 1 ID 1 1 ID 1 1 1 1 D P - - - - - - - -- - - -- - - - - - -- - -- - -- --- --GEOGRAPHIC SITE p p ppp pp p p S S S p S sppp p S S ppp p p p p p p p 99 9 999 9 19 1 1 19 1 1 19 19 1 1 99 1 19 9 9 9 9 1 2 2 2 1 2 2 1 o 2 t t t 1 t t o 1 o 1 t t 2 2 o o 2 1 1 1 1 o SPECIES ABUNDANCY DOMINANCY No OP- SL DRY LOWLAND ILOWLANDSI 11 ~hamah1s 3 3 1 ---------------------------- 33 Rhync pub 3 4 -- --------------------------- 50 Enos sp ---- 1 1 ------------------------- 153 ~rach ves 4 3 4 4 5 54 -3 2 -------4 3 ---4 1 ---1 --- 142 1Paspa pec _-:;:¡- 1 !'1 3 3 3 4 4 1 3 1 -------3 ----- 148 ¡sch1z h1r 2 3 -3 3 --2 ---------4 ---1 ---2 -3 --- 120~tem pla -1 ----4 -----3 ----- ------------- 137 Paspa cot 3 ----4 ---1 -------- -----3 ------- 160~erm sp 1 ----1 ---------1 -- ----------1 -- 21 erre ate -1 ---3 ----1 - 1 ----- ------------- 170 ume ul -----2 1 ----1 1 ----- ----2 -----1 -- 46Ehtogul 1 1 --3 ---1 - 1 1 -----1 ------------- 55~alaCJUS 1 1 1 -1 1 1 -3 1 1 -3 3 1 -- 1 2 ---1 -------- 23 Bulbo par 4 3 4 --3 ----4 2 3 3 4 -4 -------------- 76 ¡ayrso ver 1 1 1 --1 -- 1 ---1 1 1 - 1 ------1 ----1 -- 154 1Polyg pan 1 1 1 ------1 1 1 ---1 - 1 3 ------------,.r ... -- 1 2 1 -2 1 ---1 - 1 - 1 - 1 -------------18 chth ter -----1 - -- 1 1 - 1 ----- ------------- 126 esos p1t -4 3 4 --- 6 4 5 4 4 5 4 4 ~ ----4 -4 ------ 152 rach plu 3 -----6 -3 5 4 5 5 4 5 4 -- --4 -4 4 --- 1 --- 31 1Rhync glo ---3 --- --3 --4 1 4 1 -4 ------------- 54ealacgla ------- --1 1 3 ------ ------------- 155 ~orre cap ------- 1 - 1 -1 ------ ------------- 100 ram1 -?? ------- --2 -2 --24 5 3 -5 4 -- 2 --3 ---- 77 Byrsocra ------- 1 1 1 -----------1 - -1 ------ 1¡Ruell gem -- 1 -3 -- -1 ---------3 --1 - -------- 127 ~epto lan 1 --4 3 5 3 --3 3 4 4 -----3 4 4 4 3 3 1 3 ----- 53 Enos s1m - - - - - - - - 1 1 - - - - - - - - - 1 2 3 3 - 1 ------ 27 P.1chr c11 - - - - - - - - - 1 - - - - - - - - 4 - - 1 4 --2 ----- 106 ~ndro sel - - - - - - - - - - - 2 - - - - - - - - 1 4 4 ------6 - 7 hama -sp - - - - - - - - - - - - - - - - - - 3 1 - - - -------- 67 ICOnruf -----1------------111 1 - -1 -- 1 --- 133 Paspa -sp - - - - - - - - - - - - - - - - - - 4 3 3 -- ------1 - 57Macromon -----------------1 1 1 1 ----1 - 1 2 -- 681"iypt1 - sp --------3 2 -------- -1 - 1 1 4 3 ------ 69 ~ypt1 atr -----1 -- 1 --------- 1 1 -- 1 -3 -2 ---- 101 ANN~~~~~~ ¡No S ~'--,---.--.,,---.,.--.-.,..,.,.,--.,.,-.,-----,,---.-..-.--..---....-.;..--------,-, 1 rOPOGRAPHIC SITE l GEOGRAPHIC SITE 1 14 col - - - - 3 - - - - - - - - - - - - - - - 3 1 - - 1 - - - - 169 pum - - - - - - - - 1 - - - - - - - - - 1 - - 1 - - - 1 - - - 72HyptlcOn ---3--1---------- 2321-133312 aa,Pteromay ---------1---1--- -11--1-1--1- 1 1 1586BbiCVIi ----------------- -111-11-432- 83 Cilde rub - - - - - - - - - - - - - - - - - - - - 1 - 1 1 2 - - - - - 104 ~dro hyp - - - - - - - - - - - - - - - - - - - - - 5 - - 1 - 3 - - 126 Hypog w - - - - - - - - - - - - - - - - - - - - 3 - 1 3 - - - - - - 132 Pan~c ste - - - - - - - - - - - - - - - - - - - - 1 - - - - - - 4 - 5 159 S1pan pra - - - - - - - - - - - - - - - - - - - - - 1 1 1 - - - - - 173~ynsc-m ----------------- ---1---1 75 Cupe cal - - - - - - - - - - - - - 1 - - - - - - 163 Linde d1f - - - - - - 2 - - - - - - - - - - - 3 - 1 - 147 SChiZ bre - - - - - - - - - - - - - - - - - - - - - - 1 02 Andro b1c - - - - - - - - - - - - - - - - - - - - - - 165 181ttn mol - - - - - - - - - - - - - - - - - - - - - - - - - 3 3 301Rhynccor ---------------------- --1----3 ~~ ~~~;~ ~~r :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: :::: ~ ~ - 2 - 84Pesmovtl ---------------------- ---4-1-- 91 Mimos pud - - - - - - - - - - - - - - - - - - - - - - - - 1 2 - - - - 971ludwl dec - - - - - - - - - - - - - - - - - - - - - - - - - 1 - 3 - - 124Gymnofas---------------------- 2--4---- 96Sauvaere -------------1---------3-12-2-- 25 ICyper fin - - - - - - - 2 - - - - - - - - - - - - - - - - 1 1 - - - - - 1 07 IAndro Vlf - - - - - - - - - - - - - 1 - - - - - - - - - - 4 - - 3 - - - 129 btach ver - - - - - - - - 1 - 2 - - - - - - - - - - - - - 4 4 - 3 - - -~ 52Enosruf ---------1-1-----1-----31------ 8 Chama cul 1 - - - 1 1 - - - - - - 3 - - - - 1 - - - - - - 1 1 - - - - - 42 Phyll nlf 1 - - 3 - - 1 - 1 - - - 2 - - - - - - - - - - - 1 - - - 1 - 1 1 49 Oesmo bar - - - 2 - 1 - - - - - - - - - 1 - - - - - - - 1 1 - - - - - - 172 IXyns car 1 - - - - - - - - - - - - - - - - - - - - - - 3 1 - - - - - 31 1 08 Anst cap - - - - - - 2 - - - - - - - - - - - - - - - - - - - 3 5 - - -1 149 Setar gen - 3 1 - - - - - - - - - - - - - - - - - - - - - - 3 - - - - -1 150 [fhras pet - - - - 3 2 - - - - - - - - - - - - - - - - - 1 - - - - - - - 59 i"•gna iln - - - - 1 - - - - - - - - - - - - - 1 - - 1 - - - - - - - - - 61 Casea syl - - - - 1 - - - - - - - - - - - - - 1 1 - - - - - - - - - - - 105 Andro leu 4 6 4 5 4 4 4 5 5 - 1 - - - - - 3 - 4 4 3 4 - - 4 4 - 1 5 - -¡ 1351Paspa car 5 3 5 5 5 - - - - 4 2 5 2 4 4 6 5 4 3 - 4 4 - 4 4 - - - - - - 291Rhynccon13--31-4--2-5431-4-4---411 ---35 101 ¡Andro -sp - - - 3 - 3 3 - 1 - - - 3 - - 2 5 - - - - 1 - - - 4 5 - - - -\ 51 IEnos en - 1 - - 1 1 - 1 - 1 - - 1 - - - - - - - - 1 1 - - - - 1 - - -¡ 70 Hypt1 bra - - - - 1 1 - - 1 1 1 - 1 - - - - - - 1 - 1 - - - - - 2 1 - -¡ 73 Hypt1 dli - - - - 4 - 1 1 - - - 1 - - - - - 1 1 - - - - - - - 3 3 1 - - 117 ~ono pur - - - 2 - - - 3 5 - - 2 - - - - - - - 4 - - 3 - 3 - - 5 - 3 -1 102 ANNEX 2 ( CONTINUED )_ No -SAMPLETRELEVE) 1 1 1 1 1 3 1 2 2 2 2 2 2 2 1 2 1 1 3 2 1 21 1 o 2 1 5 7 4 o 4 o 2 5 5 3 1 7 9 8 7 9 4 3 6 1 6 8 2 6 3 8 9 TOPOGRAPHIC SITE S t t t d S d d d d S d t S t t t S d h f h d h f h h d h f f 1 p p p 1 1 1 1 1 1 1 IP 1 p p p 1 1 lp 1 1 lp 1 1 1 1 p p - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 1 GEOGRAPHIC SITE p p p p p p p p p S S S p S S p p p p S S p p p p p p p p p p g g g g g g g 1 g 1 1 1 g 1 1 1 g 1 g 1 1 g g 1 1 g g g g g 1 2 2 2 1 2 2 1 o 2 t t t 1 t t o 1 o 1 t t 2 2 o o 2 1 1 1 1 o ~--SPECIES ABUNDANlJY DOMINANCY No 1 1 62~asea ulm - -- 1 - 1 2 1 - 1 - -- 1 1 - 161 perm cap - 1 -- 1 -----1 ------1 - 1 --- - 1 1 2 ---- 22pyper -?? - -----4 ---3 -- -------3 -- -----4 4 - 114~onoanc - --- - 1 -- 1 -------------1 4 - 1 ---- - 112 ono -sp -- --3 -- -----------5 ---- --2 --- 1 - 17¡Eupat amy - --- ---1 - 1 --------1 ---- -- 1 --2 -- 138 1Paspa cov ------4 ------4 3 -----5 -- - ----2 -- 631rlba a-a -- -----------1 ------1 -- ----1 --- 136 ~aspa cla ---------------4 --- 1 ----3 ------ 47 1 roto -sp -----2 -------- ---------2 --3 --- - 95¡Psldl sal -- ----1 1 --------------------2 -- 2 fA.!ter -sp ----1 --- ---------- - --- --------- 64Schul -sp ------1 ----- ------ --- ---------- 67 Cypel hn -- -2 ----------------- -- - ------- 93 Ps1d1 -sp -----1 ------------------------l 94 Ps1d1 mar ----1 -------------------------- 11 o Anst ter ;::~:::::::::::::::::::::::::::! 115 Axono aur 141 Paspa par 4------------------------------ 28 Rhync bar 3---------4--4----------------- ----- -----1 --- -------- - --------3 purcu seo --- --- - ---- ---- ----- ------- --12 Chama kun 1 15 bompo - sp - - - - - - - - - 1 34 Rhync sub 431Centr ang 45phto -sp 48 prota mt 58Stylo QUI 60 !Zorm - sp 80 ,Pelta spe 90E¡ssa ova 121 D1g1t nee 145 Sacc1 myu 146 Sch1z - sp 157 ~ab1c col 171,.....yns -sp 92 Myrc1 gul 19 St1lnp p1t 76 Byrso -sp 1 31 Pan1c par 166 Meloccol 6 !Anst num 44Centrven 71 Hypt1 cap 81 S1daacut 82Sidal 101 --------------4---------------- -------2----------------------- ---------------1-1------------- -------1 ----------------------- -------1 ----------------------- -------1 ----------------------- ---------1 --------------------- ---------1 1 -------------------- ----------1 -------------------- -------------1----------------- ---------31-------------------- --------1 ---------------------- ---------44-------------------- - - - - - - - - - 1 1 - - - - - - - - - 1 - - - - - - - - - _l --------1---------3------------ --------------1---1------------ -------------1----2------------ --------1------------1--------- ::::::::::::::::::::~~:::::::::¡ ---------------------1 ---------1 ::::::::::::::::::!:::~::::::::¡ 103 Subpro¡ect TL 03 2 nat1ve pastura management e sequantial bummg and rotet1onal grazmg of the nat1ve pasturas of the flat pla1n IAitdlanura) of tha Llanos G R1ppstem (CIA T/CIRAD EMVTJ A G E Peters (U Wagemngen) J K Broekhw¡sen (U Wagenmgen) G Escobar (TLJ 1 INTRODUCTION The Llanos of Colombia cons1sts of an area of 17 mllhon hectaras coverad w1th tropical savanna Extens1ve cattla product1on 1s pract1cad m th1s area because of lack of nutnt1ve forage dunng the dry season of the year The fluctuat1ons m quant1ty of nutnt1ve forage between wet and dry season cause fluctuat1ons m hvewe1ght gam for ammals grazmg here resultmg m very low product1on f1gures of 28 kg/an/yaar at a stockmg rate of O 2 an/ha W1th the lack of mfrastructure 1n th1s area the 1mportance of product1on 1mprovement by low externa! mputs becomes clear A poss1b1hty for th1s hes m ad¡usted management systems Factors Wlthm th1s are burmng and graz1ng A lot of mvest1gat1on has been done on burmng systems and grazmg systems md1catmg degradat1on of the pastura or small mcreases m ammal product1on under these treatments Therefore a management systems of sequent1al burmng resultmg 1n regrowth every month combmed w1th rotat1onal grazmg was proposed Th1s systems was stud1ed for 1ts effect on the vegetat1on dynam1c and the development of we1ght of grazmg ammals at two stockmg rate dunng three months at Canmagua the research Center of CIA T /CORPOICA 2 MATERIALS ANO METHODS Befare burmng the vegetat1on of the plots was one year old cons1stmg of annuals and perenmals Soll was totally covered by a dry b1omass dommated by the dormant perenmals Trachypogon vestltus and Axonopus purpusu These would start the1r germmat1on at the start of the wet season or after a f1re at the end of the dry season Annuals were present mamly as seeds m the seed bank because of unfavorable env1ronmental cond1t1ons for growth Soma of these would germmate e1ther when growth cond1t1ons 1mprove (e g mcrease m water ava1lab1hty decrease m compet1t1on for growth factors) or by photopenodiCal control after occurrence of the longest days 1n June July For the study of the vegetat10n dynam1cs and evolut1on of ammal we1ght proposed management system there are 8 research quest1ons to mvest1gate concerned w1th the effect of 1 Penod of burmng on the botan1cal compos1t10n of the vegetat10n 2 Stockmg rata (SR) on the botamcal compos1t1on of the vegetat1on 3 Penod of burmng on the product1on of the vegetat1on 4 Stockmg rate on the product1on of the vegetat1on 5 Penod of burmng on the development of the we1ght of grazmg ammals 6 Stockmg rate on the development of the we1ght of grazmg ammals 7 Penod of burmng on the d1et preference of the ammal 8 Stockmg rate on the d1et preference of the ammals gam m the These are The research was performed on nat1ve savanna on clay so1l w1th pH (H20) of 4 9 and orgamc matter content of 3 1 % 1n the upper 20 cm of the soll The amounts of exchangeable cat1ons calc1um magnes1um potass1um and alum1mum of th1s so1l are O 19 O 073 O 064 and O 2 Meq/100 g respect1vely Alum1mum saturat1on 1s very h1gh 82% Dueto th1s h1gh 104 alum1mum avadab1hty the avadab1hty of phosphorus was low O 9 ppm (CIAT 1990 and Tergas 1986) Med1a annual ra1nfall 1s 2200 mm and average temperatura 1s 26 degrees Cels1us (Table 18) Tabla 18 Chmate data of February March and Apnl 1993 1n Canmagua Temoerature I°C Month Average Mm1mum Max1mum Prec1p1tat1on (mm) February 28 1 22 1 33 6 49 6 March 26 2 22 2 31 6 251 o Apnl 26 3 23 o 31 1 233 2 Source Meteorolog1cal Department of Canmagua Research Stat1on The botamcal compos1t1on of the nat1ve savanna 1n Can magua cons1sts for 39 3% of grasses 13 6% legumes and 9 0% Cyperaceae Other spec1es occurnng m th1s savanna are m decreasmg order from 5 5% to 2 2% Lab1atae Rub1aceae Melastomataceae Euphorbaaceae and Composatae Next to these there are other famahes m very low frequenc1es The tnal const1tutes of two plots one wath a medaum stock1ng rete of O 25 ammal/ha and one wath a hagh stock1ng rete of O 5 ammallha Th1s 1s explaaned an Tabla 19 For thas tnal 8 zebu cows are usad (4 an each stock1ng rete) Tabla 19 Des1gn of the tnal LSR (A) HSR (A) LSR (8) HRS (8) sub plot 2 sub plot 3 1 burnt 01 02 93 sub plot 12 sub plot 1 sub plot 10 sub plot 11 sub plot 8 sub plot 9 sub plot 6 sub plot 7 sub plot 4 sub plot 5 burnt 01 03 93 burnt 01 04 93 8 hectares 4 hectares Low stockang rate ILSRl Hagh stockang rata IHSR) LSR HSR = 4 anlmals grazang on 2 4 hectares = O 25 anamal/ha = 4 anamals grazang on 2 8 hectares = O 50 anamal/ha 105 Changes m botamcal composJtJon and the b1omass productJon after d1fferent treatments of burnmg and under d1fferent grazmg mtensJtJes were JnvestJgated lnd1cators for quahty of regrowth and spec1es are the preference of grazmg ammals for certam specJes or for certam ages of regrowth We1ght development was dehm1tated by monthly we1ghmg of the grazmg ammals commg dJrectly from the f1eld The mvest•gat•on were carned out dunng three consecut1ve months 10 the trans1t1on penad from the dry to the wet season Th1s means that f1rst measurements were made 10 February wh1ch was the end of the dry season because the ramy season started very early th1s year The second measurements were of March and the last enes of Apnl both m the mJtJal penad of the wet season The treatments of burnmg and grazmg were carned out at the same t1me m the 2 plots (HSR and LSR) The plots were dJVJded mto two t1mes 6 subplots (A and B) The treatment cons1sted of burnmg ene sub plot every f1rst day of the month After a regrowth penad of 15 days grazmg started under both h1gh stockmg rate (HSR) and low stockmg rate (LSR) m the part of the plot where the subplot w1th regrowth was sJtuated After four weeks of grazmg the ammals were moved to the other 6 sub plots w1th a new regrowth subplot Th1s 1s shown 10 Table 19 The vanables to measure after the treatments are The composJtJon and dynam1cs of the yegetatJon In each sub plot permanent transects of 20 meters were made on wh1ch the occurnng spec1es at every 20 cm were recordad Th1s was done after two weeks of regrowth (w1thout grazmg) and after s1x weeks of regrowth (four weeks grazmg) Transects were also performed at every 1 O cm on the d•agonals of f1ve protected squares of 1 x 1 meter 10 every subplot Standmg dry phytomass of the pasture befare byrnmg Th1s was measured by cuttmg the vegetat1on of f1ve squares of 1 x 1 meter befare the burnmg at a 5 cm he1ght The cut phytomass was dned and we1ghted ProdyctJon of regrowth after burmng w1th and w1thpyt grazmg Th1s was measured by regular cuttmgs of regrowth standmg b1omass at 5 cm he1ght of f1ve squares of 1 x 1 meter after two weeks regrowth (w1thout grazmg) and after SIX weeks regrowth (With four weeks graz1ng) Also standmg phytomass after s1x weeks of protected regrowth These protected plots were s1tuated w1thm the sub plots so 10 the past these were e1ther grazed under h1gh stock•ng rate or low stockmg rate The preference of graz1ng ammals for syb plots d¡ffenng m regrowth stage and grazmg ¡ntens¡ty Th1s was measured by observatJon of preference of grazmg ammals for sub plots 10 a certam regrowth stage The ammals were observad from a h1dmg place w1thout d1sturbmg the behav1or of the an1mals These observatJons took place every 14 days for 2 x 12 hours F1rst of these observat1ons were made m March when the ammals started grazmg the plot 3 RESULTS The mfluence of the treatments on the vegetat1on are shown 10 the next graphs and tables S1nce 1t was not poss!ble to analyze the data m a stat•st1cal way Jt 1s assumed that a d1fference of 2% has sorne s1gmfJcance Slmllarly changas m frequency percentage greater than 5% are taken as JOdJcatJve of 1mportant changes m the botamcal composJtJon 106 3 1 Vegetation Dynamics To g1ve more convemently arranged data these were reduced towards spec1es w1th a frequency percentage h1gher than 1 O% at any regrowth stage The spec1es apart from these are callad others 3 1 1 lnfluence of burnlng The effect of the t1me of burmng on the botamcal compos1t1on was shown by companng the regrowth vegetat1on of plots d1ffenng 1n date of burmng Wlth the1r former vegetat1on Thls was calculated m Tabie 20 These data are presentad more clearly to fac1htate companson of the effect of d1fferent dates of burmng m Table 21 and F1gures 18 to 22 Table 20 The 1nfluence of burnmg on the frequency 1%1 of spec1es after s1x weeks of protected regrowth 1 BURNING Spec1es Februarv March Aonl Stockma rate Nr N ame H1gh Low H1gh Low H1gh Low 1 Andropogon blcOmls 1 4 1 o 14 o 00 07 1 8 21 Hypt1s spp 1 6 07 1 5 42 15 4 10 5 24 Andropogon 11 8 50 70 14 7 1 7 26 29 leucostachyus 11 o 10 9 12 8 33 76 o 1 30 Otachynum vers/co/or 05 12 7 85 22 1 93 22 9 Paspalum pectmatum 33 Axonopus purpusu 95 10 1 31 6 04 65 00 37 Andropogon selloanus 44 20 3 87 85 58 1 7 43 Trachypogon vest1tus 45 8 22 3 17 6 27 3 9 1 48 5 00 Other spec1es 04 86 4 1 64 33 20 3 46 Bare so1l 28 o 25 8 00 4 1 00 00 3 1 2 lnfluence of grazlng See Table 22 and F1gures 23 and 24 3 2 Phytomass 3 2 1 Above phytomass Data of the standmg hvmg b1omass before burmng and b1omass product1on by regrowth after burmng are summanzed m Table 23 The plots are d1fferent m date of burmng and grazmg mtens1t1es 107 Table 21 The occurrence of spec1es m the s1x weeks old regrowth vegetat1on after burmng w1thout grazmg F M NR ++ + ++ + 1 21 24 X X X 29 X X 30 X X 33 X X 37 X 43 X 00 X 46 X + + = mcrease 1n fraquancy parcentaga 1n both plots + = mcreasa m fraquancy percantaga 1n ona plot = decrease m frequency percentege m both plots = decraasa m fraquancy parcantaga 1n one plot X = changa 1n frequancy parcantaga h1ghar than 5% (for namas of spac1es sea Tabla 20) A ++ + X X X X X X X X X X X X X X The effect of the t1me of burmng on the stand1ng b1omass has been analyzed by comparong average above ground dry matter m protected areas s1x weeks after burnmg ( cagas ) at d1fferent dates February was a very dry month w1th prec1p1tat1on durong the s1x weeks after burmng of 36 mm Average above ground b1omass of und1sturbed regrowth after burmng was 34 3 g/m' The results of the measurements m the ramy season w1th a prec1p1tat10n durong the s1x weeks regrowth of 304 mm m March and 476 mm m Aprol showed an average above ground stand1ng b1omass of S4 9 g/m' and 118 5 g/m2 respect1vely The b1omass product1on after s1x weeks of und1sturbed regrowth m the ramy season were three to four t1mes as h1gh as the results of the measurements taken at the end of the dry season 3 2 2 Below phytomass These data show a very h1gh b1omass for a rest of two weeks comparad w1th the other rests to graze after burmng Also s1gn1f1cant d1fference can be observad between stockmg rate w1th med1um stockmg rata we can observe very h1gh root b1omass tour t1me those of low stockmg rata (Tabla 24) The reasons of these d1fferences have to be mvest1gate {botamc compos1t1on compactat1on fore etc ) 108 Table 22 Changas 10 frequency of specoes after graz10g burnt sub plots SPECIE HIGH STOCKING RATE Nr N ame February March Apnl Average 1 Andropogon b1corms 00 1 7 o 1 05 21 Hypt1s spp 09 10 o 94 68 24 Andropogon leucostachys 11 6 12 4 37 92 29 Otachyflum vers1co/or 1 9 94 1 o 28 30 Paspa/um pectmatum 30 05 10 7 47 33 Axonopus purpusu 05 12 8 39 3 1 37 Andropogon selloanus 4 1 11 3 1 4 29 43 Trachypogon vest1tus 67 58 94 34 00 Other specoes 15 3 12 2 12 3 13 3 46 Bare sool 15 3 16 7 07 10 9 SPECIE LOW STOCKING RA TE Nr N ame February March Apnl Average 1 Andropogon b1corms 00 00 00 02 21 Hypt1s spp 30 1 5 24 07 24 Andropogon leucostachys 1 o 15 1 40 06 29 Otachyflum vers1color 06 78 20 2 1 30 Paspa/um pectmatum 11 9 25 8 20 6 19 4 33 Axonopus purpusu 24 08 00 1 1 37 Andropogon selloanus 20 44 1 5 26 43 Trachypogon vest1tus 1 8 65 21 8 57 00 Other specoes 13 9 56 1 2 69 46 Bare sool 13 9 28 7 45 15 7 3 2 3 Sote preference of grazong (Table 24) The data of the foeld observatoon of ammals grazong the regrowth are shown 10 Table 25 and Table 26 For the analysos of the specoes preference for graz10g ammals preference 10doces (PI) were calculated and reportad 10 Tablas 27 and 28 109 ~ ~ o HIGH STOCKING RATE (O 5 ANIMAL/Ha ) Andropogon b1corms • HyptiS spp A leucostachys 1 1 Otachynum vers1color 20 10 Paspalum pectmatum Axonopus purpusu Andropogon seloanus 58 Trachypogon vest1tus • Other spec1es So11 ... 60 40 20 o 20 FIGURE 18 EFFECT OF BURNING ON VEGETATION DYNAMICS 3 • • BURr.~NG MONTH f±ltebruary 0march ~aprll 40 RELATIVES CHANGES 6 WEEKS AFTER BURNING ( IN DRY ANO BEGINNING OF WET SEASON ) Andropogon btcomts Hyptts spp A leucostachys Otachynum verstco/or Paspa/um pectmatum Axonopus purpusu ~ Andropogon se/oanus ~ ~ Trachypogon vestttus Other spectes Sotl MEDIUIW STOCKING RATE (O 25 ANIMAL/Ha) • ••• •• 60 40 20 o 14 7 20 221 22. 20 3 F1gure 19 Effect of burmng vegetat1on dynam1cs 208 Relatlve changes 6 weeks after burnlng m the dry beglnnlng of wet season BURNING MONTB mJ February ~March IJI Aprll 40 Andropogon seloenus Trechypogon vest1tus Other spec1es So1/ '*' .... 223 •• 28 60 40 20 o 20 ll!li H1gh stockmg rate !11 Med1um stockmg rate F1gure 20 Effect of burnmg on vegetatlon dynam1cs Relat1ves changas 6 weeks after burnmg m dry and begmntng of wet season 40 ( MARCH BURNING Andropogon b1corms HyptiSSpp A /eucostachys Otachyrwm vers1co/or •• Paspalum pectmatum ... ., Axonopus purpusu • • Andropogon seloanus ~ ¿;; Trachypogon vest/tus 1S Other spec1es Sol/ 30 20 10 o 10 20 30 40 Efl HIGH STOCKING RATE j;8l MEDIUM STOCKING RATE FIGURE 21 EFFECT OF BURNING ON VEGETATION DYNAMICS RELATIVES CHANGES 6 WEEKS AFTER BURNING IN DRY ANO BEGINNING OF WET SEASON APRIL BURNING Andropogon btcornts 15. Hyptts spp A leucostachys Otachynum verstcolor Paspalum pectmatum 228 - -.¡. Axonopus purpusu Andropogon se/oanus % ., Trachypogon vestttus 415 Other spectes 203 o So ti o 60 40 20 o 20 40 1! H1gh stockmg rate D Med1um stockmg rate F1gure 22 Effect of burnmg on vegetat1on dynam1cs Relat1ves changes 6 weeks after burnmg m dry and begmnmg of wet season ~ ~ "' HIGH STOCKING RATE Andropr.gon b1corms Hypt1s spp A leucostechys 2 Otechynum vers1co/or Pespa/um pectmetum o' Axonopus purpusu 1 21 Andropogon seloenus 1 '• Trachypogon vest1tus Other spec1es Solf 16 10 5 o 5 10 2 .. 16 .. ' Figure 23 Vegetation dynamics under grazmg Effects of h1gh stockmg rate on the frequency (%) of spec1es m a 6 weeks old regrowths of burnt savanna BURNING IIONTH m! February 0March ~April 20 MEDIUM STOCKING RATE Andropogon b/corms Hyptts spp A leucostachys 1 Otachyr/um verstco/or Paspa/um pectlnatum 20 - 122 :p ::¡ rLLL~ lú BURNING MONTH ffitebruary Omarch ~aprll Axonopus purpus/1 D[ "' M1 Andropogon se/oanus ~ Trachypogon vestltus V/// Other spec/es 1-d So ti . 1 30 20 10 o 10 20 30 40 FIGURE 24 VEGETATION DYNAMICS UNDER GRAZING EFFECT OF MEDIUM STOCKING RATE ON THE FRECUENCY {%) OF SPECIES IN A 6 WEEKS OLD REGROWTHS OF BURNT SAVANNA Table 23 H1gh Stockmg Rate Low Stockmg Rate Standmg phytomass before burmng and regrowth after burmng w1th Wlthout grazmg February March Apnl Date of burmng g DM/m 2 Before burnmg 393 9 257 7 362 6 O week regrowth 00 00 00 2 weeks regrowth 75 88 10 6 6 weeks regrowth• 64 47 34 1 6 weeks regrowth 1n cage 35 3 88 6 136 2 Before burmng 338 8 183 9 295 7 O week regrowth 00 00 00 2 weeks regrowth 10 9 54 29 9 6 weeks regrowth 28 21 2 56 o 6 weeks regrowth 1n cage 25 9 115 3 104 7 NB 6 weeks regrowth + 4 weeks consumpt1on Table 24 lnfluence of rest before grazmg (0 2 4 8 weeks) and stockmg rate after burmng on below phytomass (roots b1omass) ROOTS BIOMASS (0 80 cm) g DM/15 9 cm• g DM/m2 kg DM/ha Duncan groupmg Rest (weeks) o o 86 540 5400 B 2 2 01 1264 12264 A 4 o 92 579 5790 B 8 o 98 616 6160 B Stock1ng rate (ha/animal) Low 6 o 67 421 4210 e Med1um 4 1 71 1075 10750 A H1gh 2 1 31 824 8240 B Means w1th the same letter are not s1gmf1cant d1fferent 117 Table 25 S1te preference of ammals grazmg 2 4 weeks old regrowth T1me spent between 6 00 a m and 6 00 p m 10 each sub plot (percentage of t1me) March March Apnl Sub plots age 12 weeks regrowth) 14 weeks regrowth) 12 weeks regrowthl of regrowth Stockmg rate (days) H1gh 1 Med1um H1gh 1 Med1um H1gh 1 Med1um 14 48 o 49 o 30 4 27 6 28 56 3 53 7 72 29 5 17 7 59 2 52 1 86 31 2 32 7 135 20 9 30 6 35 89 149 57 8 1 196 08 1 3 32 6 1 210 72 37 258 08 00 22 33 272 06 05 315 00 1 5 1 7 22 329 1 4 02 Tabla 26 Regrowth quahty Regrowth age Protem IVDMD (weeks) (% DM) (% DM) 2 77 44 6 4 62 38 o 8 46 31 2 118 Table 27 Spec1es preference (preference .ndex) Spec1es selected by cattle on tour weeks regrowth after burmng most occurnng spec1es (fp > 1 0%) BURNING FEBRUARY MARCH APRIL fp PI fp PI fp PI (%) (%) (%) (%) (%) (%) HIGH STOCKING RATE Andropogon bicornls o 1 7 52 9 1 5 100 o Hypt1s spp 29 o 2 1 61 9 31 o o Andropogon leucostachys 24 79 2 14 3 o 1 1 727 Otachynum vers1color 12 9 36 4 74 54 22 o Paspalum pectmatum o 04 100 o 45 o Axonopus purpusu 05 100 o 25 7 444 49 28 2 Andropogon selloanus 12 1 83 4 1 70 7 11 8 720 Trachypogon vest1tuts 13 3 60 9 92 18 5 90 42 2 MEDIUM STOCKING RATE Andropogon b1corms 05 100 o o o Hypt1s spp o 27 100 o 85 o Andropogon leucostachys 1 o 100 o 57 o 30 o Otachyrwm vers1color 11 5 34 8 10 3 35 o 1 1 o Paspalum pectmatum 1 5 66 7 22 773 25 1 17 5 Axonopus purpusu 05 o o o Andropogon selloanus 20 100 o 44 727 1 5 o Trachypogon vestltuts 33 5 43 3 240 49 6 28 o o fp frequency of presence (%) for each spec1e PI preference mdex (%) = number of plants grazed x 1001 number of total plant for (each seroel or frequency of consumpt1on x 1 00 1 frequency of presence In February at low stock.ng rate select1v1ty of Andropogon b1corms (1) Andropogon leucostachys (24) and Andropogon sel/oanus (37) was very h1gh and selected to a lesser extend were Paspa/um pectmatum (30) Pamcum vers1co/or (29) and Trachypogon vestltus (43) Not selected was Axonopus purpusu (33) At h1gh stock.ng rate there was less select1V1ty among spec1es Selected 1n order of PI (%) were Axonopus purpusu (33) Andropogon leucostachyus (24) Trachypogon vest1tus (43) and Pamcum vers¡color (29) PI of Andropogon selloanus (37) was very low Not selected were Hypt1s sp (21 l and Paspalum pectmatum (30) In March the d1fferences 1n select1on of spec1es between h1gh stock.ng rate and low stock1ng rate were small Spec1es selected at h1gh stock.ng rate were Andropogon b1corms ( 1 l Hypt1s sp (21 l Paspa/um pectmatum (30) Andropogon selloanus (37) Axonopus purpusu (33) Pamcum vers1color (29) and Trachypogon vestltus (43) Here Andropogon /eucostachyus (24) was not selected desp1te 1ts h1gher frequency percentage comparad to the plot at low stockmg rate 119 Table 28 Select1v1ty of mmor spec1es (FP> 10%) by grazmg ammals Spec1es (No ) February March Apnl (No see Annex 1) fp(%) PI(%) fp(%) PI(%) fp(%) PI(%) H1gh stockmg rate ANCEPS 2 o o o AXAUR 3 o o o CAMPO 4 o o o DICRO 13 67 14 9 85 60 29 o ERIO 14 o o 08 o ELlO 15 o o o ES K IZA 16 o o o 1 o GALA 18 o 09 o o GIMNO 19 o o 87 LEPTO 23 05 80 o 04 o 33 o RUYEI 34 o o o RHYNBAR 35 o o 63 o SETA 38 05 o 34 o 68 16 2 SIPANE 39 1 9 o 2 1 o 45 o TRAPE 44 1 o 100 o o 04 100 o Low stockmg rate ANCEPS 2 o 1 3 30 8 o AXAUR 3 o o o CAMPO 4 o o 04 o DICRO 13 1 o 500 64 63 1 5 o ERIO 14 o o 22 o ELlO 15 o o 22 18 2 ESKIZA 16 o o o GALA 18 o 22 o GIMNO 19 o o 56 o LEPTO 23 o o o RUYEI 34 o o o RHYNBAR 35 o o 1 9 o SETA 38 o o o SIPANE 39 2 5 o 50 o 45 o TRAPE 44 20 75 o o 26 15 4 Spec1es selected at low stockmg rate are m arder of PI(%) Hypt1s sp (21) Paspalum pectmatum (30) Andropogon sel/oanus (37) Pamcum vers1co/or (29) and Trachypogon vestltus (43) Also here Andropogon leucostachyus (24) was not selected desp1te 1ts presence Andropogon b1corms (1) and Axonopus purpusu (33) are not ment1oned here because these were not present 1n the regrowth as m the sub plot burnt m Apnl The f1rst one of these ment1oned was not present m the vegetat1on befare burnmg the other one was w1th a frequency percentage of JUStO 4% and a frequency percentage m regrowth of the cage of 08% 120 In Apnl d1vers1ty m select1on of speCies was h1ghest espeCially at h1gh stockmg rate m Apnl Andropogon b1corms ( 1) Andropogon sel/oanus (37) Trachvpogon vestltus (43) and Axonopus purpusu (33) were selected more at h1gh stockmg rate compared w1th low stockmg rate On the other hand Andropogon b1corms ( 1) and Axonopus purpusu (33) were not present at all 1n the vegetat1on of the plot at low stockmg rate and therefore could not be selected at all Andropogon selloanus (37) and Trachvpogon vestltus (43) were not selected m Apnl at low stockmg rate desp1te the1r presence m contrast to the h1gh sto¡¡kmg rate where they were present and selected In the group of spec1es w1th frequency percentage less than 1 O% the d1vers1ty m select1on was h1ghest m the plot at low stockmg rate In February d1fferences m botamcal compos1t1on between the plots at h1gh stockmg rata and low stockmg rate did not occur But select1V1ty for spec1es was very clear Trasva petrosa (44) was h1ghly selected 100% and 75% for the plot at h1gh stockmg rate and low stockmg rata respect1vely Select1on of Leptocoryphwm lanatum (23) was 80% at h1gh stockmg rate and 25% at low stockmg rate and D1chromena Cllfata (13) was selected w1th a PI of 14 9% at h1gh stockmg rate and 50% at low stockmg rate The h1gh select1on m March of Axonopus anceps (2) at low stockmg rate (PI= 30 8%) contnbuted to the d1fference (Table 28) But at h1gh stockmg rate m March th1s spec1es could not be selected because 1t was not present m the natural vegetat10n nor m the regrowth Bes1des th1s the F1gures of plots at h1gh stockmg rate and low stockmg rate are almost equal also 1n th1s penod selectlvlty of spec1es less frequently occurnng m the nat1ve savanna was very low 01fferences m select10n of Elvonorus cand1dus ( 14) and Trasva petrosa (44) m Apnl were the mam contnbutors to the d1fference occurnng between h1gh stockmg rata and low stockmg rate The presence of Elvonorus cand1dus ( 14) m the nat1ve vegetat1on befare burnmg was O 4% at h1gh stockmg rate and 2 4% at low stock1ng rate Frequency percentage of th1s spec1es was zero m the regrowth vegetat10n m the plot at h1gh stockmg rate At low stock1ng rate the regrowth of these spec1es (fp = 2 2%) was consumed w1th a PI of 18 2% SelectiVItY of the other spec1es was very low m th1s penod at both h1gh stockmg rate and low stockmg rate 4 DISCUSSION ANO CONCLUSIONS 4 1 Vegetat1on dynam1cs 41 1 Burmng Spec1e speclflc compet1t1ve ab1hty and t1me of burnmg (water avallabihty) are very 1mportant for regrowth after burmng The botamcal composltlon after burnmg changed towards e1ther spec1es res1stant to burnmg adaptad to water stress or spec1es able to benef1t from the 1mproved growth cond1t1ons by the d1sappearance of spec1es For sorne annuals growth was lnltlated by the temperatura effect of f1re and for sorne spec1es growth cond1t1ons 1mproved because of reduced compet1t1on for growth factors w1th other plants Regrowth of annuals was poor after burmng m the dry penod though Hvpt1s sp took advantage of the decrease 1n compet1t1on w1th other plants for growth factors hke water hght and nutnents Under Clrcumstances Wlth h1gher water availabihty recovery of Hvptls sp mcreased even more Spec1es w1th a germmat1on wh1ch was 1n1t1ated by the f1rst rams of the 121 wet season were k1lled by the burmng The decrease m number of annual spec1es after burnmg can not be generalizad because sorne annuals start the1r growth cycle another t1me because of the1r strateg1es adapted to the spee1f1c env1ronmental factors Good recovery under wet c1rcumstances was also shown by the spec1es Andropogon leucostachyus and Paspa/um pectmatum The key spee~es for management of th1s type of savanna T vestltus decreased m frequency percentage m the s1x weeks old regrowth after all burmng treatments Th1s was partly replaced by Paspalum pectmatum wh1ch 1s another 1mportant spec1es m the savanna of Canmagua Th1s decrease m the key spec1es m1ght stay so or may recover m t1me but data about further developments are lackmg m th1s expenment Under wet c1rcumstances when so1l was suturated w1th water and prec1pltat1on was h1gh vegetat•onal recovery after burmng was fast and considerable 4 1 2 Grazrng lntroducmg grazmg two weeks after burnmg caused mcreases m the number of spec1es under all grazmg treatments mamly contnbuted by annual spec1es An mcrease under grazmg also occurred for Hypt1s sp Th1s was probably related to the reduced compet1t1on for growth factors w1th other spec1es and 1mproved growth cond1t1ons because of the mterrupt1on of the plant by grazmg Th1s effect of grazmg on the vegetat10n and seemed to be relatad to stockmg rate 4 2 Srte preference of grazrng ammals By companng data m Table 8 the s1te preference for d1fferent stockmg rates were obtamed The b1ggest d1fferences occurred m the f1rst observat1ons m March The ammals m the plot at HSR spent 11 8% more of the1r grazmg trme m the sub plot w1th regrowth of 72 days old than the ammals m the plot at LSR These latter ammals mstead spent more t1me grazmg m the sub plot w1th 135 days old regrowth (d1fference of 1 O 5%) The same occurs m the measurements of Apnl although to a lower extent ammals m the plot at HSR spent 7 1 % more grazmg t1me m the sub plot w1th 72 days old regrowth than ammals 1n the plot at LSR These spent th1s t1me m the followmg sub plot w1th a regrowth of 135 days old Stnkmg 1s the d1fference occurnng m t1me spent grazmg m the sub plots between March and Apnl In March the ammals preferred to graze the sub plot w1th youngest regrowth of 14 days old as expected Th1s 1s contrast to the data of Apnl when the ammals preferred to graze the sub plot w1th 7 2 days old regrowth to the plot most recently burnt for grazmg D1fferences m average temperatura and prec1p1tat•on of these months were small But here one should reahze that 1n March part of the prec1pltat1on was usad to m01sten the dry ground wh1ch reduced ava1lable water for plant growth Th1s resultad m an above ground dry matter ava•labll1ty of 8 8 g/m 2 at HSR and 6 5 g/m2 at LSR m March In Apnl these f1gures were 1 O 6 g/m2 at HSR and 29 9 g/m2 at LSR 4 3 Spec1es preference Spec1es preferred by grazmg ammals at young age were Andropogon b1corms Axonopus purpusu Andropogon se/loanus and Trachypogon vestltus Spec1es decreasmg under the grazmg treatments were Andropogon leucostachyus and Paspalum pectmatum Trachypogon vestltus the key spec1es m th1s savanna type showed decreases m frequency percentage 122 under grazmg at h1gh stockmg rate and mcreases m frequency percentage under grazmg at med1um stockmg rate after burmng 1n February as well as 1n Apnl But the oppos1te effect occurred under grazmg after burmng 1n March Th1s could not be explamed w1th data from th1s tnal but m1ght be caused by env1ronmental factors not mvest1gated here Therefore 1t 1s not poss1ble to make here any conclus1ons about the mfluence of grazmg on the occurrence of Trachypogon vestltus Under grazmg the occurrence of bare so1l was 1ncreased (Table 20} wh1ch seemed to be related to water avallab1hty (prec1p1tatlonl and to stockmg rate After burmng 1n the dry season the mcrease 1n occurrence of bare so1l was equal for both stockmg rate 1n the regrowth of March the mcrease m the occurrence of bare so1l due to grazmg was h1ghest at low stockmg rate whlle 1n the regrowth of Apnl the occurrence of bare so1l dueto grazmg was very small 4 4 Phytomass and product1on Grazmg act1v1t1es also caused decreased m the above ground b1omass wh1ch also seemed to be related to water ava1lab1hty (preclp1tat1on} and to stockmg rate Absoluta reduct1ons 1n above ground regrowth b1omass due to grazmg were equal for both stockmg rate 1n February were h1ghest at med1um stockmg rate 10 March and were h1ghest at h1gh stock1ng rate m Apnl although 1n all of the months relat1ve reduct1ons were h1ghest at h1gh stockmg rate 4 5 Ammal performance These features are reflectad m tha hfewa1ght changas of an1mals grazmg undar th1s management system Ammals graz1ng s1x weeks old ragrowth after burnmg 1n February severely loose we1ght due to msuff1c1antly ava1lable good forage The same treatments 1n March laad to a h1gher ava1lab1hty of green forage though and unexpected loss 1n hvewe1ght of ammals grazmg th1s vegetat1on occurs Th1s m1ght be axplamad by the changa 1n gutf1ll wh1ch causes temporary reduct10ns m hvewe1ght whan d1ets cons1st of h1gh protem concentrat1ons after a penod of low protem contammg d1ets After the treatment 1n Apnl an1mals show compensatory mtaka and compensatory hvewe1ght gams Th1s leads to the conclus1on that although th1s expenment last for only three of 10 total 1 2 months of one year the a1m of the proposed management system 1s not reached fluctuat1ons 1n we1ght gam of grazmg ammals do st1ll occur Although graen matenal1s avallable 1n the dry month the msuff1c1ency hm1ts product1on Therefora 1t 1s proposed to mcrease the areas for burmng 10 the dry season to mcrease the quant1ty of green forage on offer durmg th1s penod Because the expenment lastad only for three months th1s proposal needs more mvest1gat1on for the effect of annual burmng on the botamcal compos1t1on by monthly measurements of all plots 123 ANNEX 1 LIST OF ABBREVIATIONS OF SPECIES Nr Abbrev1&t1on Famlly Spec1e 1 ABICOR Poaceae (a) Andropogon b1corms 2 ANCEPS Poaceae (p) Axonopus anceps 3 AXAUR Poaceae Axonopus aureus 4 CAMPO Poaceae Paspalum campestre 5 CA PICA Fabaceae Sty/osanthes cap1tata 6 CA SEA Flacourt1aceae Caseana ulmlfo/Ja 7 CASSIA ? Cass1a sp 8 CENTRO Fabaceae Centrosema sp 9 CUTO Fabaceae C!Jtona gwanens1s 10 CYPURA lndaceae Cypura pa/udosa 11 DESMO Fabaceae Desmod1um barbatu 12 DICO ? 0Jcotlfedoneas spp 13 DICRO Cyperaceae D1chromena cJ!Jata 14 EllO Poaceae Elyonurus cand1dus 15 ERIO Fabaceae Enosema spp 16 ES K IZA Poaceae SchJzachynum hJrtJflorum 17 FILAN Euphorb1aceae Phyl/anthus mrur1 18 GALA Fabaceae Ga/actJa g/auscesens 19 GIMNO Poaceae (a) Gymnopogon fo/Jousus 20 GRAM Poaceae Grammeas spp 21 HYPTIS Lab1atea (a) Hypt1s spp 22 !POMO Convulvulaceae /pomoea 23 LEPTO Poaceae (p) Leptocoryph1um lantum 24 LEUCO Poaceae (p) Andropogon leucostachyus 25 LINDA Scrophulanaceae Lmderma dlffusa 26 MELO Stercuhaceae Meloch1a VJffosa 27 MIMO M1mosaceae M1mosa spp 28 OVAL Fabaceae Desmod1um ova!Jfo!Jum 29 PAVER Poaceae (p) Pamcum versJcolor 30 PEPE Poaceae (p) Paspa/um pectmatum 31 PERATA Poaceae Imperara brasJ!JensJs 32 PODO Cyperaceae Rhynchospora podeosperma 33 PURPU Poaceae (p) Axonopus purpusu 34 RUYEI Poaceae Pamcum rudge1 35 RYNBAR Cyperaceae Rhynchospora barbara 36 RYNCON Cyperaceae Rhynchospora confm1s 37 SELOA Poaceae (p) Andropogon selloanus 38 SETA Poaceae (p) Serana gemculata 39 SIPANE Rub~aceae S1panea prarens1s 40 STELLA Poaceae (p) Paspa/um srellarum 124 41 42 43 44 45 STYLO TINCTA TRACHY TRAPE TRAPLU a = annual p = perenmal Fabaceae Poaceae (pi Poaceae (p) Poaceae (p) Poaceae (p) 126 Stylosanthes sp Aflst1da tmcta Trachypogon vest1tus Thrasya petrosa Trachypogon plumos1s CHAPTER 111 PROJECT TC 01 PROTOTYPE SUSTAINABLE CROPPING SYSTEMS FOR THE BRAZILIAN CERRADOS M1gue/ A Ayarza (TL) INTRODUCTION In the last three decades the Cerrados of Braz1l covenng about 205 m1lhon hectares have contnbuted 1mportantly to the country s crop and hvestock product1on However the mtens1ve use of th1s area part1cularly for monocroppmg and pasture development has g1ven nse to forms of land use that are ne1ther env1ronmentally nor econom1cally sustamable Alternat1ve land use systems are needed to halt and revert dechnmg product1v1ty and losses of so1l and water Among the technolog1es w1th potent1al to do th1s the combmat1on of crops and pasturas m space and t1me 1s one of the best opt1ons Th1s technology has the potent1al to mcrease overall productiVIty enhance s01l fert1hty and contnbute to 1mprove soc1oeconom1c cond1t1ons of farmers CIAT and EMBRAPA carned out p1oneenng stud1es to charactenze the Cerrados accordmg to ma¡or agroecolog1cal classes (Pro¡ect TC03) After pnont1zmg these classes the Center CPAC/EMBRAPA and the Federal Umvers1ty of Uberland1a 1n1t1ated on stat1on and on farm stud1es to develop land management opt1ons for sustamable development PURPOSE Generate technolog1es and land management strateg1es for sustamable agncultural development m the Braz1han Cerrados OUTPUTS 1 Prototype technolog1es that allow lastmg 1mprovement 1n the eff1c1ency of resource use and that control so1l and water degradat10n 2 ldent1f1cat1on of so1l md1cators of sustamab1hty ACTIVITIES A senes of on stat1on and on farm act1v1t1es were carned out dunng the last two years to develop sustamable crop pastura systems w1th the part1c1pat1on of severa! nat10nal and mternat10nal mst1tut1ons Below are reportad the act1v1t1es methodolog1es and mam results 1 Development of lmproved Agropastoral Systems The 1mpact of the nce pasture technology to recla1m degradad pasturas m the cerrados has been demonstrated by researchers from the Nat1onal Center for Research m R1ce and Beans (CNPAF) New nce vanet1es and other crop components are currently under testmg to 1mprove crop opt1ons However httle has been done to mcrease pastura components espec1ally forage legumes Legumes are 1mportant dunng the dry season when the quahty of 1mproved grasses drops and 1mprove N ava1lab1hty and so1l orgamc matter quahty avo1dmg pasture degradat1on Less work has been done to select pasture components for rotat1ons w1th crops m mtens1ve management systems A collaborat1ve effort was estabhshed between our Sect1on the Trop1cal Forage Program of CIAT and the Umvers1ty of Uberland1a to mcrease pasture opt1ons Act1v1t1es mcluded a 1 small plot tnals to select prom1smg germplasm b) quant1f1cat1on of the tradeoffs between crop and pastura components and 3) mult1phcat1on of prom1smg forage spec1es for on farm testmg 1 1 Select1on of pastura germplasm for agropastoral systems M A yarza (TLPJ and E A P1zarro (TFPJ Test1ng act1v1t1es started m 1992 plantmg several grass and legume access1on w1th nce and corn m several farms m Uberland1a m large plots The results of the work showed that Stylosanthes gwanens1s cv mme!fao a forage legume recently released by EMBRAPA was successfully estabhshed m nce pasture systems m clay and sandy so1ls In contrast establishment of the legume was poor when planted w1th corn and Pamcum Vencedor under 1ntens1ve use of 1nputs Dunng the present year we tested other prom1smg grass and legume opt10ns for nce and corn systems A m1xture of Paspalum atratum BRA 961 O and Sty/osanthes mme!fao was plantad s1multaneously w1th nce to recla1m a degradad pastura of 8 ruz¡z¡ensJs m a sandy so1l of Uberland1a S1multaneously we tested the adaptat1on and y1eld potent1al of prom1smg grass legume mixtures to the cond1t1ons of the reg1on Twenty one legumes and nmeteen grasses were plantad m small plots usmg a mod1f1ed reg1onal tnal arrangement Grasses were planted w1th a common legume (Stylosanthes gwanens1s cv mmelfao 1 and legumes w1th a common grass (8rachJafla decumbens cv 8asil1sk l Establishment of the Paspalum and Stylosanthes m1xture plantad s1multaneously w1th nce was excellent Pastura was ready for grazmg after nce harvestmg On the other hand nce y1elds were reduced by 30% due to compet1t1on from the rema1mng pasture of 8 ruz¡z¡ensJs Average y1elds were approx1mately t 1 8 ton/ha Ammal performance measurements started m May 1994 The results of the f1rst year of evaluat1on of the small plot expenment showed that Stylosanthes was established well w1th most of the brach1ana grasses Among grasses BrachJarJa decumbens 8RA 4308 was the most product1ve access1on dunng the wet and the dry penods Among legumes Stylosanthes mme!faO and Arach1s pmtoJ 8RA 41143 showed a h1gher leaf retent1on dunng the dry penod comparad to the Centrosema Calopogomum and PuerarJa access1ons mcluded m the tnal The adaptat10n tnal w1ll contmue for another year to select most prom1smg spec1es for crop pasture tnals 1 2 Tradeoffs betwaen crops and pasturas M Ayarza (C/A TJ and L Vi/ela (CPA CJ The results of our work last year showed that the establishment of legumes was markedly 1nfluenced by the crop and grass compamons Dunng the present year we plantad a small plot expenment m a sandy so1l of Uberland1a to mvest1gate the effect crops (nce and corn) and several grasses on the establishment of Stylosanthes mme!fao and Arach1s pmto1 8RA 41143 Crop and pastura components were planted alone or s1multaneously The expenment mcluded a treatment of delayed sowmg of the grass component to reduce early compet1t1on for the crops and the legumes Dry matter product1on of grasses and legumes was measured at crop harvest1ng 128 The results of the expenment w1th corn showed that establishment of Stylosanthes was more affected by compet1t1on from the grasses than by compet1t1on from the crop 1tself (Table 1) Compet1t1on from the Pamcum Vencedor was stronger comparad to the Paspa/um spec1es Grasses also reduced corn y1elds by 7% (Table 2) In turn y1eld potent1al of the pastura spec1es were reduced by the crop (Table 3) Delayed plantmg of the grasses suppressed the compet1t1on effects of the grass but retardad 1ts establishment (Table 4) Table 1 Effect of the crop and grass components on the establishment of two forage legumes 1n an 1mproved crop pastura system plantad m a sandy latossol of Uberlílnd1a MG Treatments Grass components Vencedor Paspalum Vencedor Paspalum /Arachts /Arachts /M1ne1riio /M1ne1riio kg/ha Legume + corn 453 686 1710 1919 Legume + grass 158 309 2 341 Legume + grass + corn 96 221 11 144 Grass plantad 30 days after legume and corn 545 618 723 1078 Tukey value (5%) w1th1n columns 306 w1thm rows 427 Table 2 Effect of grass and legume components on corn gram y1elds 1n a crop pastura system plantad 1n a sandy latossol of Uberlílnd1a MG Systems Gram (kg/hal Corn 6364 Corn + forage legumes 1 6400 Corn + grasses2 5837 Corn + legumes + grasses 5586 Corn + legumes + grasses 30 days after 6484 Tukey value (5%) 964 1/Legumes Stylosanthes gwanensts cv Mme•riio and Arachts pmtot 2/Grasses Pamcum maxtmum cv Vencedor and Paspalum atratum 129 Y1eld reduct1on (kg/hal 527 778 Tabla 3 Effect of several strateg1es of plantmg grasses and legumes w1th corn on D M product1on of the pastura components m a sandy latossol of UberUind1a MG Systems Pasturas Vencedor Paspalum Vencedor Paspalum +Arach1s +Arach1s +Mine~rllo +M1ne1rllo kg/ha Legume + grass 15130 14055 16789 13843 Legume + grass + corn 6257 2417 4949 2526 Legume + corn 453 686 1710 1919 Grass+Corn 5965 2128 6111 2221 Grass plantad 30 days after legume and corn 1157 917 1721 1287 Tukey value(5%) w1th1n columns 2997 w1thm rows 3029 Table 4 Effect of grass spec1es and sowmg date on gram y1eld and forage product10n of a nce pastura system planted m a sandy latossol of Uberland1a MG R1ce Pastura Systems Sowmg dates of the pastura Sowmg dates S1multaneous 30 days S1multaneous 30 days after after nce kg/ha R1ce + Paspalum + M1ne1rao 1106 2189 4808 628 R1ce +Vencedor+ M1ne1rao 194 2156 7458 1417 R1ce + Marandu + Mme1rao 1208 2556 7299 747 R1ce + Paspalum + Arach1s 1014 2445 5677 1988 R1ce +Vencedor +Arach1s 314 2203 7616 2733 R1ce + Marandu + Arach1s 1023 2570 6187 1612 R1ce alone 2666 2437 Tukey(5%) w1thm columns 2094 842 w1thm rows 2854 952 130 Grass compet1t1on was stronger for the nce crop than for corn crop Gram y1elds were reduced 50 80% dependmg on the spec1es P Vencedor and 8 bflzantha were the most aggress1ve spec1es Stylosantbes and Arachts estabhshed very well w1th most pasture spec1es except P Vencedor Delayed plantmg of grasses had the same effects observad m corn expenment These results md1cate that there are product10n tradeoffs m crop pasture systems and that there 1s a need to reduce grass compet1t1on and select more aggress1ve legumes for corn pasture rotat1ons m h1gh 1nput systems 1 3 Mult1phcat1on of prom1smg pasture spec1es (Umv Fed Uberlandta M Ayarza andE A Ptzarro) Lack of enough seed 1s always a ma]Or barner for the testmg of prom1smg forage components 1n large plots at the farm level To overcome th1s problem we made efforts dunng the present year to promote seed mult1phcat1on by farmers and reg1onal lnst1tUt1ons of Uberland1a One ha of Stylosanthes mmeuao was plantad m 1993 m one of the collaboratmg farms Part of the seed collected th1s year was dehvered to the Umvers1ty of Uberland1a and severa! farmers to plant mult1phcat1on plots As a result of th1s tour ha of Stylosanthes one ha of Arach1s pmto1 and 500 m2 of Paspalum atratum BRA 961 O have already been estabhshed dunng the present year Seed produced w1ll be purchased by the Sect1on to expand on farm testmg to other product1on systems and to support research act1v1t1es of CIA T and EMBRAPA m other areas Remamed seed wlll be put m the market by the Un1vers1ty CONCLUSIONS ANO WORK PLAN Sty/osanthes gwanensts cv Mme!fao and Arachts pmto1 have the potent1al to become components of 1mproved nce pastura systems m a short t1me Enough seed w1ll be ava1lable for further testmg by farmers m the reg10n m commg years 1f the collaborat1ve effort we are develop1ng w1th farmers and the umvers1ty proves to be successful In the mean t1me we wlll cont1nue our act1v1t1es to select forage legumes for mtens1ve management systems and expand testmg of 1mproved germplasm for other purposes ( 1mproved feedmg for m1lk product1on cover crop and not tlll systemsl Technology adJustment to reduce grass compet1t1on 1n crop pasture systems 1s expected to be conducted by the collaboratmg lnStltUtiOnS 2 Synerg1sm and tradeoffs of crop pasture mtegrat1on Management pract1ces such as land preparat1on hmmg and fert1hzat1on and ammal management mfluence the potent1al contnbut1on of crops and pasturas m mtegrated crop pastura product1on systems To measure these effects a long term expenment was estabhshed m a red dark Latossol at EMBRAPA CPAC m Planaltma Braz1l 1n 1991 The ob]ect1ve of the expenment 1s to determme the effect of mtegrat1on on crop and ammal product1v1ty and to 1dent1fy s01l key parameters relatad to 1mprovement or degradat10n The expenments 1ncludes contmuous crop and pastura systems and crop pastura and pasture crop rotat1ons cycles of f1ve years Land preparatlon methods evaluate the effect of d1skmg (convent1onalland preparat1onl and deeper land preparat1on methods (flexible land preparat1on method) on crop product1v1ty and soll phys1cal propert1es Fert1hzat1on treatments mcluded the effect of convent1onal fert1hzat10n and a correct1ve fert1hzat1on mcludmg the use of gypsum 131 on crop and pasture product1v1ty and so1l fert11ity Graz10g management treatments evaluate the effect of low and h1gh graz10g ~ntens1t1es on ammal product1on botamcal compos1t1on and lltter product1on 10 grass only and grass legume pastures Measurements are taken by a group of researchers from CIA T CPAC and the Umvers1t1es of Bayreuth (Germany) and Cornell (USA) We are momtonng the short and long term changes 10 crop and ammal product1v1ty and on s01l chem1cal phys1cal and b1olog1cal propert1es 10 the management treatments 1ncluded 10 the expenment The expenment 1s complementad w1th satelllte plots located 1n several farms 10 Uberland1a These plots were establlshed 1n 1992 on sandy and clay soll types and two product1on systems ( cont~nuous pastures and cont10uous cropp1ng systems)to determine the potent1al contnbut1on of forage legumes to 1mprove susta1nab11ity of crop pasture systems We are momtonng changes 10 a m mal product10n and so1l fert11ity 10 the 1mproved and the unrecla1med systems over t1me Below are reported the results of the work 132 INTRODUCTION PROJECT TC 03 DYNAMICS OF LAND USE M1guel A Ayarza (TLJ Land use m the cerrado reg1on has been extraordmanly dynam1c dunng the las three decades lts patterns have been alterad contmually by government pohc1es f1scal mcent1ves d1sturbance of natural env1ronments and mtroduct10n of new crops and technolog1es These patterns have not been systemat1cally charactenzed and ne1ther the reasons for rates and d1rect1ons of change well understood To contnbute to the lastmg product1v1ty of savanna ecosystems 1t 1s 1mportant to understand the 1mpact of human act1v1t1es on the resource base and consequently formulate appropnate pohc1es and technolog1es PURPOSE 1 Understand the dynam1cs of land use 2 Asses the 1mpact of new technolog1es on land use 3 Appra1se pohcy alternat1ves for 1mproved land use OUTPUTS 1 Georeferenced data bases on the spat1al and temporal dynam1cs of land use 2 Typolog1es of land use classes to help to 1dent1fy domams of extrapolat1on 3 Dynam1c and mult1ple ob¡ect1ve models to asses the econom1cal and ecolog1cal 1mpacts of land use ACTIVITIES Dunng the last three years severa! act1v1t1es were conducted to charactenze the dynam1cs of land use m the cerrado reg1on and select representativa study areas to conduct detallad stud1es The operat10nal strategy cons1sted m the bread class1f1cat10n of the reg1on m mayor zonas followed by a more deta1led charactenzat1on of severa! representativa areas Th1s report summanzes the contnbut1on of many members from CIAT EMBRAPA and NGO orgamzat10ns workmg m Braz1l Below are descnbed the mam act1v1t1es Methodolog1es followed and the mam results The hst of pubhcat1ons can be found 1n append1x 1 1 Dynam1cs of Settlement and Agr1culture 1n Brazll s Forest Margms and Savanna Ecosystems Th1s act1v1ty was conducted by the Instituto de Soc1edade Populacao e Natureza (ISPN) under a short term contract w1th CIAT m 1991 The procedure cons1sted m the zomng of the two reg1ons us1ng EMBRAPA map of Braz1l s vegetat1on and supenmpose upon 1t a map w1th the same scale mformat1on contammg the country s state and m1croreg1onal subd1vls1on lnformat1on was collected from the censuses from 1 970 to 1985 and the prehmmary results of the 1991 demograph1c census The study showed that the occupat1on of the cerrados started spontaneously m the 1 960s However 1mportant changes occurred dunng the economiC expans1on of Braz1l between 1960 and 1980 A combmat1on of federal pohe~es h1gh mflat10n and speculat1on were the factors 133 respons1ble for the very s1gmf1cant mcreases 1n area and agncultural pnces m the cerrados The growth of domest1c and fore1gn markets and the gradual expans1on of transportat1on also played a fundamental role m the agncultural expans1on and modermzat1on The analys1s of ava1lable 1nformat1on allowed to d1stmgU1sh tour zonas m the reg1on 1) a zone dynam1c usmg modern technology anda dynam1c agncultural system encompassmg most of the south of the cerrados (Mato Groso do Sul M1nas Gera1s G01as and part of Mato Groso) 2) a fa1rly new area w1th low agnculturalmtens1ty but h1gh techmcallevels around Bras11ia and the west of M1nas Gera1s 3) an area of recent agncultural expans1on w1th low agnculture 1ntens1ty and 4) an almost empty area where agnculture 1s st1ll mc1p1ent and have low techmcal levels These are remete areas m wh1ch cond1t1ons for expans1on of commerc1al agnculture 1s st1ll precanous 2 Area class1f1cat1on and mappmg for the Cerrado reg1on of Brazll Th1s study was camed out m 1992 by P Jones and h1s team w1th the collaborat1on of J Macedo and B Pmhe1ro from EMBRAPA The ob¡ect1ve of the work was to subd1v1de the cerrado reg10n mto d1st1nct1ve agroecolog1cal classes So1ls climate and land use data were comb1ned usmg the GIS fac11it1es at CIAT Twelve areas were selected by f EMBRAPA as the most mterest1ng areas to conduct collaborat1ve research on resource management The role of CIA T was to descnbe the agroecolog1cal classes present m each area and 1dent1fy the1r s1ze m other areas w1thm the cerrado reg10n The work allowed to d1v1de the cerrado reg1on mto eleven agroecolog1cal classes and to rank the 12 areas of mterest accordmg to the number of classes and the and the area covered by each class Gurup1 Campo grande Go1ama Paracatu Uberland1a and R1o Verde were listed as the most representat1ve areas 3 Soe~oeconom1cal charactenzat1on of tour reg1ons of the cerrado Th1s act1v1ty cons1sted 1n the analys1s of the soc10econom1c mformat1on of tour reg1ons of the Brazi11an cerrados ( Campo Grande Uberland1a Rondonopolis and R1o Verde) These areas were des1gnated as havmg h1gh pnonty w1th1n the EMBRAPA/CIAT pro¡ect to develop sustamable land use systems The work was conducted by L R1vas from CIAT and G Pere1ra from EMBRAPA They analyzed the s1m11itudes and d1fferences between reg1ons usmg mformat1on about trends on land mventory and use crop and livestock product1on mechamzat1on mfrastructure There were found large d1fferences 1n land use mtens1ty farm s1ze and mfrastructure among the four reg1ons Campo grande has most of the three fourths of 1ts area m extens1ve cattle ra1smg actiVItles R1o verde and Rondonopolis have the largest agncultural areas and Uberland1a the lower average farm s1ze h1gher mtens1f1cat10n levels and better capac1ty of storage of gra1n at the farm level 4 ldentlf1cat1on of land use patterns and problems In order to charactenze the mam product1on systems detect trends of mtegrat1on and 1dent1fy problems and demands for technology two short term v1s1tS were camed out m 1991 and 1993 by a multldiSCiplinary team composed by sc1ent1sts from EMBRAPA and CIAT The methodology used m the v1s1ts was the Rap1d Rural Appra1sal The VISit mcluded mterv1ews 134 w1th farmers and local mst1tut1ons to account for the problems and research demands from the farmer perspectiva The tnps across the cerrados allowed to observe a cons1sted relat1onsh1p between landscape and product1on systems Annual croppmg systems w1th h1gh use of mputs and technology are present m the flat areas (chapadoes) of med1um to h1gh clay content wh1le extens1ve hvestock systems are dommant on gently slopmg lands over more frag1le so1ls Although average farm s1ze he between 500 1 000 has there are a S1gn1f1cant number of small farms located on h1gher base status so1ls on slopmg landscapes They produce most of the m1lk m the cerrados All product1on systems have problems of b1ophys1c and soe~oeconom1c sustamab1hty WORK PLAN The mformat10n generated so far by the proJect has led to develop an obJeCtlve class1f1cat1on of the cerrado agroecosystem m several classes and the select1on of the Uberland1a reg1on 1 M G) a case study area for detallad stud1es on land use dynam1cs and development of prototype sustamable systems However to ach1eve the mam obJeCtlves and to develop the outputs proposed by the project 1t 1s necessary to 1mprove the current agroecolog1cal class1f1catlon and to strength our understandmg on soc1oeconom1c and ecolog1c 1mpact of land use m the cerrados The followmg are the act1v1t1es planned for the next three years 1 lmprovement of the agroecolog1cal class1flcat1on system P Jones 8 Be// (LMJ J Sm1th and M Ayarza (TLJ Th1s act1v1ty w111 mclude the refmement of the current agroecolog1cal class1f1cat1on systems and a ground truth venf1cat1on of agroecolog1cal classes across the cerrado reg1on Most recent mformat1on w1ll be mcorporated mto the current database to 1mprove the prec1s1on on the agroecolog1cal classes F1eld v1s1ts to selected reg1ons of the cerrados w1ll allow to determme the cons1stency of the class1f1cat10n 2 Evolut1on of farmmg and reg1onal systems and 1mpact of land use on the env1ronment J Sm1th (TLJ W Doppler (U Hohenhetm) and M Ayarza (TLJ The ObJectlve of th1s act1v1ty w111 be to generate bas1c knowledge on the dynam1cs and sustamab1hty of current product1on systems m the cerrados and to prov1de tools for assessmg the poss1ble future 1mpact of 1mproved technology marketmg and pohc1es Through th1s approach 1t w1ll be poss1ble to generate hypothes1s pertammg the development of sustamable systems and 1dent1fy research and pohcy measures for successful 1mplementat1on of sustamable farmmg systems Th1s 1n1t1at1ve has been presentad as a JOint proJeCt between CIAT and the Umvers1ty of Hohenhe1m (Germany) for fundmg by BMZ 3 Vahdat1on of dynam1c modela to asses the econom1c and ecolog1c 1mpact of land use J Smtth (TLJ Th1s act1v1ty cons1der the testmg and vahdat1on of hohst1c models encompassmg econom1c ecolog1c and soc1al mteract1ons for the development of sustamable systems Sorne of these models have already been developed However 1t 1s necessary to test the1r usefulness for the cond1t1ons of the cerrados Ava1lable mformat1on from selected product1on systems w111 be used to test the models and to adJuSt parameters 135 SUBPROJECT TC51 SOIL INDICATORS OF SUSTAINABLE AGROPASTORAL SYSTEMS M A Ayarza (TL) INTRODUCTION There ts evtdence that sorne extsttng cropptng systems tn the savannas of Lattn Amenca are not sustatnable etther tn btophystcal or economtc terms Sotl eroston compactton and general low levels of fertthty hmtt agncultural productton on the actd sotl savannas CIA T has prototype technologtes combtmng ustng legume based pasturas and crops whtch can tmprove sotl condtttons tncrease agncultural producttvtty and reversa degradatton Thts pro¡ect wtll determtne the relattonshtp between sotl orgamc matter and sotl phystcal properttes tn sotls under dtfferent land uses PURPOSE Generate baste knowledge on sotl phystcal degradatton and enhancement processes and develop effecttve cropptng systems for savanna agroecosystems OUTPUTS 1 Quanttftcatton of the dynamtcs of sotl organtc matter tn alternattve cropptng systems 2 ldenttftcatton of tndtcators of degradatton and enhancement 3 A conceptual model of sotl phystcal changes tn alternattve productton systems 4 Esttmatton of sotl losses at the watershed leve! ACTIVITIES The pro¡ect started wtth the arnval of two PhD students last September to conduct the work One of the students ts worktng on the effect of land use on sotl orgamc matter composttton Sotl samples were taken from severa! selected sttes tn two sotl types (clay and sandy texturas) tn Uberlandta The selected land use systems are Conttnuous annual cropptng systems conttnuous pastura systems nattve cerrado planted forests and crop pasture rotatton systems Severa! techmques tncludtng 13C NMR and GC spectroscopy to wtll be used to tdenttfy the composttton of SOM tn vanous fracttons of the sotl Further studtes wtll be conducted to determtne the dynamtcs of most senstttve fracttons tn relatton to sotl management The second acttvtty ts related wtth the water and nutnent dynamtcs m tntegrated and not tntegrated crop pasture systems Measurements wtll be carned out wtthtn our crop pasture tntegratton expenment at CPAC Severa! treatments tncludtng the nattve cerrado and the crop and pasture systems were selected for thts study The ob¡ecttve of the work ts to study the effect of tntegrated and not tntegrated systems on water avatlabthty and dynamtcs of nutnents denved from orgamc matter mtnerahzatton (N P and S) Selected plots wtll be sampled over ttme to asses the short term effect of crop pastura rotatton next year 136 CHAPTER IV TA 01 PROTOTYPE SUSTAINABLE CROPPING SYSTEMS FOR FOREST MARGINS M1chae/ Thung CIA T/EMBRAPA/CPAF Acre Freid research actrvrtres rn the Forest Margrns srte of Brazrl begun rn the second tnmester of 1994 wrth the transfer of a cropprng systems specrahst to Rfo Branco Acre and development of a mrnrmal logrstrcal rnfrastructure whrch took longer than antrcrpated Dunng thrs rnrtral penad expenments were setup pnmanly rn the State of Acre others rn cooperatron wrth an ICRAF senror staff located rn Porto Velho Rondonra As prevrously agreed wrth partners rn the Alternatrves to Slash and Burn consortrum CIA T agronomrc expenments concentrated rnrtrally rn the rntroductron and testrng of germplasm wrth the arm of rdentrfyrng components for future expenmentatron rn cropprng systems In addrtron the cropprng systems specrahst partrcrpated rn the charactenzatron of the regron reportad elsewhere Desprte delays rn developrng the requrred rnfrastructure a number of expenments were conducted rn the Expenmental Statron of CPAF/Acre and on farmer frelds rn the settlement project P A D Pedro Perxoto dunng the frrst semester of 1994 as follows 1 Beans regronal tnal for Carraca and Prnk seeded grarn type 2 lntroductron of leadrng bean cultrvars from southern Brazrl 3 Natronal Web Bhght nursery 4 Adaptatron tnal of rnterracral crosses of beans 5 Cultural practrces for reducrng the web bhght rncrdence and P fertrhzatron Further expenments estabhshed later rn the year wrth vanous other crops are not rncluded as yet rn the present report All nursenes rn the CPAF/Acre statron except the mulchrng expenments could not be evaluated because of the heavy web bhght rncrdence On farmers freid all nursenes were planted wrthout fertrhzer rnputs Web bhght rncrdence was mrlder than rn the expenmental statron and most of the nursenes could be harvested Unfortunately there were too many mrssrng plots due to web bhght resultrng rn hrgh coeffrcrents of vanatron and no statrstrcal srgnrfrcance 1 Results from CPAF/Acre Experrmental Statron The result of the mulch expenment usrng Velvet bean (Mucuna sp ) Puerana sp marze resrdues and nce resrdues and 6 levels of P fertrhzatron rangrng from O to 400 kg P/ha can be seen rn Tables 1 4 Puerana gave the best protectron agarnst web bhght because rts leaves drd not decompose as raprdly as those of velvet bean hence grvrng longer protectron agarnst web bhght The longar the bean plants are protected from splashes of sorl partrcles whrch carry the web bhght spores or myceha the hrgher are bean yrelds The hrghest bean yreld was obtarned from plots wrth Puerana as mulch followed by nce velvet bean and the lowest yreld was obtarned usrng marze as mulch Rrce and marze resrdue drd not decompose as raprdly as velvet bean but due to low resrdue quantrtres protectron agarnst web bhght was poor Table 1 P level o 25 50 100 200 400 Mean cv (%) The effect of s1x P levels and d1fferent velvet bean management (mulch green manure and manure m1xed w1th hme) on bean y1elds (kg/ha) Cover crop velvet bean Mulch lncorporated lncorporated Mean and hmed 194 8 839 4 777 2 603 8 102 7 952 8 983 9 679 8 229 3 918 9 775 6 641 3 177 3 940 6 1006 1 708 o 175 8 987 8 826 1 663 2 191 3 941 7 642 2 591 7 178 5 930 2 835 2 648 o 423 LSD 15%) NS Plots w1th mcorporated velvet bean produced h1gher y1elds than plots w1th velvet bean as mulch (Table 1 l whereas mcorporated Puerana or ma1ze res1dues were assoc1ated lower w1th y1elds than those usmg mulches (Table 2 and 3) The effects of mulches and of mcorporat1on of the res1dues on s01l parameters were evaluated but analyt1cal results are st1ll pendmg 8ecause of the heavy mc1dence of the web bhght no response of P fert1hzat1on was observed although the soil was very low m P (1 6 ppm P Bray lll (Tables 1 4) Web bhght dec1mated bean plants at random so that the response to P and N fert1hzat1on could not be evaluated (Table 5) The 16 segregatmg populat1ons now m F8 were y1eld tested under 1rngat10n dunng the dry season (July August) m the expenment stat10n of CPAF/Acre Canoca p1touco (Canoca w1th growth hab1t 1 l was used as local check All the populat1ons tested y1elded h1gher or at least equal to Canoca p1touco (Table 6) The number of days to flowenng was reduced from 45 (1n the Cerrados reg10n) to only 33 days under the cond1t1ons encountered at CPAF/Acre R1o Branco Web bhght 1nc1dence was relat1vely m1ld durmg the dry season (July August) us1ng furrow 1rngat1on Th1s opens the poss1b1hty to produce seed of good quahty m the dry se asan to supply the demands of the farmers m the reg1on Currently bean seeds for the State of Acre come from as far as sao Paulo 138 Table 2 Bean y1elds (kg/hal as affected by d1fferent P levels and Puerarta as mulch and green manure (mcorporated before plantmg) P level Cover crop Puerana sp Mulch 1 lncorporated 1 Mean o 1160 o 581 3 870 7 25 1256 7 843 3 1050 o 50 1210 o 792 7 1001 4 100 1290 7 796 o 1043 4 200 1314 o 818 o 1066 o 400 961 3 466 o 713 7 Mean 1198 8 716 2 957 5 cv (%) 34 5 LSD 15%) NS Table 3 Effects of s1x P levels on bean y1elds (kg/ha) when plantad on plots Wlth ma1ze as mulch or mcorporated mto the so1l Cover crop ma1ze P level Mulch 1 lncorporated 1 Mean o 308 o 385 2 346 6 25 448 7 536 7 492 7 50 395 3 378 o 386 7 100 357 3 505 3 431 3 200 365 3 292 o 328 7 ' 400 374 7 506 o 440 4 Mean 374 9 433 9 404 4 cv (%) 38 3 LSD (5%1 NS 139 Table 4 Bean y1eld (kg/ha) as affected by d1fferent P levels on plots w1th nce res1due as mulch Cover crop nce P level Y1eld lkg/hal o 834 o 25 646 7 50 749 3 100 834 o 200 906 7 400 905 3 Mean 812 3 cv (%) 144 LSD (5%) NS Table 5 The effect of N and P fert1hzat1on on bean y1eld (kg/ha) CPAF/Acre N levels P level (kg P20 5 P/hal (kg N/ha) 1 J l Mean 00 25 o 50 o 100 o 00 682 2 811 1 1080 o 500 o 768 3 25 o 770 2 428 9 775 6 422 2 599 2 50 o 1080 o 566 7 548 9 515 6 677 8 100 o 733 3 464 4 684 4 357 8 560 o CV(%) 60 2 LSD (5%) NS 2 Results from farmers f1eld The reg1onal tnal for Canoca and Pmk gram type on farmers f1eld were planted 1n March 1 994 when the ramy sea son was decllmng All advanced breedmgllnes tested 10 the reg1on showed no adaptat1on problems to the hot and hum1d growmg cond1t1ons The only d1fference observed was the number of days to flowenng wh1ch was about 1 O days shorter than 1n the southern reg1ons The average y1eld of the tested advanced bean breed1ng llnes was h1gher than the average y1eld of the reg1on Th1s may be the results of the small testmg plot s1ze wh1ch vaned between 5 to 1 O m 2 In the Canoca group only ESAL 588 y1elded less that the check (Table 7) In general the pmk advanced breedmg lmes y1elded less than the Canoca gram types Sorne lmes llke P1 and Saf1ra performed well 1n th1s reg1on but the local check Rosmha y1elded st11l h1gher than all the advanced breed1ng lmes tested 1n th1s expenment 140 (Tabla 8) Severa! released hnes from southern Braz1l were testad under hot and hum1d grow1ng cond1t1ons Soma cult1vars produced shghtly h1gher or at least equal to the Canoca local check The Canoca cult1var mult1phed 1n the Go1ílma expenment stat1on was out y1elded by the Canoca local check (Tabla 9) All breed1ng l1nes that composed the Nat1onal web bh9ht nursery were out y1elded by the Canoca local check and had poorer sconng 1n web bhght res1stance (Table 1 0) Table 6 ldent BZ9497 BZ9508 BZ9482 GX9519 BZ9503 GX9520 BZ9513 GX9521 BZ9498 BZ9502 BZ9511 GX9522 MX9526 BZ9500 MX9514 BZ9501 CARIOCA PITOCO cv (%) LSD (5%1 Y1eld (k9/ha) of 16 segregat1ng populat1on (bulk harvested) comparad to the local check Canoca P1toco (Growth hab1t 1) under 1mgated system Plantad July 7 and harvested Sept 29 1994 1n CPAF/Acre Expenment Stat1on Y1eld (kg/hal Days to Number of Web bhght Adaptat1on 1 flowermg pod f1lled at matunty 1028 4 33 o 50 60 70 1017 o 33 o 50 40 70 1006 o 33 o 60 60 70 961 6 300 60 60 70 929 6 30 o 60 40 70 927 2 35 o 60 70 60 890 2 33 o 50 50 70 886 6 32 o 60 60 70 855 2 35 o 70 70 60 806 8 30 o 50 50 70 799 o 33 o 70 50 70 683 2 33 o 60 50 70 676 4 33 o 60 70 70 636 6 33 o 70 70 80 6244 33 o 70 70 70 607 8 33 o 60 60 80 599 2 30 o 70 70 70 18 3 NS Web bilght and adaptatoon scorong 1 = Very good and 9 = dead 141 Table 7 Reg1onal y1eld tnal of Canoca gram type testad on farm m P A D Pedro Pe1xoto settlement/ Acre No ldentif1cation Yield (kg/hal 1 ESAL 588 1418 8 2 CARIOCA MG 1254 7 3 AN 910518 1178 1 4 LR 720982 1173 4 5 AN 730340 1145 3 6 AN 730038 1085 9 7 AN 910236 1060 9 8 ESAL 586 1037 5 9 ESAL 579 943 8 10 AN 910523 928 1 11 LR 720982 CP 785 9 12 AN 910234 714 1 13 AN 910522 710 9 14 BZ 3815 1 415 6 CHECK 1 CARIOCA (LOCAL CHECK) 1329 7 CV(%) 528 LSD 15%1 NS Th1s f1rst set of beans expenments show the d1ff1cult1es assoc1ated w1th the crop m the hot and hum1d reg10n where web bhght const1tutes the only 1mportant b1ot1c constramt Rondoma has part1c1pated 1n the Nat1onal Bean Evaluat1on Network for many years but up to the present no cult1var recommendat1on has been made because all the y1eld tnals des1gned for the other bean growmg reg1on 1n Braz1l could not produce rehable data A new strategy w1ll be used for y1eld testmg 1n th1s reg10n To start w1th only the best recommended cult1vars from the northern part of the Cerrado wlll be evaluated the plot s1zes wlll be mcreased to 1 00 m2 and more rephcates w1ll be mcluded sub¡ect to seed ava1lab1hty Only few advanced breedmg lmes w1ll enter the y1eld expenments wh1ch w1ll subst1tute gradually the poorest hnes of the expenment W1th th1s new strategy 1t IS expected that w1thm 2 years Acre and Rondon1a w111 be able to recommended new cult1vars In the meant1me efforts by both government mst1tut1ons and the pnvate w1ll be reqUired to develop local seed mult1phcat10n fac1ht1es 142 Table 8 Reg1onal y1eld tnal of p1nk gra1n type tested on farm 1n P A D Pe1xoto settlement/Acre No 1 ldent•f•catlon Y1eld lkg/hal 1 P1 1061 o 2 SAFIRA 1007 8 3 AN 730630 975 o 4 P13 975 o 5 RUBI 939 1 6 P77 895 3 7 P36 887 5 8 P80 854 7 9 P3 843 8 10 P99 796 9 1 1 p 71 711 o 12 FE 732325 690 7 CHECK 1 ROSINHA 1164 1 2 CARIOCA 1112 5 cv 1%1 349 LSD 15%1 NS 143 Table 9 lntroduct1on of the best cult1vars from Southern Brazll comparad to local check Canoca under farmers f1eld cond1t1on 1n P A O Pedro Pe•xoto settlement/Acre ROXO 90 JALO PREC NOVO JALO IAPAR 31 IAPAR 16 GOYTACAZES SAFIRA ldent1f1catJon CARIOCA Go1ílma A PO RE EMGOPA IAPAR 57 CARIOCA cv 1%1 LSD 15%1 CHECK 144 Y1eld lkg/hal 1330 o 1107 o 1083 o 1082 o 1043 o 1023 o 922 o 845 o 807 o 772 o 613 o 1080 o 21 8 NS Tabla 10 The Nat10nal Webbhght 1n nursery w1th 41 entnes comparad to local checks testad under farmers cond1t1ons Number 1 ldent1flcat1on Y1eld (kglhal 1 Web bhght' 1 OURO NEGRO 775 6 7 2 MUS80 766 7 7 3 MUS49 753 3 8 4 MUS 52 693 3 7 5 8AT 76 624 4 8 6 ICTA 8126 6044 7 7 RAB 27 600 o 7 8 MUS 23 591 1 7 9 XAN 112 584 4 8 10 MUS3 582 2 8 11 BAT 67 526 7 8 12 MUS 9 524 4 8 13 TALAMANCA 522 2 8 14 BAT 1579 517 8 8 15 BAT 1449 513 3 8 16 ICT A T AMAZULAPA 502 2 7 17 MUS 21 497 8 8 18 MUS 50 491 1 8 19 LM 21303 491 1 8 20 BAT 1155 488 9 7 21 MUS 51 475 6 8 22 ICTA 8131 475 6 8 23 A 237 471 1 8 24 S 630 B 468 9 8 25 RAI70 466 7 8 26 MUS 19 451 1 8 27 TURRIALBA 1 448 9 8 145 Number T ldent1f1cat•on Y1eld (kg/hal 1 Web bbght1 28 ESPARZA 9 446 7 8 29 EX RICO 23 442 2 8 30 BAT 64 442 2 8 31 ICTA QUETZAL 440 o 8 32 BAT 1297 4400 8 33 PORRILLO 70 428 9 8 34 MUS 26 426 7 8 35 MUS48 4244 8 36 MUS6 4200 8 37 MUS 20 417 8 8 38 ICTA 8164 417 8 8 39 RAB 34 415 6 8 40 MUS 71 397 8 8 41 XAN 90 368 9 8 CHECK 1 ROSHINA G 2 742 2 7 2 DIAMANTE NEGRO 497 8 8 3 EMGOPA OURO 471 1 8 4 A PO RE 448 9 8 5 CARIOCA LOCAL 1000 o 8 CV(%1 35 o LSD (5%1 NS 1 Webbhght score 1 = very good and 9 = dead 146 CHAPTER V PROJECTS TL 02 ANO TC 02 MECHANISTIC UNOERSTANOING ANO MOOELS OF SOIL CHEMICAL PHYSICAL ANO BIOLOGICAL PROCESSES IN AGROPASTORAL ANO SEOUENTIAL CROP PROOUCTION SYSTEMS The purpose of pro¡ects TL02 and TC02 1s to develop through the use of models a mechamst1c understand10g of s01l chem1cal phys1cal and b10log1cal processes 10 agropastoral and sequent1al cropp10g systems 1n the llanos of Colombia and cerrados of Braz1l The act1v1t1es pursued 10clude the establishment of two long term expenments on stat10n 10 both Colombia (w1th CORPOICA) and 1n Braz1l (w1th EMBRAPA CPAC) NARS partners are actlvely 1nvolved 10 the development planmng and execut1on of these expenments Crop product1on fert1llzer opt1m1zat1on nutnent cycllng soll propert1es and weed 10festat1on form the bas1s of the shared actiVItles w1th10 these expenments Work of a more strateg1c nature 1s conducted w1th10 these expenments and also w1th10 satelllte expenments and 1s a1med at 1mprov10g our understand10g of key processes for susta10able product1on and develop1ng 10d1cators of susta10ab1llty Two approaches are used for susta10ab1llty 10d1cators one 1s a comparat1ve assessment determ1010g and companng the performance of one system w1th alternat1ves and the second approach 1s a dynam1c one 1n wh1ch the dynam1cs of the system are a measure of 1ts susta10ab1llty e g measunng sorne attnbute of so1l quallty over t1me The nat1ve savannas are used as controls or benchmark s1tes for companson Wlth cropp10g and agropastoral systems Here we report program act1v1t1es undertaken w1th10 the long term expenments together w1th sorne key process stud1es 10clud10g nutnent cycllng v1a lltter crop res1dues cattle excreta and m1crob1al b1omass phosphorus transformat1ons s01l phys1cal propert1es the dynam1cs of soll fauna populat1ons and act1v1t1es LONG TERM EXPERIMENT 1 SUSTAINABLE CROP ROTA TION ANO LEY FARMING SYSTEMS FOR THE ACID SOIL SAVANNAS OF COLOMBIA (CULTICORE EXPERIMENT) O Fnesen fiFDCJ H Carmen H Delgado E Owen (CORPOICAJ R Thomas M F1sher A G1Jsman 1 Rao 8 Volverás H F Alarcón (CIA T TLJJ G R1ppstem (CIA T/CIRAD EMVTJ P Lave/le (ORSTOM/TSBF/Umv PansJ A Moreno J J J1ménez (Umv Complutense MadndJ lntens1f1cat1on of agncultural product1on on the ac1d s01l savannas of Lat10 Amenca 1s constra10ed by the lack of d1vers1ty 1n ac1d (alum10um) tolerant germplasm and poor so1l fert1llty The use of h1gh levels of 10puts espec1ally 10 monocropp10g s1tuat1ons 1s thought to be unsusta10able s10ce 1t results 10 detenorat10n of so1l phys1cal propert1es as well as escalat1on of pest problems lmproved legume based pastures cons1dered by many as least harmful to the soll resource base reqUire 10vestments 1n 10puts for establishment wh1ch are unattract1ve or beyond the means of extens1ve graz1ers Establishment of pastures 1n assoc1at1on w1th nce (to defray the cost of 1nputs) has proven to be a very attract1ve alternat1ve wh1ch IS rap1dly be10g adopted 10 front1er areas of the Colombmn llanos However as farmers see the prof1ts to be made from nce th1s development could easlly detenorate to one of cont10uous monocropp10g w1th potent1al problems of land degradat1on such as those already observed w1th monocropp10g 1n the Braz11ian cerrados Alternat1ve systems mcorporatmg components wh1ch attenuate or reverse the deletenous effects of monocultures are reqUired and b1ophys1cal measures of sustamab11ity need to be developed as pred1ctors of system performance and health Gram legumes green manures mtercrops and leys are poss1ble components wh1ch could mcrease the stab11ity of systems mvolvmg annual crops The long term expenment descnbed here mvest1gates a select1on of alternat1ves usmg these components at two levels of mtens1f1cat10n based on lime and through 1ntens1ve momtonng under controlled cond1t1ons to 1dent1fy md1cators of susta1nab11ity of rotat1onal and agropastoral systems 1n the ac1d savannas Smce many processes contnbute to and mteract m determm1ng the stab11ity of any part1cular system a pnnc1pal ob¡ect1ve of the pro¡ect 1s the development of mtegrated models wh1ch would s1mulate the effects of system components and management on system sustamab11ity as exemplif1ed by the 1dent1f1ed pred1ctors and enable the evaluat1on and extrapolat1on 1n space and t1me of the effect of components and management perturbat1ons on system stab11ity Recogmzmg that the deletenous (or benef1c1al) effects of vanous agncultural pract1ces are often subtle and only mamfest themselves over long penods the proposed expenmentat1on 1s mtended to extend through at least two rotat1onal cycles MATERIALS ANO METHODS The expenment was established on a loamy clay Ox1sol on the Corp01ca/CIA T Research Stat1on at Canmagua on the eastern plams of Colombia Typ1cal of Ox1sols the so1l 1s charactenzed by a h1gh degree of exchangeable Al saturat1on and low pH and a very low level of ava1lable nutnents (Table 3) The bas1c approach mvolves a companson of product1on systems rangmg from potent1ally resource degradmg cereal monocultures to the hypothet1cally resource 1mprov1ng agropastoral ley systems all at two levels of 1nputs The select10n of systems was based on whether lime 1s applied as a ferti11zer (to supply Ca and Mg to Al tolerant crop and pasture spec1es) or as a so1l ac1d1ty ameliorant (to enable product1on of more Al sens1t1ve spec1es) In the fert11izer lime systems upland nce 1s grown m cont1nuous monoculture or 1n rotat1ons w1th green manures cowpeas or adapted m1xed pastures wh1le the remed1al lime systems mvolve ma1ze 1n contmuous monoculture or 1n rotatlons w1th green manures soybeans or less adapted m1xed pastures A descnpt1on of the treatments 1s g1ven 1n Table 1 The expenment 1s a split plot des1gn m four random1zed blocks 1n wh1ch levels of lime are ass1gned to ma1n plots and systems to subplots Plot s1zes (0 36 ha or for pastures O 72 ha) were based on the total area reqUired to support a mm1mum of three ammals stocked at approx1mately two ammals per hectare when rotated through the replicat1ons after prov1s1on for d1v1d1ng at a later date D1mens1ons were chosen for ease of handling convent1onal machmery and w1th a v1ew to a probable need to d1v1de the plots 1n the futura All systems are managed to opt1m1ze product1on and mm1m1ze s01l degradat1on that 1s crop res1due conservat1on 1s pract1ced so1l fert11ity levels are ma1ntamed weeds and other pests controlled etc The expenment has been Implementad 1n two stages rotat1ons based on nce and low mputs of lime were 1n1t1ated 1n M ay 1 993 wh1le rotat1ons based on ma1ze and h1gher levels of lime were 1n1t1ated 1n Apnl 1994 Sequent1al measurements and observat10ns are bemg made of cr1t1cal so1l phys1cal chem1cal and b1olog1cal propert1es and the 1mpact of changes 1n them on crop/system product1v1ty and prof1tab11ity resource and mput use eff1c1ency and the env~ronment Among the observat1ons and measurements currently bemg recordad are the followmg 148 Table 1 Treatment descnpt1on F1rst agropastoral cycle (f1ve years) Treatment Mam Plots Sub plots Descr1pbon No 1 Fert•llzer lime R1ce monoculture R1ce grown '" monoculture one crop per year 1n the flrst semester second semester weedy fallow turnad 1n wrth early land prepn at end of ra1ny season 2 R•ce cowpea rotat1on R1ce (1st semesterl and cowpea (2nd semesterl•n 1 year rotat1on res1dues 1ncorporated pnor to plant1ng 1n followmg season 3 R1ce green manure rotat10n R1ce (1st semester) and green manure (2nd semester) '" 1 year rotat1on Legumes 1ncorporated at max1mum standing b1omass levels 1n late ramy season 4 Nat1ve savanna lspare plotl Managed trad1t1onally by burnmg annually durmg dry season 5 R•ce agropastoral rotat1on Brach1ana humidlco/a 1 Centrosema acutlfoi/Um 1 Sty/osanthes cap1tata 1 Arach1s pmto1 cocktail sown w•th nce '" year 1 grazed to mamtam legume content rotated every 4 or 5 years depend1ng on pastura compos1t1on ~ 6 Remed1al L1me Ma1ze monoculture Ma1ze grown '" monoculture one crop per year 1n the flrst semester second .¡. "' semester weedy fallow turnad '" w1th early land prepn at end of ra1ny season 7 Ma•ze soybean rotat1on Ma1ze (1st semester) and soybean (2nd semester) 1n 1 year rotat1on res1dues mcorporated pnor to plant1ng 1n follow1ng season 8 Ma1ze green manure rotat1on Ma•ze (1st semesterl and green manure (2nd semesterl 1n 1 year rotat1on Legumes 1ncorporated at max1mum stand1ng b1omass levels 1n late ra1ny season 9 Nat1ve savanna (spare plotl Managed trad1t1onally by burmng annually dunng dry season 9A Ma1ze soybean rotat•on (no t1ll) Ma1ze (1st semesterl and soybean (2nd semester)m 1 year rotat1on tlllage only to lnltlate croppmg on nat1ve savanna res1dues left on sml surface (no mcorporat•on) 10 Ma1ze agropastoral rotat1on Ma1ze monocrop m year 1 Pamcum max1mum 1 Glycme w1ghtu 1 Arach1s pmto1 pastura sown w1th nce 1n year 2 grazed to ma1nta•n legume content rotated every 4 or 5 years depend•ng on pastura compos1t10n System and component product1on Gra1n y1eld and above ground b1omass product1on are determinad for crops and green manures Standing biomass feed on offer and botamcal composition of pasturas are assessed at three monthly mtervals while root biomass production and distr~bution are estimated on an annual bas1s Records of ammal stockmg rates and livewe1ght gams (3 month mtervals) are also mamtamed In add1t1on the nutr~ent compos1t1on of crop and pasture components and green manures 1s measured to est1mate nutr~ent exports and recycling through res1dues Soil phys1cal propert1es Parameters wh1ch are likely to reflect soil phys1cal degradat1on as a result of agr~cultural pract1ces are momtored annually These mclude bulk dens1ty poros1ty and water mf1ltrat1on rate hydraulic conduct1v1ty and aggregate stab11ity Soil chem1cal propertles Parameters are measured wh1ch are 1nd1cat1ve of changmg s01l fert11ity loss of SOil orgamc matter and hence so1l structural stabi11ty and wh1ch perm1t compar1sons of the relat1ve rates of nutrlent cycling m the d1fferent systems These include s01l ac1d1ty and exchangeable cat1ons and so1l N P and S pools accordmg to published fractionatlon procedures Soil b1ology Populat1ons and dynam1cs of earthworms are momtored RESUL TS ANO DISCUSSION Crop and pastura production and establishment D Fr1esen (/FDCJ H Carmen H Delgado (CORPOICAJ Dueto logiStlcal d1ff1cult1es the low lime mput r1ce basad systems were 1mplemented m 1993 wh1le the h1gh lime mput ma1ze based systems were 1mt1ated m 1994 Average r1ce gram y1eld m the first year was about 3 t ha 1 and was not s1gmf1cantly affected by undersowmg 1t w1th pastura (Table 21 Nevertheless forage spec1es were well established at harvest and pasturas were put to grazmg approximately 3 ~ months later at the onset of the dry season (December 1 993) Durmg the subsequent 3 months ammals stocked at 3 ha 1 gamed an average of 333 g day 1 livewe1ght Without placmg undue pressure on the pasturas wh1ch emerged for the dry season dommant 1n S cap1tata Consequently the stocking rate was mcreased to 4 ha 1 m June 1994 to try to reduce the amount of standing b1omass and encourage a more balanced grass legume compos1t1on Rice m the low mput rotation systems was followed 1n the second semester of 1993 by an erect cowpea var1ety (ICA Cabec1ta negra) for gram product1on (Treatment 2 Table 11 and a prostrate cowpea var~ety OCA Menegual for green manure (Treatment 3) The green manure was mcorporated at flower~ng 52 days alter plantmg Cowpea gram yield was approx1mately 1 1 t ha 1 wh1le b1omass (res1duesl turnad into the so1l was approx1mately 1 6 t ha 1 for both the gram crop and the green manure (Tabla 2) Both res1dues and green manure hada marked tmpact on mmeral N levels m the so1l profile dur~ng the followmg dry season (see subsequent sect1on) Pr~or croppmg w1th cowpea whether for gram or green manure product1on had a SIQnlflcant impact on the product1on of r1ce m rotat1on treatments m 1994 (Table 2) R1ce gram yields were 300 750 kg ha 1 greater than m r1ce monoculture Part of th1s may be explamed by the improved N economy m the legume rotat1on treatments alluded to above Add1t1onally weed populations were much lower 1n the rotat1ons than m the monocropped r1ce treatments largely 150 because the formar treatments were culttvated and cover cropped dunng the second semester of the prevtous ratny season whtle the nce monocrop treatment was left as a weedy fallow The matze based systems were tmttated wtth the actd sotl tolerant matenal (Stkuam 3) recently released by CIMMYT Gra1n ytelds (Table 2) were low at approxtmately 2 5 t ha 1 approxtmately 1 t ha 1 less than the best obtatned 1n satelhte expenments adJacent to thts expenment Thts dtsappotntlng performance can probably be attnbuted to the poor and uneven stand obtatned as a result of dtfftculttes expenenced wtth the machtnery at sowtng Moreover unhke the expenence wtth nce establishment 1n the precedtng year where weeds were not a problem weed populattons prohferated 1n the matze plots espectally dunng the later growth stages We speculate that thts was largely the result of the ttllage operattons applied 1n the prevtous year dunng whtch the enttre expenment was ploughed 1n preparatton for an anttctpated complete tmplementatton 1n 1993 Addtttonally the wtde row spac~ng for matze may not have provtded sufftctent competttton to weeds early 1n the season Soll chemtstry/ferttlity The unamended Oxtsol under nattve savanna at the expenment stte 1n charactenzed by a htgh exchangeable Al saturatton low pH and low contents of avatlable nutnents (Table 3) However exchangeable Al decreases and sotl pH tncreases wtth depth 1n the proflle below 40 cm Applicatton of 500 kg ha 1 of dolomtttc lime pnor to tmplementatton of the nce based systems 1n Apnl 1993 dtd not affect exchangeable Al or soll pH (measured 4 months after applicatton after the nce harvest) but dtd reduce Al saturatton by about 8% as a result of tncreased exchangeable Ca and Mg denved from the lime An addtttonal applicatton of 200 kg ha 1 1n March 1994 befare sowtng causad a further 8% reductton 1n exchangeable Al saturatton (measured 3 weeks later) There was no stgmftcant effect of these amendments on soll properttes at depth Phosphorus ferttlizer applicattons (60 kg P ha 1 to nce and 40 kg P ha 1 to cowpea tn 1993) tncreased avatlable P from 1 5 to 6 O ppm as measured by Bray 11 pnor to sowtng 1n 1994 Applicatton of 2 t ha 1 of dolomtte to the matze based systems reduced Al saturatton to approxtmately 45% wtthtn 3 weeks (Table 3) Although exchangeable Al decreased and exchangeable Ca and Mg tncreased s01l pH was only margtnally affected Levels of tnorganic N tn crop pastura and rotattonal cropptng systems D Fnesen (IFDCJ R Thomas (TLJ As part of the efforts to determine nutnent balances and recycling 1n the expenment levels of tnorgamc N have been momtored tn the sotl proftle of selected treatments dunng the year for use 1n crop models and as an assessment of nutnent use efftctency Treatments tncluded cowpea as a gratn crop and as a green manure sown 1n the second semester after nce The green manure was tncorporated tnto the s01l 52 days after sowtng at a rate of 1 5 t DM ha 1 For the gratn cowpea around 1 7 t DM ha 1 was returned to the sotl after the harvest dates of Jan 56 1994 Sotl N levels were measured 1n the monocrop nce nce cowpea green manure and savanna control 1n January 1994 Levels were low for the savanna treatment and slightly greater for sotl under monocropped nce whtch had a marked tncrease 1n mtrate N only at a depth of 80 100 cm (Ftgure 11 Levels of ammomum N and espectally mtrate N were greatest tn the nce cowpea green manure treatment at all depths wtth a total of about 100 kg mtrate N ha (Ftgure 11 A stmtlar trend 1n sotl N levels was observad 1n early March 1994 wtth a marked tncrease 1n mtrate N 1n the nce cowpea green manure treatment of up to 135 kg mtrate N ha (Ftgure 151 2) Table 2 Gram and dry matter product1on m monocultures and rotat10ns m the long term cropp1ng systems tnal ( Cult1core ) at Canmagua 1993 1994 Cultural System R1ce Cowpea Dry Dry Gram matter Gram matter R1ce Ma1ze k1lograms per hectare Monoculture 2470 Cereai/Gram legume 2850 40201 1130 1720 2780 23301 rotat1on Cereai/Green manure 1540 3120 rotat10n Agropastoral rotat1on 3140 4270 2570 treatments harvested as one umt per repllcate In late March levels of s01l N m savanna and nce pasture treatments remamed low but m all other treatments elevated levels of mtrate N were observad espec1ally 1n the nce cowpea green manure treatment where up to 177 kg n1trate N ha was measured 1n the 1 m so1l prof1le (F1gure 3) S1m1larly about 113 kg n1trate N ha was measured m plots where cowpea had been grown for gram and where crop res1dues had been 1ncorporated (Figure 3) S01l wh1ch had been plowed 1n preparat1on for ma1ze sowmg also had elevated levels of both ammon1um N and espec1ally mtrate N (up to 70 kg mtrate N ha) comparad w1th und1sturbed savanna s01l (Figure 3) Th1s result md1cates an 1ncreased rate of N mmerahzat1on and accumulat1on of n1trate followmg s01l d1sturbance wh1ch IS m add1t1on to the mtrate denved from cowpea green manure and cowpea res1dues mcorporated after gram harvest Data from September 1 994 after a second success1ve nce crop md1cate that n1trate was also present m the nce monoculture treatment after the second nce crop but at depths between 60 100 cm (F1gure 4) A s1m1lar pattern was observad for the nce cowpea treatments and treatments where ma1ze had been grown as the f1rst crop m the h1gher mput system (F1gure 4) In companson w1th the March data S1gn1f1cant amounts of n1trate appeared lower down the prof1le at a depth of 60 100 cm 1n September comparad w1th O 40 cm m the monocropped nce treatment The source of th1s N IS not known but could anse e1ther from fert1hzer apphed to the second nce crop or as a result of further m1nerahzat1on of so1l orgamc matter followmg cult1vat1on and land preparat10n procedures Unfortunately no samples were taken from the und1sturbed savanna control wh1ch m ay have g1ven an 1nd1cat10n of the effect of s01l d1sturbance on N accumulat1on 152 Table 3 lnfluence of hme and fert1hzer mputs on so1l chem1cal propert1es m nce and ma1ze based systems Depth Org e Total N Bray P pH Exchangeable cat1ons meq/1 OOg Al satn System (e mi (%) (%) (ppm) (H20) Al H Ca Mg K ECEC (%) Nat1ve savanna 010 3 24 014 1 5 4 51 2 35 o 31 o 25 o 11 o 09 3 11 75 6 10 20 2 55 o 10 1 8 4 56 2 20 o 29 014 006 004 2 73 80 5 20 40 1 2 4 80 1 86 o 07 o 11 o 04 o 04 2 12 87 4 40 60 1 o 4 85 1 32 007 o 11 004 o 04 1 57 84 1 60 80 09 5 13 o 93 o 04 o 11 004 003 1 15 81 1 80 100 1 o 5 22 o 46 o 03 o 12 o 03 003 o 68 68 5 R1ce systems 1 010 3 31 o 15 4 30 2 31 o 12 o 55 o 28 012 3 39 68 o (Sep 1993) 10 20 2 57 o 10 4 32 244 o 12 019 o 08 005 2 88 84 6 ~ "' R1ce systems2 o 10 3 08 67 4 18 2 00 o 24 o 56 o 37 o 19 3 36 59 6 "' (March 1994) 10 20 2 33 1 4 4 16 2 12 o 22 o 25 o 14 009 2 82 75 1 Ma1ze systems' o 10 313 20 4 44 1 54 o 19 1 06 o 62 o 11 3 52 43 9 (March 1994) 10 20 2 35 1 2 4 23 2 21 o 25 o 23 010 006 2 86 774 1 alter flrst semester nce (hmed m Apnl 1993 w1th 500 kg/ha dolom1te and fert1hzed) 2before second year sowmg (re hmed w1th 200 kg/ha dolom1te 3 weeks prev1ously) 3before f~rst year sow1ng (hmed w1th 2000 kg/ha dolom1te 3 weeks prev1ously) li! • i ~ ~- !! JI !i! J § ~ ~ z = ~ :¡: ;¡¡¡¡ -= ~ o o !i! ~ : ~ 8 ;¡ ~ $ = '""" ~ - li 1 e: (J) E il ¡: ~ (J) 2 a. JI X !i! (J) l Q¡ L .... r ~ o o ::: ::::1 o (.) o ~ % ! 1 8 a ;; ¡ e: = '""" 1/l (J) > (J) li! z i 2 e: as li C> ~ .... !! o e: :1! JI !i! o ~ ~ en " r~ ~ ..-J;L! Q> u. o o lil ~ ! z 8 ;; ~ $ = '""" 154 il !ª 1i - :1 e: 11 CD E ;::: V gi CD i a. E X 1 CD .. T l' a I~ CD .... o o a ~ ~ ~ ~ 8 a .:!:: ;;; ¡ ::J w .- (.) a e: 1/) lil a; > ..9? i z J o e: 11 cu C) .... gi o e: r; o lt l' UJ h C\1 ' C) ¡ ~ ~ 8 a ¡¡: ¡ = .- 155 1 1 1 ti 1 R- R i I JI t ~ lil lil 1 l J !! !'! ¡ J 1 8 o o o ~ ~ 1 o ~ ~ 1 8 ó ¡ ó ¡ ... ...., e w=») paptpu ,.,Jet -e: Cl) il E ¡:: Cl) 1 1 Q. )( 1 1 Cl) = R- R- Cl) .... 1 i o JI o lil lil .:2 :J () = !'! J !'! e: 1 ~ § -! !11 o o Cl) !'! ~ ~ 1 1 8 o ~ ~ ~ 1 ~ > o ¡ ó 8 Cl) ... ......, .......... z 2 il il e: ca C) .... 1 ~ o e: 1 1 R- o R 1 i en lili Jl (1) lil C) f u: !'! J !'! f 1 1 i i o o o ~ ~ ' 1 ~ o ~ ~ 1 8 ó ó ¡ 8 -- ~- 156 ~ "' ..... ,... __ D-101 if .... 10.20: ¡zo-co ! ~ ! -1 80-1001 l. o 10 -- D-10 -10.20 o 10 ·-- 20 30 ..... 1 40 110 ·-- 1 1 20 30 40 110 kgNha 1 ,...._.. s..,temb e2IM D-101 ~ ""' 11).20 ~ !- - eo-aol + 110-100I..L. o 10 20 30 40 80 eo ...... 1 Malz.tofbHn(vr-t m.IUJ ) Sept.mbfi2/M •"llf-10.20 20-<40 5 ""' !«~-«) eo-ao SD-1001"' • o 10 20 30 40 110 ... 1 F1g 4 So1l morgamc N levels 1n Cult1core" expenment flce.co...- (green miRUI' ) -- 1 5 = ! 1 70 o 10 20 30 40 80 110 ... 1 ... .._ - 1 D-10 10.20 E 20-<40 !- 1 eo-ao ' 1 BD-100 110 10 o 10 20 30 40 80 .... 1 The data md1cate the accumulat10n and potent1alleakage of large amounts of mtrate N from the croppmg systems Th1s N may be recovered 1f a pasture 1s sown s1multaneously w1th the crop thus md1catmg a more eff1c1ent use of the fert1hzer Further momtonng at more frequent mtervals 1s planned for 1994/95 to ga1n a better understandmg of the N economy of crop and crop pasture rotat1onal systems Decompos1t1on of crop res1dues R Thomas Decompos1t1on of the crop res1dues m the expenment 1s bemg measured w1th ma1ze nce soybean and cowpea The latter two are also bemg used as green manures m some treatments Prev1ous mformat10n 1s ava1lable for the decompos1t1on rates of ma1ze (79 day half hfe) nce (71 day half hfel and cassava (68 day half hfe) and nutnent release patterns (B1annual report Savanna program 1992 93) Rates of nce stubble decompos1t10n were more than tw1ce as great when the plant matenal was mcorporated m htter bags m the so1l at a depth of 5 1 O cm comparad w1th bags left on the s01l surface (Table 4) R1ce stubble left on the s01l surface decomposes at a rata (0 0053 d '1 s1m1lar to that of e1ther roots of the nat1ve savanna grasses (0 0056 d '1 or a mixture of roots and above ground htter (0 0059 d ') The nce stubble m these expenments was placed 1n htter bags of two mesh s1zes 1 and 4 mm but the rates of decompos1t1on were 1dent1cal w1th no effect of mesh s1ze Therefore the data m Table 4 are presentad as means of the mesh s1ze treatments lncorporated cowpea green manure decomposed at rates greater than mcorporated nce stubble (Table 41 and although the hgmn N rat1os were s1m1lar for cowpea (5 2) comparad w1th nce stubble (6 8) other nutnent concentrat1ons such as N and P were much greater 1n cowpea comparad w1th nce (2 8% N O 2% P m cowpe'a comparad w1th O 92% N and O 08% P m nce) Stud1es on the decompos1t1on of the green manure treatments (soybean and cowpea) are contmumg m 1994 but results are not yet ava1lable Further d1scuss1on of these f1ndmgs appears below under recychng of nutnents v1a decompos1t1on of htter and crop res1due Table 4 Rates of decompos1t1on of crop res1dues Spec1es Treatment K half hfe day' days R1ce Superf1c1al o 0054 ±o 0016 104 (nce monoculture) R1ce Superf1c1al o 0051 ±o 0017 108 (nce pasture) R1ce lncorporated o 0131 ±o 0021 48 (nce monoculture) Cowpea lncorporated o 0186 ±o 0039 32 (green manure) Mean of htter bag mesh s1zes 1 and 4 mm 158 Effect of cropp.ng systems on earthworm populat1ons and d1str1but•on J J Jtménez M Ftsher D1stnbut1on of earthworms 1n the nat1ve savanna buffer area 1n the expenment was evaluated dunng November 1993 and found to have the same charactenst1cs as 1n the Brachtana decumbens based pastures(see belowl Further measurements are bemg taken to document the spat1al d1stnbut1on of the earthworm commumty and to follow recolomzat1on followmg cult1vat1on w1th the ob¡ect1ve of determmtng the mechamsms To ach1eve th1s the followmg contrastmg treatments have been chosen and are sampled each 3 4 months monoculture nce R1ce pastura rotat1on and corn pastura rotat1on The samples on each occas1on are taken at the same pomts on a gnd layout to determme the dynam1cs of the earthworm commumty across the d1fferent systems and the effects of seasonal vanat1ons These data have been not yet processed but observat1ons at the f1eld s1te 1nd1cate a very h1gh reduct1on of earthworm commumt1es wh1ch w11i make demonstrat1on of spat1al changes d1ff1cult Long term expenment 2 Agropastoral systems for the cerrados of Brazll M A yarza (CIA TJ L Vtlela J C Mtranda A Cardoso (EMBRAPA CPAC) H Neufeldt R Westerhof W Zech (Bayreuth Umverstty) J Duxbury (Come// Umverstty) lntroduct1on The charactenzat10n work conducted by CIA T and EMBRAPA dunng the last two years h1ghhghted the need to halt or revert so1l degradat1on and dechmng s01l product1v1ty of the current product1on systems 1n the cerrados Among ava1lable technolog1es to mcrease product1v1ty and recla1m so1l fert1hty the combmat1on of crops and pasturas 1n space and t1me IS one of the best opt1ons Th1s technology not only mcraases overall product1v1ty but also contnbutes to 1mproved soc1oeconom1c cond1t1ons of farmars Management practicas such as land preparat1on hmmg and fert1hzat1on and ammal managament mfluance the potant1al contnbut1on of crops and pasturas 1n mtagrated crop pastura product1on systems In order to measure these effects a long term expenment was estabhshed 1n a red dark Latossol at EMBRAPA CPAC 1n Planaltma Braz1l1n 1991 The ob¡ect1ve of the expenment 1s to determine the effect of 1ntegrat1on on crop and ammal product1v1ty and to 1dent1fy so1l key parameters related to 1mprovement or degradat1on The expenment mcludes contmuous crop and pastura systems and crop pastura and pastura crop rotat1ons cycles of f1ve years Land preparat1on methods evaluate the effect of d1skmg (convent1onal land preparat1on) and deeper land preparat1on methods (flexible land preparat1on method) on crop product1v1ty and so1l phys1cal propert1es Fert1hzat1on treatmants study the effact of convent1onal fert1hzat10n and a correct1ve fert1hzat1on tncludmg the use of gypsum on crop and pastura product1v1ty and so1l fert1hty Grazmg managament treatments evaluata the effect of low and h1gh grazmg 1ntens1t1es on ammal product1on botamcal compos1t1on and htter product1on 1n grass only and grass legume pasturas A group of researchers from CIAT CPAC and the Umvers1t1es of Bayreuth (Germany) and Cornell (USA) IS measunng the short and long term trends on crop and ammal product1v1ty and so1l fert1hty 1n the expenment 159 The on stat1on work 1s complementad by satelhte plots located m several farms m Uberland1a These plots were estabhshed m 1992 on sandy and clay so1l types to determ1ne the potent1al contnbut10n of forage legumes to 1mproved sustamab1hty of crop pasture systems m hvestock product1on systems and m mtens1ve management systems w1th annual crops Changes m ammal product1on and so1l fert1hty m 1mproved and unrecla1med systems are bemg mon1tored over t1me Below are reported the results of the work On Stat1on research 1 Crop performance L VJ/ela and M A yarza A fter three consecut1ve crops results show that so1l fert1hty management rather than land preparat1on methods 1s havmg a greater mfluence on crop y1elds Correct1ve fert1hzat1on 1ncreased corn y1elds by 30% compared to the convent1onal fert1hzat10n treatment dunng the present year Th1s trend was s1m11ar to the observed w1th soybeans last year (Figure 5) Above ground corn b1omass also mcreased Slgmflcatlvely w1th fert1hzat10n (Table 5) The enhanced crop growth obtamed m th1s treatment reduced weed competltlon wh1ch affected gram product1on 1n the convent1onal fert1hty management system Penmsetum ped1cellatum and Acantospermun australe were the dom1nant weed spec1es found 1n th1s system These weeds are common m annual croppmg systems m the cerrados Table 5 Effect of s01l fert1hty and land preparat1on methods on performance of corn and weed b1omass m the crop pasture expenment Fert1hty Land preparat1on Gram (kg/ha) F1 F1 F2 F2 Convent10nal Flexible Mean Convent1onal Flexible Mean 5285 89 5289 90 5287 90 7326 65 7199 10 7262 87 1/biomass determmed at flowermg time 2/blomass determmed at corn harvestmg t1me F 1 Low fert1hty F2 H1gh fert1hty Convent1onal D1sk harrow land preparat10n Flexible Deep plowmg land preparat1on Corn b1omass 1 (kg/hal 4577 03 4623 54 4600 28 6545 55 6551 40 6548 47 Weed b1omass 2 (kg/ha) 2549 01 1972 00 2260 50 1322 99 964 32 1143 65 Although there were no d1fferences m gram y1eld between the tradltlonalland management system (d1skmg) and the flexible management system (plowmg) there were sorne d1fferences 1n the d1stnbut1on of corn roots w1th depth and on s01l phys1cal propert1es Measurements camed out down the O 60 cm s01l profile m several p1ts dug m the expenment showed an mcreasmg mechamcal res1stance w1th depth m both land preparat1on methods (Figure 6) However 1t was greater m the O 25 cm depth of the s01l prepared w1th a d1sk (Figure 6) The use of a plough m the flex1ble land preparat10n method enhanced root growth and d1stnbut1on 160 1n the 7 25 cm (Figure 7) D1fferences 1n management system due to the amehorat1ng effect of gypsum on subso1l ac1d1ty and calc1um def1c1ency Although weed b1omass was lower w1th the use of the plough d1fferences w1th respect to the use of the d1sk were not stat1st1cally s1gmf1cant due to the h1gh vanabll1ty 10 weed populat1on 1n the expenment 2 Ammal peñormance L Vtle/a and M A yarza The supenor ammal performance 1n the grass legume assoc1at1on observad dunng the f1rst year of grazmg remamed the same dunng the second year (Figure 8) Stylosanthes mmetrao st1ll contnbuted s1gmf1cantly to the total ava1lable forage b1omass ( > 60%) and to the n1trogen supply to the grass v1a htter decompos1t1on Th1s resultad 1n a h1gher mtrogen content 1n t1ssue of A gayanus 10 the grass legume pastura comparad to the grass only pastura ( 1 59 + O 2 vs 1 13 +0 01) Mamtenance fert1hzat1on and grazmg 1ntens1ty mfluenced more botamcal compos1t1on than ammal performance L1tter product1on data w1ll be reportad next year A marked mortahty of Stylosanthes plants has been observad smce the last dry season (Apnl 1994) The results of the evaluat1on of damage carned out at the end of the dry season 1nd1cated that 50% of the standmg plants showed symptoms of a d1e back d1sease or mechamcal damage produced by ammal tramphng Damage was lesser 1n Sty/osanthes plants growmg 10 1nt1mate assoc1at1on w1th the grass Th1s problem 1s reducmg legume b1omass rap1dly 10 the expenment Therefore we delayed the 1mt1at10n of grazmg 10 the wet season and reduced grazmg pressure as an attempt to mamtam the legume component 1n the expenment Sol! phys1cal propert1es under pasturas followed the same pattern observad last year Mechamcal res1stance values 10 the O 1 O cm topsoll of the grazed pasturas were h1gher comparad to the cropp1ng systems as a result of to ammal tramphng (Figure 9) However absoluta values were below 12 mpascals that have been found to hm1t root growth accordmg to prev1ous work conducted at CPAC (Tropsolls Bulletm No 91 01) Roots of Andropogon were more evenly d1stnbuted 10 the s01l prof1le than the corn roots and the nat1ve cerrado vegetat1on (F1gure 1 0) Changas 10 so1l chem1cal propert1es are not reportad smce laboratory analys1s results are not ava1lable yet 3 MycorrhiZ8 populatlon and actiVIty J e Mtranda The a1m of th1s work IS to evaluate the effect of crops and pasturas over t1me on mycorrh1za populat1on and act1v1ty Propagule dens1ty ( number of spores/ 50 gm so11) and act1v1ty (percentage of roots 1nfected) 1s under evaluat1on smce the begmmng of the expenment So1l samples from the O 20 cm sol! depth conta101ng fresh roots are collected three t1mes each year (at the begmmng at the peak and at the end of the ramy season) Bes1des spore countmg a prehmmary assessment of most dommant mycorrh1zal genus 1s done 1n each management system Nat1ve populat1on 1s low ( 12 spores/ 50 gm s011) and composed of several genera (Gtgaspora Scutellospora Acau/ospora and Glomus) The results of the f~rst year of evaluat1on 1nd1cated that mycorrh1zal populat10ns mcreased by twenty f1ve fold dunng pastura establishment and by threefold under the soybean crop (F1gure 11) Root colomzat1on followed the same patterns Dunng the second year populat1on under pasturas dropped wh1le 1t mcreased under crops Changas are probably relatad to the h1gh mycorrh1zal dependence of pasturas dunng the establishment phase and to d1fferences 1n root morphology between crops and pasturas 161 ~ ~ e i! " 7 8 a b CPiowlog []Ma•ntenance fertlltzat1on 6 l:l!Disklog [¡JCorrect•ve fertdwhon 6 5 a 4 4 J 2 o U..-...J Soybean 1991 Soybean 1992 Coro 1993 Soybean 1991 Soybean 1991 Corn 1993 F1gure 5 Effect of land preparat10n and two fertJhzatton strategtes on gram ytelds ofthree crops planted consecutJvely 10 dark red latossol ofCPAC, Planaltma (Crop pasture mtegratJon expenment mean oftwo reps) 162 25 12 5 -e CJ 25 o _.. .e .... c. ~ "' - 35 o ... = OO. 49 5 57 o o 02 Cone mdex (MPa) 04 06 08 1 12 ::íS..E -e-Plowmg •D•skmg Figure 6 MechaDicalimpedance pro file of th1rd year effect of two tillage treatments mcluded m crop-pasture mtegration expenment 163 -75 15 25 35 45 55 e 65 CJ '-' - .. = 15 00 25 35 45 55 65 Mamtenance fertihzatmn D1skmg Plowmg Correchve ferhhzation ================;---, D1skmg Plowmg 50 40 30 20 10 o 10 20 30 40 50 Root distnbution (%) Ftgure 7 Coro root dtstrtbutton as functton oftwo fertthzatton strategtes and two ttllage methods (Crop-pasture mtegratton expertment) 164 120 100 rS..E - -e- Andropogon+legumes = 80 e -e- Andropogon - = ~ } 60 -~ e 40 = ~ ..... -= 20 ~ ... ~ ~ o == ~ -20 -40 M ay Jul Sep Nov Jan Mar M ay Fagure 8 L1vewe1ght ga1ns ID Andropogon gayanus alone and ID assoc1ataon wath a cocktail offorage legumes planted ID dark red latossol ofCPAC, Planaltat1a-DF (Crop-pasture IDtegrataon exper1ment) 166 Cone mdex (MPa) o 02 04 06 08 l 25 125 F1gure 9 Mecbanu~alunpedance profdes of tbree sml management systems mcluded m tbe crop-pasture expenment 168 -e Col - .e .... c. ~ "C -= = OO. 5 15 25 35 45 55 60 50 40 30 20 10 o 10 20 30 40 50 Root dtstnbution (0/o) F1gure 10 Relat1ve dJstnbuhon of corn and Andropogon gayanus roots m the O 65 cm profile (Crop-pasture mtegrat10n exper1ment) 167 350 * Grass - 300 + Grass+Legumes bl) = ..... crop ~ 250 = o NatiVe Cerrado -J. 200 41 .J:I e 150 = = 41 lOO J. o * c. 50 00 o 9/91 1/92 4/92 9/92 1/93 4/93 9/93 Samphng time F1gure 11 Var1at1on of NV A mycorrh1zal populatlons m a dark red latossol as a funct10n of the management system and samphng time ( Crop- pasto re Int exper1ment) 168 4 Soal Orgamc matter dynam1cs IComplementary Sub proJect TC511 H Neufeldt R Westerhof and W Zech Considerable attent1on 1s g1ven to the role of orgamc matter m sustamable land use systems m neotrop1cal savannas Bes1des to the benef1c1al effect on so1l structure orgamc matter 1mproves water holdmg capac1ty cat1on exchange capac1ty and soll b1olog1cal act1v1ty A collaborat1ve pro1ect between CIA T EMBRAPA CPAC and the Umvers1ty of Bayreuth was recently 1n1t1ated to study the short and long term effects of several land management systems on soll orgamc compos1t1on and dynam1cs Two PhD students arnved last September at CPAC One 1s study1ng the long term 1mpact of several land management systems on the compos1t1on of organ1c matter and the other the short term effects of land use on water and orgamc matter assoc1ated nutnents The effect of land use on so1l organ1c matter compos1t1on started w1th the s01l samphng m several selected s1tes m two so1l types (clay and sandy texturas) m Uberland1a (M Gl The selected land use systems are Contmuous annual croppmg systems contmuous pasture systems nat1ve cerrado plantad forests and crop pasture rotat1on systems Several techmques mcludmg 13C NMR and GC spectroscopy to w1ll be used to 1dent1fy the compos1t1on of SOM m vanous fract1ons of the so1l Further stud1es w1ll be conducted to determme the dynam1cs of most sens1t1ve fract1ons m relat1on to so1l management Water and nutnent dynam1cs stud1es w1ll be carned out w1thm our crop pastura mtegrat1on expenment at CPAC Several treatments mcludmg the nat1ve cerrado and the crop and pastura systems were selected for th1s study The ob]ect1ve of the work 1s to study the effect of mtegrated and not mtegrated systems on water ava•lab1hty and dynam1cs of nutnents denved from organ1c matter mmerallzat1on (N P and SI Selected plots w1ll be sampled over t1me to asses the short term effect of crop pastura rotat1on next year 5 Effects of agncultural development on greenhouse gas fluxes 1n the cerrados J Duxbury andA Cardoso Trop1cal ecosystems are globally 1mportant sources of a number of atmosphenc gases mcludmg carbon monox1de mtnc ox1des and methane Moreover trop1cal land use change mcludmg both the convers1on of trop1cal forest to pasture and agnculture and more mtens1ve management of ex1stmg agncultural land are takmg place m the trop1cs The 1mpact of these land use changas on b1ogeochem1cal cycllng and trace gas 1s not well understood The ob]ect1ve of th1s proJect 1s to determme the mfluence of agnculture development on fluxes of greenhouse from/to so1ls of the cerrado and to learn about the processes controlhng fluxes The focus of the work 1s on the em1ss1on of mtrous ox1de and methane consumpt1on and ox1dat1on Several plots from the Crop pasture expenment at CPAC were selected to conduct the study 1) nat1ve cerrado 21 contmuous grass pastura 3) contmuous grass legume pastura and 4) contmuous annual croppmg system Gas flux measurements started m July 1994 and are expected to contmue for two years Prehm1nary expenments usmg mcubated s01l samples from the selected plots to determme the effect of CH4 concentrat1on and s01l m01sture on methane ox1dat10n showed a clear decrease m CH4 concentrat1on m the O 60 cm so1l depth 1n all the treatments (Figure 1 2) 169 Methane concentrat1ons were less than 1 ppm at the 1 O cm depth suggestmg a strong smk for methane m the s01l surface Flux measurements dunng a two week penod showed that methane ox1dat1on was h1ghest m the corn plots and lowest m the cerrado plots (F1gure 13) The pasture treatments were m between Th1s results suggest that under the cond1t1on of the expenment there 1s not yet a reduct1on of the capac1ty of the s01l to ox1d1ze methane lncreasmg so1l mo1sture decreased methane ox1dat1on capaCity m both nat1ve cerrado and corn plots However the effect was greater m the corn plots md1catmg that changas m pore s1ze d1stnbut1on brought about by cult1vat1on may leave the sollless buffered m 1ts ab1hty to ox1d1ze methane N1trous ox1de fluxes were very low 1n all management systems dunng the dry season However 1t 1s expected large changes dunng the wet season On farm research M1guel A yarza and Lounval Vlfe/a After several months of graz1ng results are showmg the benef1c1al 1mpact of Sty/osanthes mme"ao m 1mprovmg the productiVIty of nce pastura systems m hvestock product1on systems w1th low mputs Ammal gams were cons1stently better m the legume based pasturas than m the Brach1ana pasture recla1med w1th nce and the ongmal sown but degrad1ng grass pastura (Table 6) Legumes planted w1th corn and Pamcum cv Vencedor d1d not estabhsh because of crop and grass compet1t1on In sp1te of the fa1lure to estabhsh the legumes the pure grass pasturas are growmg v1gorously and supportmg h1gh carrymg capac1t1es and adequate ammal gams Th1s must be assoc1ated w1th the residual fert1hty left by several years of annual crops Evaluat10n of legume populat1on m these pasturas 1s showmg a gradual mcrease m the Calopogomum and Neonotoma spec1es over t1me m the Fazenda Sta Terezmha The evaluat1on of changes m so1l fert1hty and so1l phys1cal propert1es under 1mproved and unrecla1med systems started recently Results are not ava1lable yet Conclus1ons Long term controlled expenments offer the best opportumty to measure the 1mpact of 1ntegrated crop pastura systems on product10n and s01l fert1hty The expenment at CPAC allows the opportumt1es for the study of changas m so1l chem1cal phys1cal and b1olog1cal parameters under a w1de range of management vanables After three years of cont1nuous croppmg so1l fert1hty management 1s mfluencmg crop growth and y1elds more than land preparat1on methods So1l compact1on and loss of aggregat1on under the convent1onalland preparat10n systems usmg a d1sk 1s not affectmg crop performance yet Weed pressure 1s mfluenced by fert1hty management and land preparat1on methods Satelhte plots 1n Uberland1a are demonstratmg the benef1c1al 1mpact of forage legumes on ammal product1v1ty of crop pasture systems and the feas1b1hty of mcludmg forage legumes m crop pasture rotat1ons More than 200 farmers of the reg1on have v1s1ted the plots Future plans W1th the 1n1t1at1on of the pro¡ect on so1l organ1c matter we expect to strengthen our act1v1t1es on strateg1c research m our long term controlled expenments Contacts have already been made w1th the Center of Nuclear Energy at P1rac1caba to study the contnbut10n of crops and pasturas to s01l orgamc matter usmg stable 1sotopes A s1m1lar 1mt1at1ve 1s m progress w1th the Nat1onal Center of Agrob1ology at R1o de Jane1ro to study the dynam1cs of n1trogen 1n crop 170 pastura systems usmg ' 5N labelhng Table 6 Ammal performance on 1mproved pasturas estabhshed w1th crops m two contrastmg management systems and s01l types 1n several farms of Uberland.a MG Stockmg rate liV8W81ght gams Pasturas (AU/hal Ammal lg/day) Sta Terezmha (IMS'I sandy soll (94 days of graz1ngl P max1mum cv Vencedor3 Vencedor+ Legumes 2 87 1 73 483 622 Bom Jard1m (IMSI clay soll (99 days of graz1ng) Vencedor+ legumes 2 94 655 Sta ln6z IELS21 el ay soll 1118 days of grazmg) 8decumbens 8 decumbens + legumes Degradad pastura (control) 1 41 2 09 641 825 456 Cachoe1ra IELSI sandy soll (176 days of graz1ngl 8 ruz¡ztensts 8 ruztztensts + legumes Degradad pastura (control) 1 31 1 34 1 28 387 414 272 Are a (kg/hal 164 136 391 162 299 168 192 142 1/ IMS = lntensove management systems woth hogh onputs Pasturas plantad after severa! years of crops 2/ ELS = Extensove hvestock systems woth low onputs Pasturas plantad on degradad pasturas 3/ Cocktail of legumes S gwanensts cv Moneorao + Neonotoma W1ghtu + Ca!opogonwm mucun01des Our current act1V1t1es of momtonng of productoon systems m Uberland1a woll be expanded to studoes on pastura degradatoon The prehm1nary phase of the work woll consost of the selectoon of the several pasturas undergomg degradatoon and the complete charactenzat1on of the1r productovoty and s01l fert1hty Thos w1ll be carned out by one master student from the Unoversoty of Gottmgen The study w1ll be complementad woth the economocal evaluatoon of degradad pasturas and omproved agropastoral systems Th1s work woll be done by a second M Se student from the same Umversoty lt 1s expected that the pasturas selected on thos study can be used m the futura stud1es assessmg the 1mpact of pastura degradat1on on SOM accumulat1on s01l losses and gas fluxes 171 -o o 10 20 ! 30 -1- 40 Q 50 60 70 Methane concentration (ppm) 05 1 -e-Coro ..... Grass pasture +Cerrado 15 F1gure 12 Sml atmosphere concentrat10ns of Clll under severa! management systems of the crop-pasture exper1ment 172 - 35 -= *Coro = 30 -= +Pasture -u b.O '-" 25 -e- Leglpasture = 20 •Cerrado Q -.... = -= 15 -l>'l o lO ~ = = 5 -= .... ~ ~ o 1417/94 17/7/94 2017/94 2317/94 Date F1gure 13 Methane oxJdatJon fluxes mfluenced by the management systems mcluded 10 the crop-pasture exper1ment 173 STRATEGIC RESEARCH ON KEY PROCESSES Recychng of nutnents v1a decompos1t1on of htter and crop res1due Richard Thomas N Asakawa H F Alarcón (TL) lntroduct1on The modern v1ew of the management of s01l fert1hty 1n the trop1cs 1s one wh1ch rehes more on b1olog•cal processes adapts germplasm to adverse so1l cond1t1ons enhances so1l biolog1cal act1v1ty and opt1m1zes nutnent cychng to m1mm1ze externa! mputs while max1m1zlng the eff1c1ency of use of any externa! mputs A part of th1s effort w1ll be the ab1hty to pred1ct the rate of release of nutnents from plant htter and crop res1dues by the use of models basad on a sound understandmg of the key processes nutnent pools and fluxes mvolved In grazed trop1cal pasturas recychng of nutnents v1a plant htter •s probably more 1mportant than recychng v1a the ammal when pasturas are moderately grazed erops generally rece1ve fert1hzer and the1r res1dues therefore are an 1mportant source of nutnents 1n low fert1hty so1ls where pools of nutnents are small and/or unava1lable for plant uptake Therefore a ma¡or effort has been made to charactenze the decompos1t1on processes mvolved w1th both herbaceous and shrubby forages and crop htter Decompos1t1on of herbaceous and shrubby forages A htter bag techn1que (Thomas & Asakawa 1993) was used to compare the •mt1al compos1t1on and rates of decompos1t10n of htter from herbaceous and shrubby legumes and one grass spec1es w1th the add1t10nal treatment of htter s1ze The latter cons1sted of cutt1ng leaf htter of all spec1es down to a s1ze Similar to that of Stylosanthes cap1tata about 1 O mm length L1tter bags were placed 3 cm below the so11 surface m large plast1c contamers 150 x 30 x 25 cm) m a glasshouse contammg an ox1sol from eanmagua lclay loam from lntroduct1ons 11) Table 7 shows the spec1es used and the1r Jmt1al compos•t•ons As there was httle or no d1fference between cut or mtact htter the data for the two treatments were pooled (Table 8) Rates of decompos1t1on and half hves 1nd1cated that there was a w1de vanat1on amongst the spec1es w1th A pmto1 S cap1tata e flonbunda and O velutmum decomposmg the fastest and F macrophylla O oval!foi!Um and e acutlfol!um the slowest (Table 8) Spec1es w1th fastest rates of decompos1t10n generally had low hgmn N e N rat1os or low amounts of tanmns conversely spec1es w1th slowest rates had h1gh hgmn N e N or h1gh amounts of tanmns when the compos1t1onal factors were cons1dered together L1gmn content (r2 = O 87) hgn1n N (r2 = O 67) and hgmn hem1cellulose rat1os Ir' = O 68) gave the best lmear correlat1ons w1th k (all p E ~.e ~ 200 150 .... ~ca, 100 -Q) > a; -E so ::J (.) Brach1ana d1ctyoneura 1- 23 F1g 14 Response of herbage to unne apphcat1on ttl S: -z C» .lo: ttl S: -z C» .lo: ttl S: - 110 o Scm 40 :so 20 10 110 5 10cm 40 30 20 10 IIOir-----------------------~ 10 2Dcm 40 z :so C» .lo: 20 10 o 2 7 14 21 28 35 42 134 Days after unne aphcatton § 3 § 3 F1g 15 lnorgamc N 1n so1l after unne aphcat1on 181 Levels of morgamc N (ammomum and mtrate) were momtored at O 5 5 1 O and 1 O 20 cm depths over t1me Ammon1um mcreased m the so1l 1mmed1ately after apphcat1on (Figure 1 5) and then decreased gradually over 28 days There was httle or no appearance of n1trate o ver the 1 55 days momtored md1catmg httle apparent mtnf1cat1on Most of the mcrease m s01l ammomum N levels occurred 1n the top 5 cm of the so11 prof1le w1th httle apparent movement downward (Figure 14) These results awa1t conf1rmat1on from the ' 5N labelhng data 1n s01l organ1c matter Conclus1ons The results demonstrate the faster recychng of N v1a ammal unne than v1a plant htter and a greater % recovery w1th unne comparad w1th htter Apparent losses of unne N were relat1vely low comparad w1th data from temperate systems The data md1cate that although the returns of N to the so1l may be greater v1a htter than unne the actual amounts of N ava1lable for plant uptake may be greater from ammal excreta Phosphorus transformat1ons m 1mproved pasturas A Oberson (U Saskatchewan) D K Frtesen (IFDCJ H T1essen J Mo1r (Umv of Saskatchewan) G Barrero (TL) lntroduct1on Phosphorus 1s among the nutnents most hm1tmg plant product10n 1n the ac1d savannas Ac1d tolerant pastura germplasm cannot be estabhshed Wlthout add1t1ons of purchased P and there are msuff1c1ent amounts ava1lable 1n the unamended so1ls to 1mprove recychng lmproved pastures usmg ac1d tolerant grass and legume vanet1es have resultad m a 1 O to 1 5 fold mcrease 1n beef product1on per ha m f1eld expenments 1n Canmagua The fact that th1s dramat1c product1on 1ncrease was obtamed w1th modest P fert1hzer mputs ra1ses quest1ons about the perce1ved meff1c1ency of P fert1hzer 1nputs on h1gh weathered h1gh P sorbmg so1ls and about transformat1ons of s01l and fert1hzer P occurnng when nat1ve savanna 1s replaced by 1mproved grass only or grass legume pasturas The present mvest1gat1on attempts to answer these quest1ons by means of sequent1al chem1cal P fract1onat10n to assess the s1gmf1cance of d1fferent orgamc (P ) and morgamc (P1) P fract1ons and by est1mated P budgets for nat1ve savanna and 1mproved pastura so1ls The fract10nat1on method was augmented w1th the charactenzat10n of the C/N/P rat1os of the orgamc matter extractad dunng d1fferent steps Data are d1scussed m terms of an 1mproved understandmg of P transformat1ons m so1ls under 1mproved pasturas The usefulness of d1fferent P and P fract1ons as susta1nab1hty md1cators IS evaluated Matenal and Methods So1l samples were taken 1n a long term pastura expenment estabhshed m 1978 on an Ox1sol at the ICA CIAT Canmagua research stat1on m the eastern plams of Colombia Treatments mcluded were nat1ve savanna (SA) grass only pastura (GO) (BrachJarJa decumbens cv Bas1hsk) and grass legume pastura (GL) (BrachJarta decumbens cv Bas1hsk w1th Puerarta phaseo/o¡des CIAT 9900 (Kudzu)) D1fferent P and P fract1ons were sequent1ally determmed usmg the followmg extractants H20 w1th amon exchange resm O 5 M NaHC03 O 1 M NaOH 1 O M HCI hot conc HCI The so1l remammg at the end was d1gested w1th perchlonc aCid to assess res1dual P lnorgamc P m extracts was determmed by the molybdate ascorb1c ac1d method Total P m the H20 NaHC03 NaOH and hot HCI extracts was measured after d1gest1on w1th K2S20 8 Orgamc P was calculated as the d1fference between total P and P Fract1ons are 182 abbrev1ated as follows H20 P Resm P B1c P P NaOH P P HCI P HCI. P P Res1d P, Orgamc C m H20 NaHC03 NaOH and HCI" extracts was determmed usmg a mod1f1ed K2Cr20 7 H2S04 method Total N m H20 NaHC03 and NaOH extracts was determmed by d1gest1on w1th K2S20 8 m an autoclave Total N 1n the HCI" extracts was determmed after KJeldahl d1gest10n on an autoanalyser RESUL TS ANO DISCUSSION Effects of 1mproved pasturas on orgamc and morgamc soll P fract1ons Wh1le both P and P fract1ons were affected by 1mproved pasture treatments (Table 11) the GL pasture mduced more pronounced changes than the GO pastura when comparad to the SA Th1s was a consequence of the overall s1gn1f1cantly h1gher total P (SumPT and PT Table 11) content of the GL pastura The 1mportance of P m the so1l solutlon 1s reflectad m the almost equal amounts of P and P extractad m the res1n H20 step In extracts of temperate so1ls B1c P IS cons1dered to contam mamly lab1le act1vely cychng P compounds wh1le NaOH P was mterpreted as moderately lab1le P compounds wh1ch reflect the slower long term transformat10ns of s01l P occurnng dunng s01l format1on or prolongad cult1vat1on In the present study lablle B1c P was not affected by the convers1on of nat1ve savanna mto 1mproved pasturas wh1le NaOH P was mcreased under the 14 year old GL pasture but not the GO pastura Consequently these results support a rev1sed mterpretat1on of these two fract1ons for trop1cal so1ls B1c P appears to be at a constant level regardless of the croppmg h1story m sorne trop1cal s01ls wh1le the NaOH P fract1on reflects the overall changes m s01l orgamc matter and P levels when the s01l 1s stressed by cult1vat1on and net P export Th1s fract1on m ay therefore represent a relat1vely act1ve reserv01r (source or smk) of P under trop1cal cond1t1ons The observad mcrease m extractable so1l P m GL was probably not due to mert plant maten al s1nce the legume Puerana phaseo/o1des 1s known to decompose at s1mllar rates as the grass Brach1ana decumbens G1ven the mcrease m lab1le P fract1ons (Resm P, B1c P Table 11) espec1ally m the GL pastura P m plant res1dues and ammal excreta seems to be recycled eff1c1ently Th1s IS supported by the observat1on of h1gher ac1d phosphatase act1v1ty m 1mproved pasturas than m nat1ve savanna ( 181 ± 8 213 ± 14 239 ± 17 mg p mtrophenol kg 1 h 1 for SA GO and GL respect1vely) Phosphatase act1v1ty was at a h1gh level for all treatments thus showmg the s1gn1f1cance of orgamc P m the so1l P dynam1cs of 1mproved pasturas and nat1ve savanna All extractable morgamc fract10ns were mcreased m the GL pasture but not the GO pasture (Table 11) The mod1f1cat1ons observad m the GL pasture m lab1le fract1ons 1nd1cate an 1mprovement m P ava1lab1hty wh1ch 1s also reflectad m the results of the Bray 11 P soll test values (Table 11) Although the1r content 1s st1ll very low the mcrease m lab1le P fract1ons 1s surpnsmg g1ven the low P mputs made by fert1hzers (T able 1 2) and the med1um to strong P sorpt10n capac1ty of the s01l In all treatments d1fferences m P fract1ons were largely m absoluta values rather than relat1ve terms Th1s suggests that the prehmmary part1t1onmg of apphed P among so1l P fract1ons was not greatly mfluenced by 1mproved pasture spec1es e1ther m GO or m GL systems S1gn1f1cantly the observad absoluta mcreases m fract1on s1zes mcludmg fract1ons cons1dered as read1ly ava1lable occurred almost ent1rely m the GL pasture 183 -CD ~ Table 11 lnorgamc P fract10ns m the O 1 O cm so1l !ayer 1n nat1ve savanna and 1mproved pasturas P fracbon Res111-P H,C>-P BIC P BIC p NaOH-P NaOH-P HCI"' P HCI P HCI"' P Res1d P, SumPTi Extr P.,§ ~ Brayll IT'g p kg 1 MeanSA 35b 27 b 33b 96 231 b 431 b 13 541 b 83 509ab 200b 638 b 185 b 13b SE 02 01 01 07 07 18 01 16 1 9 14 2 26 3 01 Sum% 17 14 17 48 11 6 216 06 271 42 254 100 319 923 MeanGO 40b 43a 34b 99 228 b 414 b 1 8 606b 127 499b 211b 684ab 193 b 14b SE 01 05 01 06 08 21 03 31 26 1 1 8 36 6 02 Sum 0/ 19 21 1 6 47 10 8 196 09 288 60 236 100 323 915 MeanGL 52 a 35ab 46a 96 311 a 519a 15 686a 12 1 549a 240a 771 a 227 a 22a SE 04 03 02 05 12 31 01 1 5 25 09 6 38 5 03 Sum% 22 1 5 19 40 130 217 06 286 50 229 100 312 939 IIF test# - - NS - - NS - NS - - - tMean and SE for 8 (GO and SAl and 7 (GL) samples per treatment Means followed by the same letter are not S1gmf1cantly d1fferent (P =O 05) by Tukey s mult1ple range test *Su m of all fract1ons § Sum of extracted orgamc P fract1ons ,Total P determmed by perchlonc ac1d d1gest Table 12 Est1mated P budget over 15 years (1978 1993) m nat1ve savanna and 1mproved pasturas and companson w1th measured total P (PT) mcreases dueto fert1hzat10n 1 reatment tst1matea 1' ouaget Measurea 1'1 t:la1ance mput 1 t:xport J 1 t:lalance d1fference t unaccounted for kg 1' na SA 2 1 2 1 GO 103 24 +79 10 4 70 7 GL 106 36 +70 54 6 17 5 t D1ffarence 1n PT content (from tabla 3) batween 1mproved pasturas and nat1ve savanna ; P export of 8 g P kg 1 of hva we1ght gam (LWG) LWGs of 15 20 200 and 300 kg ha 1 yr 1 10 SA GO and GL respect1vely P budgets 10 nat1va savanna and 1mprovad pasturas P budgets were est1mated to determme the relat1on between P mput m 1mproved pasturas and P output v1a beef cattle and to better understand the mod1f1cat1ons m the s1ze of P fract1ons or total P that occurred espec1ally m the GL pastura P budgets were calculated by subtractmg the P removed from the system by ammals from the P apphed m mmeral fert1hzer (Table 12) Pos1t1ve P budgets were obtamed for both 1mproved pasturas whlle a small negat1ve budget was found for the unfert1hzed SA (Table 12) Clearly d1fferences 1n PT as well as ava1lable P fract1ons between the GO and the GL pastura were not due to d1fferences 1n P budgets The calculated d1fferences m the P budget are 81 kg P ha 1 between GO and SA and 72 kg P ha 1 between GL and SA (column 4 table 12) Assummg that all (non exportad) apphed P remamed 1n the O 1 O cm surface so1l layer and usmg a bulk dens1ty of 1 3 g cm 3 the measured PT d1fferences (column 5 table 12) are smaller than the values est1mated from the budget for the GO as well as for the GL treatment Espec1ally for GO the d1screpancy (column 6 table 12) clearly exceeds d1fferences that m1ght be explamed by analyt1cal errors One poss1ble explanat1on for the observad d1screpanc1es 1s that the P budgets were much less than est1mated for both 1mproved pastura types However overest1mat1on of the P budget could only occur by underest1mat1on of P exportad m ammal hve we1ght gam (LWGl and 1t 1s unhkely that the error would be so large Moreover P exportad would be expected to be s1m1lar for both systems or perhaps even greater from GL pasturas g1ven the h1gher P content m legumes than grasses The alternat1ve explanat1on 1s that P moved out of the surface so1l 1nto the so1l prof1le and that there was a largar net movement of P out of the top so1l !ayer m the GO pastura comparad to the GL pastura lt 1s very Improbable that any s1gmf1cant movement of 1norgamc P would have occurred through leachmg g1ven the substant1al sorpt1on capac1ty of th1s s01l However losses of soluble orgamc P as well as of part1culate P may occur cons1denng the h1gh ramfall mtens1ty the h1gh P contents 1n the water of the resm extract1on step (Table 11) and the fme textura of the so1l Although pos1t1ve P budgets and mcreases 1n most P fract1ons were observad for both 1mproved pasturas the companson of PT and the est1mated P budget md1cates P losses from the O 1 O cm layer espec1ally m the GO pastura Consequently the mtegrated cons1derat1on 185 of P fract1ons PT and P budget shows that the GO pastura 1s less resource conservmg than GL C/N and C/P rat1os of d1ffarent orgamc P fract1ons D1fferences m the compos1t1on of so1l orgamc matter extractad m the d1fferent steps or remammg m the so1l res1due are lllustrated by the1r e/N and e/P rat1os (Tabla 13) S1nce pasture type d1d not s1gmf1cantly affect these rat1os only mean values for each extract are presentad In general the e/N/P rat1os (Tabla 13) espec1ally e¡p were.h1gh when comparad to values known from the hterature wh1ch are mamly based on stud1es on temperate so1ls e/N rat1os vaned more between extractants than e¡p Tabla 13 e/N e/P and N/P rat1os m d1fferent extracts of the sequent1al fract1onat1on procedure as well as etN/P for N = 10 Geometnc means over all f1eld treatments H20 B1c NaOH He l. Extrt Res1d Total* e/N 16 o 65 16 90 11 5 28 14 e¡p 229 197 192 354 212 176 345 N/P 14 33 12 4 39 18 4 63 20 e/N/ 160/10/ 65/10/ 160/10/ 90/10/ 115/10/ 280/10/ 137/10/ p o 71 o 33 o 80 o 26 o 54 1 6 o 51 t RatiO between sum of extractad e N and P raspect1vely t Rat1o between total e total N and the sum of extracted orgamc P The e¡p rat1o m so1l and plant res1dues added to so1l has been used to pred1ct net 1mmob1hzat1on and mmerahzat10n m so1l w1th P mmerahzat10n occurnng at rat1os of 200 or less and P 1mmob1hzat1on when rat1os are 300 or more eonsequently fert1hzer P added to any treatment of the present study would be 1mmob1hzed m orgamc matter However the mcrease m avallable morgamc P fract10ns (Table 111 espec1ally m the GL pastura occurnng at the sama t1me as an mcrease m total e and N (data not shown} suggests an mcons1stency m the pnnc1ple basad on e¡p mdex Plant htter of 1mproved grass and legume spec1es has a C/P rat1o rangmg from ca 400 to 2700 eonsequently w1der e¡p rat1os m these so1ls are hkely an md1cator of recent orgamc matenal and therefore an md1cator of P w1th a h1gh ava1lab1hty when P 1s mmerahzed concom1tantly to e and N dnven by the need for energy The h1gh phosphatase act1v1ty m all treatments mcluded md1cates the 1mportance of b1ochem1cal or b1olog1cal processes m P mmerahzat1on Based on the h1gh e¡p rat1os observed m the present study not only m stable fract1ons but also m lablle P fract1ons a re evaluat1on of the cnt1cal e/P concept 1s reqUired espec1ally when used as a parameter for modelhng P turnover eg m the eentury model (see contnbut10n 186 of G1¡sman et al m th1s annual report) Factors enhanc1ng P avallab1hty m GL pasturas For many of the vanables mcluded to mvest1gate P transformat1ons GL shows a more d1stmct separat1on from SA than GO does lncreases m P, fract1ons cons1dered as read1ly plant ava1lable show that P added to the soll v1a fert1hzer plant htter and ammal excreta cycles more eff1c1ently m the GL pasture Explanat1ons for the fact that only small P mputs v1a fert1hzers enhanced P cychng m the GL more than m the GO pasture may be found m the d1fferences m so1l floral and faunal act1v1ty as well as by the mtervent1on of plants m P cychng A recent diploma thes1s wh1ch stud1ed these pasturas showed that macrofaunal b1omass m the GL treatment was tw1ce that of the GO treatment wh1ch was double that of SA w1th earthworms bemg the dommant macrofaunal spec1es P ava1lab1hty IS apparently enhanced m casts of earthworms The greater 1mportance of processes hnked to b10log1cal act1v1ty are also md1cated by the shghtly h1gher phosphatase act1v1ty m GL than GO pasturas Furthermore m1crob1al P wh1ch 1s cons1dered to be the second most act1ve soll P pool after solut1on P was larger m the GL pasture comparad w1th GO wh1ch m turn was shghtly h1gher than SA (data not shownl Root growth and charactenst1cs may also affect P cychng eff1c1ency In th1s respect legumes may be more eff1c1ent cyclers or at least promote more eff1c1ent cychng m m1xed pasturas than grasses Ev1dence of th1s may be seen m the comparat1vely h1gher cat1on concentrat1ons (espec1ally Ca) m the surface so1l of the GL pastura (data not shown) In summary therefore 1n1t1allow P mputs through fert1hzers (Table 12) and N mput by legumes have enhanced nutnent cychng Excreta from grazmg hvestock make a more 1mportant contnbut1on to the nutnent ennchment m the top most so1l layar of 1mproved pasturas than m the SA as a result of mcreased plant product1on stockmg rate and h1gher nutnent contents espec1ally m legumes Together w1th the plant htter excreta lead to 1mproved nutnent cond1t10ns for soll flora and fauna As a result of mteract1ons between chem1cal and b1olog1cal so1l propert1es P ava1lab1hty 1s mcreased Nutnent cychng through the m1crob1al b1omass A G1¡sman A Oberson (U Saskatchewan) D Fnesen (IFDC) J Sanz R Thomas (TL) lntroduct1on So1l m1crob1al b1omass has been shown to be a sens1t1ve 1nd1cator of changas m s01l organ1c matter (SOM) quahty and quant1ty lt responds relat1vely rap1dly 1 e comparad to total SOM to changas m orgamc mputs to tha so1l or m so1l managament and d1fferances m m1crob1al b1omass can be detectad before they can be measured m total SOM Bes1des bemg useful as an md1cator of the orgamc mattar status of a so1l the 1mportance of m1crob1al b1omass 1s found m 1ts central role m nutnent cychng m a s01l Strong correlat10ns have been found batween tha amount of nutnents held m tha m1crob1al b1omass and levels of mmarahzable nutnents m the so1l 1nd1cat1ng that nutnent cychng m a so1l 1s t1ghtly hnked to the turnover of m1crob1al b1omass The ob¡act1ve of the present study was to determme wh1ch changes m m1crob1al b1omass occur when nat1ve savanna whose orgamc matter content m ay be expected to be at aqUihbnum level 1s opened up and brought under cult1vat1on M1crob1al C N and P contents and C and N mmerahzat1on ratas were determ1ned m so1l from 5 year monocrop nca and vanous forms of nce pastura rotat1on 187 Matenala and Methods So1l samples were taken from the O 1 O cm so1l layer of a f1ve year old nce pasture rotat1onal expenment (Table 14) on a clay loam Ox1sol at Matazul farm m the eastern Llanos of Colombia M1crob1al b1omass C N and P content (fum1gat10n extract1on method) and of C and N mmerahzat1on rate ( 1 O da y mcubat1on) were measured Table 14 Crop rotat1on schedule of the expenments Code unt1l 1988 1989 1990 1992 1993 R GL nat1ve savanna nce grass legume grass legume RG nat1ve savanna nce grass grass R GL R nat1ve savanna nce grass legume nce RGR nat1ve savanna nce grass nce CR nat1ve savanna nce nce nce SAV nat1ve savanna nat1ve savanna nat1ve savanna nat1ve savanna Resulta and DISCUSSIOn M1crob1al b1omass nutnent contenta M1crob1al b1omass C d1d not vary much among treatments all treatment averages bemg w1thm a 1 3% dev1at1on from the overall average (Table 15) No d1fferences were found between pasture or cropped plots M1crob1al b10mass N was cons1derably lower for the CR treatment than for the other treatments Consequently the m1crob1al b1omass C/N rat10 was more than one and a half t1mes as w1de under contmuous nce as under any other rotat1on treatment and the contnbut1on of m1crob1al N to total orgamc N content (Nm /N 9) was also low Th1s probably reflects the 1nput of low N nce res1dues at harvest Left over mtrogen from fert1hzat1on obv1ously d1d not compensate for the matenal s h1gh C/N rat1o lt may also be that fung1 havmg a w1der C/N rat1o than bactena made up a larger fract1on of the m1crob1al populat1on poss1bly because the vertical d1stnbut1on of freshly added res1dues m the CR treatment was d1fferent as most crop res1dues at harvest were depos1ted on top of the soil Fung1 are relat1vely more 1mportant for the decompos1t1on of surface htter The data on m1crob1al b1omass P showed a d1st¡nct1on between the contmuous nce and nat1ve savanna on the one hand and the nce pastura and nce pastura nce plots on the other the former group showmg lower b10mass P values W1th1n each of these two groups d1fferences were non s1gmf1cant B1omass C/P rat1o s showed the same two groups however w1th the except1on of the R G treatment m wh1ch the C/P rat1o was relat1vely h1gh The mlcrob1al C/P rat1os are at the very h1gh end of the range commonly found suggestmg that the m1crob1al populat10n may have adaptad 1tself to the low P cond1t1ons of th1s so1l Desp1te these w1de C/P rat1os the contnbut1on of m¡crobial b1omass P to total SOM P agrees well w1th values commonly found m temperate s01ls w1th a much h1gher P fert1hty The s¡gn¡f¡cance of th1s cons1sts m 1t that P absorbed m mlcrob1al cells becomes part of a pool wh1ch turns over dynam¡cally whereas w1thout th1s m1crob1al uptake 1t m1ght have become unava1lable to the plant due to the strong P sorpt1on capac1ty of these so1ls The m1crobes thus compete successfully for morgamc P and may contnbute s1gmf1cantly to P cychng 188 C and N m•nerallzat1on and microb1al act1vlty The carbon mmerallzat1on rate dunng the 11 day 1ncubat1on penod was h1ghest m the R GL pasture followed by the R G pastura (Tabla 1 6) The cropped plots espec1ally the contmuous nce and the nat1ve savanna showed a lower e mmerallzat1on rate than the pasture plots The same p1cture was found for the fract1on of orgamc e that was mmerahzed per day 1em /e 01 In most samples net mtrogen 1mmob1hzat1on occurred as amounts of extractable N after mcubat1on were lower than befare mcubat1on Only m the savanna treatment net N mmerahzat1on was pos1t1ve on average (Table 16) but th1s value was strongly b1ased by one pos1t1ve outher Re1ectmg th1s one value would bnng the savanna average down to O 27 pg g 1 d 1 wh1ch 1s w1thm the ranga of the other treatments values Obv1ously the mcubat1on penad had not been long enough to result m a net N mmerahzat1on earbon n1trogen rat1os of the decomposed orgamc matter consequently could not be calculated Table 15 Mlcrob&al C N and P contents and the1r contnbut1on to total orgamc C N and P contents 1n so1l from vanous nce pastura rotat1ons and nat1ve savanna, Treatment e N p CN CP C~IC,.. Nnm:/Nor P...JP~ lpg g 1 1%1 A GL 486 8 66 3 13 o 76 36 5 26 52 16 o AG 520 1 63 2 11 5 83 45 5 24 47 11 9 A GL A 441 1 52 9 13 o 85 34 1 22 37 12 1 AGA 543 7 66 4 13 6 82 39 2 24 46 14 1 CA 434 1 34 3 96 13 o 48 o 22 25 68 SAV 473 4 53 1 96 93 49 6 26 44 11 7 The stat•st•cal s1gmf1cance of the results 1s not md1cated smce th1s part of the analys1s has not been done yet For treatment codes see Table 14 C~/C Nm /N P m /P rg = contnbut1on of m1crob1a1 C N and P to orgamc C N and P Table 16 M1nera1tzateon of C and N dunng an 11 day mcubat1on of s01l from vanous nce pastura rotat•ons and nat1ve savanna• Treatment A GL AG A GL A AGR CA SAV e 6 74 5 81 4 65 480 3 36 4 51 N (pg g d 1 o 65 o 55 o 25 053 o 54 002 c_¡c,. qCO 1 (%d 1 lpg co2 e mg blomass e hr 0033 o 58 0026 047 0023 044 0024 o 37 o 017 o 32 0023 o 40 1 ne stat1St1ca1 Slgniflcance OT tne resu1ts 1s not 1na1catea s1nce tms pan oT tne ana1ys1s nas not oeen aone yet For treatment codes see Tabla 14 Ctnl ICor = fract1on of orgamc carbon mmerahzed per day qCO = metabohc quot1ent 189 The e mmerahzat1on rate per umt b1omass e lmetabohc quot1ent qe02 ) was h1ghest m the R GL treatment and lowest m the eR treatment The two treatments mvolvmg a legume m the rotat1on IR GL and R GL RJ had a h1gher qC02 than the comparable treatments w1thout a legume IR G and R G R) A h1gher qe02 md1cates that e1ther a greater fract1on of the total m1crob1al populat1on was act1ve (due to a greater substrate avallablhty) or the energy use eff1c1ency of these orgamsms was lower The latter 1s not hkely here smce the treatment whose plant res1dues were most N nch and thus could be decomposed w1th a relat1vely h1gh metabohc eff1c1ency IR GU yet had the h1ghest qe02 lt 1s therefore assumed that a greater fract1on of the m1crobes was act1ve under pastura than under croppmg probably because of d1fferences m orgamc matter Input Whereas m croppmg systems orgamc matter 1nput to the so1l largely occurs as an annually or sem1 annually repeated pulse m pasturas there IS a cont1nuous mput from dead roots grazmg losses faeces and unne Moreover the amount of plant res1dues returned to the s01l IS usually lower under croppmg smce part of the plant product1on 1s harvested and thus removed from the f1eld Th1s makes 1t probable that the m1crob1al populat1on under pastura stays m a more act1ve form Nutr~ent cychng through m1croblal b1omass turnover Assummg a m1crob1al turnover t1me m the range of O 5 to 1 O year the amount of nutnents cychng through the m1crob1al b1omass was calculated Table 17 shows that the m1crob1al b1omass can contnbute cons1derably to plant nutnent supply part1cularly when takmg 1nto account that the calculat1ons only refer to the O 1 O cm so1llayer Moreover the contnbut1on of m1crob1al metabohtes to nutnent cychng should also be added to these numbers lt 1s also clear however that th1s nutnent cychng pathway can supply much less nutnents (espec1ally Nl 1n the nce monocrop than 1n the other treatments Th1s 1s relatad to the fact that m the monocrop a smaller fract1on of the orgamcally bound nutnents 1s m the m1crob1al b1omass (Table 15) suggestmg there 1s less young orgamc matter The reduced res1due mput to the so1l due to removal of gram and straw from the f1eld and the h1gher orgamc matter decompos1t1on rate due to frequent so1l d1sturbance by t1llage are at the root of th1s Tabla 17 Nutnent flux through m1crobtal btomass (kg ha yr ) at two turnover ttmes 1n sotl from vanous nce pastura rotat ons and nattve savanna' t=O 5 t=1 o Treatment N p N p R GL 165 8 32 6 82 9 16 3 RG 157 9 28 6 79 o 14 3 R GL R 132 1 32 4 66 1 16 2 RGR 166 o 33 9 83 o 17 o CR 85 6 23 9 42 8 11 9 SAV 132 7 24 1 66 4 12 1 The stattsttcal s gmf1cance of the results IS not tndlcated smce thts part of the analys1s has not been done yet For treatment Godes see Tabla 14 t = mtcrobtal turnover t1me (yr) 190 In strongly P sorbmg solls such as the Ox1sol of the present study very httle morgamc P (P l m ay be read1ly exchangeable and the ava1lab1hty of P may depend more on the turnover of eas1ly decomposable so1l orgamc matter than on the release of adsorbed morgamc P The 1mmob1hzat1on of P by m1crobes and 1ts gradual release by m1crob1al turnover prov1de a mechamsm by wh1ch P 1s protected from phys1c chem1cal adsorpt1on react1ons In the present expenment more P was held 1n m1crob1al b1omass than was res1n exchangeable (except for the CR treatment wh1ch was annually fert1hzed w1th P Oberson et al unpubhshed data) The s1ze of the P flux through the m1crob1al b1omass (between 12 and 34 kg ha 1 yr 'lmd1cates that th1s could be a major pathway of P cychng m these low P so1ls Evaluat1on of sorne Phys1cal Propert1es of an Ox1sol after Convers1on of Nat1ve Savanna mto Legume based or Pure Grass Pastures A J G1Jsman (U Saskatchewan) and R J Thomas (TLJ lntroduct1on L1ttle mformat1on 1s ava1lable on short or long term development of the phys1cal cond1t1ons of so1l under pasturas estabhshed m prev1ously nat1ve savanna lnstead most work m th1s f1eld has been done w1th pasturas estabhshed after cleanng a tropical forest where a considerable changa m so1l phys1cal charactenst1cs may be expected g1ven the drast1c nature of the land convers1on and the methods used for the cleanng Earher 1t was shown that there was no d1fference m so1l aggregate d1stnbut10n between grass only and grass legume pastures estabhshed m prev10usly nat1ve savanna but that the aggregates were somewhat more stable m the presence of a legume Here data are reportad on a number of other so1l phys1cal charactenst1cs of these pasturas Matenals and Methods S01l samples were taken from two adJacently s1tuated expenments on grass only and grass legume pasturas on a clay loam Ox1solm Canmagua (Table 181 A nat1ve savanna control was also mcluded Plots were 1 ha each but m expenment 2 they were d1v1ded 1nto two halves for alternate grazmg only one of the two halves was used for samphng Expenments 2 and 3 were rephcated but for expenment 1 the rephcate plot was no longer ava1lable However because the plots m th1s expenment were not d1v1ded mto two as 1n the other pasture expenment a pseudo rephcate was created here by d1v1dmg the plots mto two halves for the samphng The followmg so1l phys1cal parameters were determmed Water retent1on charactenst1cs Unsaturated hydrauhc conduct1v1ty as calculated from the water retent1on charactenst1cs usmg a curve f1ttmg computer program for water retent1on curves Apparent bulk dens1ty Total poros1ty and pore s1ze d1stnbut1on as calculated from the water retent1on curve Penetrometer res1stance Water mf1ltrat1on rate w1th a double nng mf1ltrat1on system 191 Table 18 Treatment schedule Treatment B decumbens 8 decumbens + Puerarta phaseolo1des 8 humldlcola + 2 an /haA 8 humldlcola + Arach1s pmto1 + 2 an /ha 8 humldlcola + 4 an /ha 8 humldlcola + Arach1s pmto1 + 4 an /ha Savanna an /ha = antmals per hectare Results Apparent Soll Bulk Dens1ty and Penetrometer Res1stance Cede Bd Bd/Pp Bh+2 Bh/Ap+2 Bh+4 Bh/Ap+4 Sav The apparent so11 bulk dens1ty vaned from 1 23 to 1 30 g cm 3 (Table 19) w1th no d1fferences between treatments The penetrometer res•stance mcreased steeply w1th depth m all treatments down to a depth of about 40 to 50 cm (F1g 16) The data show a h1gher res1stance 1n the topsoll of the savanna comparad w1th the pasturas the oppos1te bemg found m the subso1l (stat1st1cal analys1s of th1s d1fference was not poss1ble) lnclus1on of a legume 1n the pasture d1d not affect the penetrometer res1stance m the 8 decumbens + 1 P phaseolo1des expenment but 1n the 8 hum1dlco/a + 1 A pmto1 expenment the penetrometer res1stance was slightly h1gher •n the grass legume treatments than 1n the grass only treatments (P 30pm Layar O 5 cm Bd 1 30 476 232 103 059 082 Bd/Pp 1 29 466 234 109 059 063 Bh+2 1 27 487 220 119 068 080 Bh/Ap+2 1 27 485 221 112 064 088 Bh+4 1 28 485 192 125 075 093 Bh/Ap+4 1 28 473 186 122 074 090 Sav 1 30 483 177 120 074 112 Layar 6 11 cm Bd 1 30 456 186 126 075 070 Bd/Pp 1 30 458 208 113 065 b72 Bh+2 1 24 469 194 115 069 092 Bh/Ap+2 1 24 475 199 107 064 104 Bh+4 1 24 479 164 119 076 121 Bh/Ap+4 1 23 474 161 118 077 118 Sav 1 29 465 174 135 081 075 ASee Table 18 for explanat1on of treatment codes 195 The pore s1ze d1stnbut1on (Tabla 191 reflects the d1fferences observad m the pF curves settmg apart the low from the h1gh stockmg rate treatments the latter havmg a relat1vely small volume of small pore space and large volume of large pore space These d1fferences were s1gmf1cant for the pores 30 pm m the soil layer O 5 cm of the 8 hum!d!cola + 1 A pmt01 expenment No legume effect was found Calculated Unsaturated Hydraulic Conduct1V1ty The calculated unsaturated hydrauhc conduct1v1ty (K) of the pasture so1ls (F1g 181 was generally lower than that of the nat1ve savanna soil the d1fference bemg most pronounced for the pasture treatments w1th a relat1vely low stockmg rate (¡ e the 8 decumbens + 1 P phaseolo1des expenment and the low stockmg rate treatments of the 8 hum!d!cola + 1 A pmto1 expenmentl An except1on was the 6 11 cm so1llayer of the 8 decumbens + 1 P phaseolo1des treatment wh1ch was largely due to one outher sample w1th a relat1vely low K at low values of O No cons1stent legume or stockmg rate effects were found although for the 6 11 cm layar of the 8 hum!d!cola + 1 A pmto1 expenment K was S1Qn1f1cantly lower for the + legume treatments than for the legume treatments at h values less negat1ve than 1 50 kPa Water lnfiltrat1on The measured cumulat1ve water mfiltrat10n and the f1tted curves are shown 1n F1g 19 The f1t between curve and data was t1ght w1th Ffs of md1v1dual mf1ltrat10n measurements varymg from O 988 to 1 000 A d1fference 1n water ¡nfiltratlon rate was observad between the grass only and the grass legume pasturas the latter showmg a more rap1d water mf1ltrat1on (SIQnlflcant at P< O 005 for the 8 hum!d!cola + 1 A pmto1 expenmentl Th1s could be attnbuted toa h1gher termmal water mf1ltrat10n rate (Al and except for the 8 humldlcola + 1 A pmto1 treatment at low stockmg rate also a h1gher sorpt1v1ty (SI 1n the grass legume pasturas A shght (non s1gmf1cantl declme 1n water mfiltrat10n rate w1th mcreasmg stockmg rate was observad m the 8 hum¡d¡cola + 1 A pmto1 expenment Th1s was due to a declme m termmal mf1ltrat1on rate the results on sorpt1v1ty bemg vanable Water mf1ltrat10n m nat1ve savanna so1l was slower than m the grass legume pasturas lt was also slower than the grass only pastura of the 8 decumbens + 1 P phaseo/o1des expenment but faster than or equal to the 1nf1ltrat•on 1n the grass only pasturas of the 8 hum!dlcola + 1 A pmto1 expenment D1scuss1on The mclus1on of a legume m the pasturas d1d not have much effect on the s01l s bulk dens1ty total poros1ty pore s1ze d1stnbut1on water retent10n charactenst1cs or calculated unsaturated hydrauhc conduct1v1ty (the latter two bemg a result of the poros1ty parametersl Although the legume effect on penetrometer res1stance was s1gmf1cant 1t was not b1g enough to be of much pract1cal 1mportance In stnk1ng contrast 1s the mcrease m the water mf1ltrat1on rate m grass legume comparad w1th grass only pasturas The two grass only treatments 1n the 8 humld!cola + 1 A pmto1 expenment al so showed a reduced water 1nf1ltrat1on rate comparad w1th the nat1ve savanna The apparent lack of correspondence between the data on water mf1ltrat10n rate and poros1ty can be explamed from Po1seu1lle s Law wh1ch relates the rate of flow through a porous med1um to the 4" power of the cap1llary rad1us Thus small changes m pore d1ameter can have a profound 1mpact on the hydrauhc conductiVItY of a s01l wh1le 196 hardly affectrng total poros1ty Moreover small pares can conduct large volumes of water as long as they are contrnuous but they do not contnbute much to total poros1ty D1fferences of scale m the methods used may conceal sh1fts m poros1ty srnce larger pares or so1l cracks have a low chance of berng sampled m the small samphng cyhnders but are more hkely to be rncluded m the much largar rnf1ltrat1on cyhnders The legume effect on the so1l phys1cal cond1t1on may be relatad to the strong rncrease m earthworm b1omass m the presence of a legume as earthworms create vertical channels 1n the so1l thus st1mulatrng water rnflltrat1on Another explanat1on may be found m the fact that many legumes espee~ally A pmtot have a relat1vely coarse root system comparad w1th that of grasses The channels left behrnd m the so1l by old legume roots may contnbute to creatrng relat1vely larga pares throughout the prof1le prov1drng condu1ts for by pass flow of water The more rap1d water ralease w1th rncreasrng pF m the h1gh stockrng rate treatments of the 8 humtdtcola + 1 A pmtot expenment than m the low stockrng rate treatments rnd1cates a largar fract1on of relat1vely large so1l pares 1n the formar (Tabla 19) Total poros1ty was not affected by stockrng rate Th1s may be relatad to a h1gher root turnover due to the more frequent and thorough defohat1on resultrng m a relat1vely large number of old root channels At h1gher stockrng ratas th1s effect may be counteracted by compact1on effects However desp1te th1s rncrease m large pare fract1on water rnf1ltrat1on rate had decreased m the h1gh stockrng rate treatments wh1ch may reflect the earher ment1oned d1fferences of scale 1n the methods used Conclus1ons Although the phys1cal cond1t1on of a s01l after convers1on of nat1ve savanna rnto pasture 1s hkely not to changa as dramat1cally as m case of convers10n of pnmary forest 1nto pasture the results presentad here have 1mphcat1ons for so11 management The so1ls under 1mproved pasture generally had a lower calculated unsaturated hydrauhc conduct1v1ty than under nat1ve savanna at g1ven value of O or formulated d1fferently were wetter at a g1ven value of K Th1s 1mphes that w1th a g1ven amount of rarn thus at a particular K under cond1t1ons of steady rnf1ltrat1on the pastura s01ls are more susceptible to berng damaged by ammal tramphng or t1llage On the other hand addrng a legume to a pasture pos1t1vely affects the s01l phys1cal cond1t1on The h1gher water rnf1ltrat1on rata under grass legume pasturas may add to reducrng the nsk of so1l loss by water eros1on A caut1onary note about the use of the CENTURV soll orgamc matter model for ecosystems on trop1cal low P solls A J Gl}sman A Oberson H Ttessen (U Saskatchewan) and D K Frtesen (IFDCJ lntroduct1on One of the most w1dely used models on orgamc matter dynam1cs m a s01l plant ammal system 1s the CENTURY model Although developed for use m temperate ecosystems the model has now also been adoptad for use m the trop1cs by the Trop1cal S01l B1ology and Fert1hty (TSBF) Programme and the Slash and Burn Pro¡ect However many trop1cal ecosystems are on 197 strongly phosphorus sorbmg solls where P and not N 1s the hm1tmg nutnent Th1s has 1mportant 1mphcat1ons for the apphcat1on of CENTURY Modelhng of nutnent cychng reqUJres a deta1led knowledge about the s1ze and transformat1ons of the vanous orgamc and morgamc nutnent pools and the flows between them For carbon and mtrogen a large body of data IS ava1lable on these top1cs but for phosphorus mformat1on 1s far more scanty and Wlth a few except10ns 1s umquely obtamed from solls of the temperate zone wh1ch are not as strongly P sorbmg as certam trop1cal s01ls Data on mteract1ons between orgamc and 1norgamc P are rare We ran the model for an ac1d low P Ox1sol w1th nat1ve savanna vegetat1on m the eastern plams of Colombia S1mulat1ons of SOM content were maccurate both m quant1ty and quahty desp1te the fact that deta1led so1l P fract10nat1on data were ava1lable to parametenze the model As we beheve that th1s could be related to poor s1mulat1on of soll P dynam1cs by the model we analyze here the P submodel of CENTURY w1th spec1al reference to 1ts use for trop1cal low P solls The P cycle and plant avallable P Phosphorus ex1sts m the soll m morgamc (P 1 and orgamc form (P 1 both of wh1ch are mvolved m the cychng of P among so1l plants m1croorgamsms and ammals F1gure 20 shows a conceptual P cycle w1th 1ts components Whereas at least for the temperate zone the contnbut10n of P, from orgamc phosphorus mmerahzat1on to plant ava1lable P, 1s cons1dered of mmor 1mportance w1thm a smgle growth cycle the need to look at P cychng (consequently mcludmg P 1 rather than ava1lable P pool s1ze has been recogmzed for perenmal plants and natural ecosystems Under trop1cal cond1t1ons P was shown to contnbute S1gn1f1cantly to crop P uptake and several stud1es md1cate that orgamc P m1ght play a major role 1n P fert1hty of trop1cal so1ls Assessment of P, and P by sequent1al extract1on The only approach wh1ch has g1ven moderate success m the evaluat1on of short term ava1lable P and P 1s a sequent1al P fract1onat10n method (F1gure 211 lt a1ms at quant1fymg lab1le (plant ava1lablel P Fe and Al assoc1ated P Ca assoc1ated P, as well as lab1le and more stable forms of P usmg the followmg extracts res1n O 5M NaHC03 O 1M NaOH 1M HCI and hot concentrated HCI Due to the react1v1ty of P w1th the soll mmeral phase determmat1on of a potent1ally mmerahzable P pool analogous to the m1nerahzable N or S pools measured by mcubat1on and leach1ng techn1ques 1s not feas1ble The natura of d1fferent extractable P pools 1s even less well defmed than that of the P, pools The1r turnover and ava1lab1hty often depend on the mmerahzat1on of C durmg wh1ch P 1s released as a by product although soluble P w1ll be rap1dly mmerahzed by s01l enzymes Th1s complex s1tuat1on means that there 1s presently no sat1sfactory method for measunng ava1lable P beyond the rather emp1ncal sequent1al extract1on combmed w1th conceptual models and poss1bly separata organ1c matter stud1es The structure of the P submodel of CENTURY The P submodel of CENTURY shown m F1gure 22 cons1sts of an orgamc P andan 1norgamc P sect1on (respect1vely the left and nght parts of the f1gure) the two bemg hnked when morgamc P enters the orgamc cycle v1a P uptake by roots or m1crobes or P adsorpt1on onto orgamc matter 198 ~ .¡:; !ti 4 ..,- o::J o E ::J () 2 o 1 S A -- -- • --- Bd 028 o 07 o-- Bd/Pp 040 008 r,. Savanna 024 004 .,.,.. ___. .....--· ./ ... ...-/ r,. r,. ,./ A ./ 1:>. .A e// A A ./ (l. JlY A A A /e" A • A A 'A (A) S - •--- Bh+2 023 lE-- Bh/Ap+2 020 •-- Bh+4 017 e- - Bh/Ap+4 029 tJ. Savanna 024 A 003 o 08 002 o 05 004 ---------- .......... --------- A A 6 .A A At:> ..e-A _..._._....--.--...- (B) ¡,- o 10 20 30 40 50 60 70 80 90 T1me (m1n) T1me (mm) F1gure 19 Cumulat1ve water 1nf1ltrat1on over a 90 m1nute penod 1n an Ox1sol under pasture or nat1ve savanna (A) = B decumbens +/ P phaseoiOides (B) B hum1d1cola +/ A p1nto1 For an explanat1on of the treatment cedes see Table 18 S = sorpt1v1ty (cm mm 112) A= termmal water mfiltratlon rate (cm mm 1) 90 LIVING rn~erbhzer 1 1 Pnmary ~ ITlmerals 1 r--------,' . . Lab1le morgamc P t ,., 1 Secondary m1neral P o ~ t 1 Occluded P ·-, F1gure 20 A conceptual model of the P cycle 1n an ecosystem Phys1cally or chem1cally rotected oraamc P Po DOOIS Po fract1ons P1 fract10ns PI pools 1 05gsoll 1 ~ 1 water extractable Po 1+- extrae! Wlth res.n stnp ---+1 resin extractable Pi in water ~' ~ 1~ NaHC03 extractable Po +- extrae! Wlth NaHC03 ---+ NaHC03 extractable Pi 1 "' 1 o m 1 "' ~¡ -1 ~ 1- 1 1 1 NaOH extractable Po 1+- ---+1 1 1 o 1 extrae! wlth NaOH NaOH extractable Pi 1 r Q) 1 e 8 ~ :::1 Q.l Q. >-1 111 ~1 extract With dllute HCI ---+ dllute-HCI extractable Pi 1; n m 1-~1 ~1 ~ 1 01 hot concentrated HCI extrae! Wlth hot hot concentrated HCI 1 t< 1 >-1 extractable Po +- concentrated HCI ---+ extractable Pi 1 13' llS ~1 1 1 lg il ~ '~ [_ residue pt 1 Figure 21 The P fractions obta~ned dunng a sequential P fractionation and an approXImation of how to link P fractions to P pools "' o c.:> 1 UVIHG SHOOT P : .1 LMNGROOTP l 1 ~ (llllago) go) (tilla, 1 STANDING DEAD P l (ti llago) + SURFACE UTTER SOILLIT1ER . 1 METABOUC P : STRUCTURAL P STRUCTURALP 1 METABOUCP ; 1 ISURFACE MICROBE p 1 AC'nVE ORGANIC P 1 1 ""' ~ SLOW ORGANIC P 1 PASSWE ORGANX: P i LEACHJNO -, 1 F1gure 22 The structure of the P submodel of CENTURY PAAENTP FalnLIZERP 1 LA81LEP 1~ SORBEOP 7 ldOODGrlaviiiBI ,¡. ~ lABILE SORBED P BTROHGLY SORBED P JLAYERZI ...... (LAYER3) 1 IL&YI!R 1 1 OCCLUDEDP Ltmttattons to the apphcatton of the P submodel P, pools vs P, fracttons The morgamc P sectton of the CENTURY modal mcludes ftve P, pools labtle sorbed strongly sorbed occluded and parent P the stze of whtch has to be esttmated for a proper parametenzatton of the model lt would be tdeal tf one could stmply asstgn dtfferent P fracttons from a sequenttal fracttonatton to dtfferent P pools However a poolts usually defmed as a homogenous enttty of substance charactenzed by tdenttcal kmettcs of transport and transformat10n whereas an extractable P, fractton ts only charactenzed by tts extractabthty wtth a certam extractan! wtthout thts latter necessanly bemg an mdtcatton of havmg stmllar kmettcs or even structural properttes Phosphorus present m a certam extractton fractton therefore m ay ongmate from vanous P pools and consequently constst of P forms wtth dtfferent plant avatlabthty By companng data from sequenttal P fracttonattons wtth expenmental observattons on P avatlabthty and transformat10ns emptncal relattonshtps can be estabhshed correlatmg fract10ns wtth pools and esttmatmg net transfers between them The P status of a s01l can then be charactenzed relattve to a conceptual model of P pools and thetr transformattons For the dommant sotl types of the temperate zone a substanttal databa se extsts on P fract10natton m relatton to P avatlabthty but for the low P sotls commonly found tn the troptcs such a database ts only begmmng to bUIId Thts makes tt vtrtually tmposstble to parametenze thts part of the model rehably even tf sequenttal P fracttonat10n data are avatlable Flows between P, pools In the CENTURY model the flows between the P pools are formulated as ftrst order reacttons of the type Flow = rate constant * pool SIZB * ablot1c reductlon factor [1) Whereas deftntng the P pools already poses constderable problems esttmattng the rate constants for the transfers between these pools ts even more uncertam The dtfftculty ts due to the numerous compettng reacttons morgantc and orgamc whtch are occurnng both concurrently and consecuttvely m sotl systems Reacttons tn whtch P moves from the sohd phase or from organtc pools to morgantc pools may be expected to follow dtfferent kmettc equat10ns Enzyme medtated dephosphorylatton reacttons are hkely to follow htgher arder kmettcs than the ftrst arder processes assumed m the modal Bestdes these hmttattons the greatest dtfftculty remams esttmatmg ktnettc parameters of tndtvtdual processes m the complex mtheu of the sotl The abtottc reductton factor (range O 1) m equatton [ 11 conststs of a temperatura and a motsture related componen! and ts the same factor used as multtpher on the organtc matter decomposttton rate However the effect of temperatura and motsture on the rates of phystc chemtcal processes dnvmg phosphate ton transfer between morgamc pools ts very dtfferent from the effect on btologtcal and btochemtcal processes whtch mvolve the decomposttton or transformatton of orgamc P compounds or the asstmtlatton of morgamc P mto orgamc pools Rates of btologtcal/btochemtcal processes governed by mtcroorgantsms often double for each 1 0°C nse m temperatura but there extsts no venftcatton that stmtlar rules apply for such processes tn sotl systems Rates of P uptake by plants may be expected to exhtbtt a Mtchaehs Menten dependence modtfted by motsture and temperatura although factors controlhng supply of P to the root surface (dtffuston water flow P solutton concentratton etc l may be more tmportant m controlhng uptake than tnflux ktnettcs 204 Sorpt1on equat1on The eqUihbnum relat1onsh1p between lab1le P and sorbed P IS descnbed 1n the model 1n terms of an equat1on mvolvmg coefflc1ents for sorpt1on aff1n1ty (sorpaf) and sorpt1on max1mum (sorpmx) Thus accordmg to the model s source code (vers1on 4 0) the amount of lab1le P 1s calculated as e = O 5 • sorpmx • (2 O - sorpa.f) (2) b = sorpmx - (lab1le P, + sorbed P,) + e (3) lab1le P, = O 5 • ( -b + yb' + 4 • e • (lablle P, + sorbed P1)) (4) The theoret1cal bas1s for choosmg th1s formulatlon IS not clear and 1t seems to bear no relat1onsh1p w1th standard 10n adsorpt1on equat10ns such as those of LangmUir or Freundhch Furthermore methodolog1es for charactenzmg the relat1onshlp and defmmg the coeff1c1ents are not documentad 1n the model Fract1on of lablle P1 ava1lable to the plant In the eENTURY model the fract1on of lab1le P that 1s avallable for plant uptake has been formulated as a hnear funct10n of the mmeral N pool s1ze (h1gher P ava1lab1hty at h1gher mmeral N levels) dehm1ted by a m1mmum and a max1mum ava1lable fract1on No explanat1on 1s g1ven for th1s very unconvent1onal relat10nsh1p In strongly sorb1ng so1ls P ava1lab1hty 1s mamly determ1ned (apart from plant rootmg charactenst1cs) by physlcochemlcal processes and a dependence on mmeral N ava1lab1hty 1s very unhkely Moreover as plant product1on on these s01ls 1s generally P hm1ted mstead of N hm1ted such a relat1onsh1p does not apply Res1due part1t1omng between structural and metabohc pools The part1t10mng of plant and ammal res1dues between the metabohc and structural res1due pools 1s formulated 1n the model as a linear funct1on of the hgnm mtrogen rat1o (l N) FractJon of restdues gomg mto metaboltc pool = O 85 O 018*L N [5) Th1s relat1onsh1p 1s based on data from a large number of plant res1due analyses separatmg cellulose and hgmn as structural components from water soluble metabohc components However plants from the mfert1le South Amencan savanna so1ls have developed nutnent conservmg strateg1es by rap1dly hgmfymg and L N rat1os 1n plant res1dues can be so h1gh that the metabohc fract1on as calculated w1th th1s equat1on becomes negat1ve eJearly the relat1onsh1p cannot be used for res1dues w1th such h1gh L N rat1os One m ay also wonder whether 1t should be expected at all that such a relat1onsh1p ex1sts In systems wh1ch are strongly P hm1ted such as natural savanna vegetat1on on trop1cal Ox1sols and Ult1sols a relat1onsh1p wh1ch IS based on the N content of plant or ammal res1dues seems not very log1cal Nutnent contents of structural and metabohc pools When part1t1omng res1dues between the metabohc and structural pools the latter 1s ass1gned a f1xed e P rat1o of 500 leav1ng the remammg res1due P for the metabohc pool However 1n analyzmg leaf and root htter from grasses legumes and nce on an Ox1sol 1n the eolomb1an savannas whole htter e P rat10s up 205 to 2700 were found In the CENTURY model th1s would result 1n negat1ve P concentrat1ons 1n the calculated metabohc fract1on For mtrogen s1m1lar problems ex1st be 1t 1n a less extreme from In pnnc1ple the opt1on ex1sts 1n the modal to mod1fy the C Por C N rat1os of the structural matenal However 1n order to cover the most extreme values reportad above a C P rat1o of around 3000 would be needed No expenmental data are ava1lable to assess how reahst1c such a value 1s and how 1t vanas among spee~es soll cond1t1ons seasons etc lncreas.ng such an 1mportant .nput parameter s1x fold w1thout any support.ng data clearly bnngs 1nto quest1on the rehab1hty of the part1t1onmg of orgamc P mputs m the modal Structure of the orgamc P submodel g The pools and flow structure of the orgamc matter submodels of CENTURY are the sama for orgamc carbon mtrogen and phosphorus the three bemg hnked v1a C N and C P rat1os Such a structure 1mphes that orgamc C N and P are stab1hzed and mmerahzed m the sama way the s1ze of the flows relat1ve to each other dependmg only on the nutnent content of the decomposmg matenal and the potent1al 1mmob1hzat10n of morgamc nutnents However 1t has been suggested that orgamc C and N are stab1hzed together and mmerahzed through b1olog1cal processes whereas P may be stab1hzed mdependently of tha mam orgamc mo1ety and mmarahzed through b1ocham1cal procassas Th1s hypothes1s 1s supportad by data on phosphorus fract1onat1on and m1crob1al b1omass from Ox1sols m Colombia (Tabla 20) Table 20 Nutnent content of vanous fract1ons from a sequent1al P fract1onat1on and of m1crob1al b10mass m an Ox1sol under nat1va savanna So1l fract1on P' p ' e C/P 11 C/P 11 t CIN' =====pgg' ==== = Water/resm 2 1 1 5 510 358 * 10 5 NaHC03 1 5 13 2 1770 137 99 Na OH 13 9 53 6 6080 111 13 4 HCI 07 * HCI hot 68 o 16 7 2670 219 74 Res1due 45 9 8800 * 191 36 5 Whole SOII 86 2 19830 237 91 17 1 M1crobes 96 470 * 50 93 P P and P = morgamc orgamc and total phosphorus Tha rat1o averagas were calculated as gaomatnc maans of treatments so values may d1ffer from thosa calculated by d1v1dmg tha md1v1dual alemant means 206 The orgamc matter collected m the water fract1on after resm extract1on as f1rst step of the sequent1al P fract1onat1on can be cons1dered as easlly decomposable and should thus be part of the act1ve SOM pool The same may hold for the orgamc matter m the NaHe03 fract1on For the eolomb1an Ox1sol of Tabla 20 these two fract1ons together conta1n ca 2280 pg e and 1 5 pg P per g s01l The act1ve SOM pool1s often est1mated as two to three t1mes the amount of m1crob1al b10mass so as to also account for m1crob1al metabohtes Usmg the m1crob1al b1omass data of Tabla 20 th1s would mean an act1ve SOM pool of 940 141 O pg e and 19 29 pg P per g so1l The two est1mat1on methods for the act1ve SOM pool thus y1eld qu1te d1fferent results espec1ally concermng the e P rat1os of the maten al ( 137 358 when est1mated accord1ng to the sequent1al P fract1onat1on and 50 accordmg to the m1crob1al b1omass est1mate) Th1s makes 1t doubtful whether mdeed the m1crob1al b1omass and the eas1ly extractable orgamc matter from the sequent1al fract1onat1on can be cons1dered to form one pool w1th one common turnover rate In the modal a e P rat1o between 30 and 80 1s suggested for the act1ve SOM Wh1le the m1crob1al e P ratiO falls w1thm th1s ranga 1t IS far too low for the H20 or NaHe03 extractable fract1ons The very h1gh e P rat1os of the extractable fract1ons md1cate that the1r decompos1t1on may be largely governed by b1ochem1cal rather than m1crob1al processes lt may thus be too restnct1ve to calculate orgamc P flows m the same way as orgamc e and N flows Conclus1ons and recommendat1ons Whereas m so1ls from the temperate zone N 1s generally the plant growth hm1t1ng element many tropical so1ls are strongly P hm1ted For modehng SOM dynam1cs th1s 1mphes that accurate representat1on of the P and P pool structures and transfer kmet1cs 1s cruc1al for the modal s apphcat1on to such solls The structure of the organ1c P sect1on of eENTURY 1s basad on the hypothes1s that orgamc P and orgamc e are stab1hzed and mmerahzed together m an kmet1cally Jdent1cal manner Th1s hypothes1s 1s quest10nable The bas1c structure of the modal s morgamc P sect1on 1s largely m agreement w1th expenmental data but the vanous equat1ons used lack theoret1cal JUStlfJcatJon or do not appear to be basad on current knowledge of processes and kmet1cs 8oth sect1ons suffer from the fact that 1t 1s v1rtually 1mpossJble to properly est1mate pool s1zes and rate parameters even 1f detallad s01l P fract10nat1on data are ava1lable Several of the equat1ons used m the modal for est1matmg the part1t1omng between P pools are based on data from a hm1ted ranga of so1ls m the temperate zone Applymg these to solls w1th strongly d1fferent charactenst1cs as e g many trop1cal so1ls results m extrapolat1ve calculat1ons beyond the ranga for wh1ch the relat1onsh1ps were testad Th1s can (and does) lead to totally erroneous results We conclude therefore that great care should be exerc1sed 1n usmg the eENTURY model for trop1cal low P solls Th1s 1s of spec1al 1mportance to the part1c1pants of the Trop1cal S01l 810logy and Fert1hty Program and the Slash & Burn Pro]ect both of wh1ch adopted the model for the1r orgamc matter stud1es m the trop1cs We recommend the followmg research 1mperat1ves w1th respect to the P submodel 1 Data are urgently needed to bulld up a database wh1ch would allow us to equate P fract1ons w1th P pools both orgamc and morgamc Bes1des chem1cal P fractJonatJon Jsotop1c dJiut1on kmet1cs should be used to obtam mformat1on on the mtens1ty quant1ty and buffer capac1ty of d1verse so1ls w1th respect to avallable P Blolog1cal parameters related to P dynamJCs such as m1crobJal b1omass P turnover or phosphatase act1v1ty 207 should also be mcluded 1n th1s database 2 Pools are hnked by transfer rates whose present model values were est1mated usmg data from a temperate zone s01l P database There 1s a need to determme transfer rates for a range of so1l types both temperate and tropical to assess whether the use of umversal values 1s appropnate As the present values were partly approx1mated through a tnal and error approach the1r est1mat1ons could be 1mproved cons1derably by usmg tracer stud1es desp1te the fact that measurements w1th 32P and 33P are hm1ted to a max1mum penod of about 3 to 6 months Such stud1es would at the same t1me g1ve mformat1on about orgamc P turnover rates 3 The theory that P 1s stab1hzed and mmerahzed mdependently of the mam orgamc mo1ety would have 1mportant 1mphcat1ons for the bas1c structure of the P submodel lncubat1on stud1es assessmg e N and P mmerahzat1on m the same samples are reqUired to test th1s hypotheSIS 4 The well known ab1hty of m1croorgamsms to adapt to changes m env1ronmental cond1t1ons could 1mply that the range of m1crob1al e P rat1os found m low P solls 1s h1gher than that commonly found m non P hm1ted so1ls Results obtamed on eolomb1an Ox1sols md1cate that th1s mdeed m1ght be true More data on th1s subJect are needed 5 The concentrat1on of orgamc P m the so1l solut1on of a trop1cal Ox1sol may be much h1gher than the concentrat1on of 1norgamc P wh1ch w1th the h1gh amount of ramfall m hum1d trop1cal zones 1s hkely to result m P leachmg Although th1s opt1on does ex1st m the model very few expenmental data are ava1lable to quant1fy the 1mportance of such a P flow lt 1s suggested therefore to mclude the measurement of d1ssolved P flows 1n leachmg and runoff stud1es Dynam1cs and Short Term Effects of Earthworms m Natural and Managed Savannas of the Eastern Pla1ns of Colomb1a M Ftsher (TL) J J Jtménez (U Complutense) T Decaens (U Parfs) A G Moreno (U Complutense) J P Rosst (ORSTOM/U Parfs) P Lave/le (ORSTOMJ R Thomas (TLJ lntroduct1on Stud1es on so1l b1ology (espec1ally macrofauna) were 1n1t1ated at the eORPOieA eiAT eanmagua Research Stat1on on the eolomb1an Eastern Plams 1n July 1993 The work started w1th a prehmmary survey of so1l fauna commumt1es m contrast1ng land facets and product1on systems us1ng the bas1c TSBF methodology (Decaens 1994) The survey was followed by the 1n1t1at1on of more deta1led stud1es of taxonomy of the spec1es mvolved and the dynam1cs of the1r populat1ons 1nclud1n9 momtonng of the cult1core expenment ObJect1ves The ma1n obJeCtlves of the program m the long term are To evaluate the d1vers1ty and ecolog1cal funct1ons of earthworm spec1es present m the area and 208 descnbe the dynam1cs of the1r commumt1es 1n arder to mampulate earthworm commumt1es to 1mprove the product1v1ty and susta1nab1hty of agrosystems denved from the nat1ve savanna A senes of 1ntermed1ate ob¡ect1ves were 1dent1f1ed Evaluate the 1mpact of d1fferent types of land use on the abundance and structure of so1l macrofauna commumt1es Descnbe the spat1al and temporal patterns of earthworm commumty structure 1n the nat1ve savanna and 1n contrast1ng management systems Descnbe the dynam1cs of recolomzat1on of earthworm commumt1es 1n a success1on of annual crops and 1mproved pastures Ouant1fy the effect of earthworms on parameters of so1l funct10n such as select1ve 1ngest1on of orgamc and m1neral part1cles and release of ass1m1lable N and P 1n casts Rat1onale Earthworms are a dom~nant component of so1l macrofauna 1n natural savannas and denved pastures and cultures at Canmagua (Table 21) Earthworm b1omass (EB) however 1s s1gmf1cantly lower 1n these ecosystems than 1n many other comparable ecosystems In the gallery forest EB IS 3 4 g fresh we1ght m 2 wh1ch 1s Similar to that at Lamto lvory Coast but lower than 1n trop1cal ra1nforests of Amenca (8 7 g f w m 2 ) EB 1n nat1ve savannas at Canmagua (30 5 g f w m 2 ) 1s lower than at Lamto whlle 1n most grazed savannas EB 1s less than 1 O g fresh we1ght m 2 lmproved pastures have much h1gher EB w1th a max1mum value 1n Brachtarta decumbens and Puerarta phaseolotdes assoc1at1on (41 2 g fresh we1ght m 2 ) comparable to other trop1cal reg1ons 1 e 1n Mex1co (35 8 to 55 5 g fresh we1ght m 2 ) and 1n lnd1a (30 2 to 56 g f w m 2 ) The 1ncreas1ng amount of pasture and legume product1on and nutnent supply of cattle faeces prov1des the necessary energy to support earthworm populat10ns at h1gher levels In nce crops EB 1s drast1cally reduced (2 3 g f w m 2 ) as 1s commonly observed 1n annual trop1cal crops due to the use of pest1c1des the reduct1on of so1l orgamc matter and the phys1cal effects of soll t1llage lt 1s therefore 1mportant to understand the dynam1cs of these changes and the1r 1mpact on such bas1c processes of s01l fert1hty as well as on so1l orgamc matter dynam1cs and so1l phys1cal structure Research Program The planned research compnses 1 A rap1d samphng of earthworm commumt1es tw1ce a year to draw maps of the spat1al d1stnbut1on 1n the CUL TICORE expenment at the d1fferent stages of success1on lt w1ll be poss1ble to follow the spat1al (w1th1n and across plots) and temporal (across seasons and alternat1ng types of land use) patterns of the dynam1cs of earthworm commumt1es 209 Tabla 21 Lost of specaes of earthworms found an dofferent ecosystems at Canmagua Specoes Famaly Locatoon' Ecologacal category• And1odrt/us sp Glossoscolecadae NS IP GF Endogeoc And1orrhmus sp 1 Glossoscolecodae NS IP Endo anecoc And1orrhmus sp 2 Glossoscolecodae NS Endo anecoc Mamodrtlus sp 1 Glossoscolecodae NS IP Anecoc Mart1odn/us sp 2 Glossoscolecodae CG Anecoc P corethrurus Glossoscolecodae AP Endogeoc Small polyhumoc 1 Glossoscolecodae NS IP GF Endogeoc Pheret1ma sp Megascolecodae CL Epogeoc D1chogaster sp 1 Octochaetodae NS IP Epogeoc Ocnerodnhdae 1 Ocnerodnhdae NS IP Endogeoc Epageoc 1 ? NS IP GF Epogeoc Epageac 2 ? CL Epageac Epogeac 3 ? GF Epogeoc 1 NS natove savanna IP omproved pastura GF gallery lorest AP Alrocan palm culture CL Caromagua lake CG Cano Gloroa near Caromagua Anecoc love on the sool but leed on the sool surlace Epogeoc love and leed on the sool surlace Endogeoc love and leed on the sool 2 An accurate 15 month samphng of natural populatoons to detaol theor demographoc parameters and seasonal and spatoal domensaons of these dynamacs Two dofferent types of land use are comparad the natove savanna and the assocoatoon legume pastura (Brachlana decumbens + Kudzu) chosen because of the prevaous survey results (Decaens 1994) (Fogures 23 and 24) 3 Studaes under laboratory condotoons of severa! aspects of earthworm ecology The dofferent specoes are culturad to measure growth (ate fecundoty mortahty angestoon of soal organoc and manera! partacles Studoes on N and P dynamocs are beang cerned out an collaborataon woth two thesos students of the Unoversodad Nacaonal Palmar a R Thomas and D Fnesen 210 "' - - g m2 70 60 50 Savanna ~ lntroduced pastures --B decumbensl kudzu Grazed 40 30 ' ' --~---~ _ _. Stock1ng rata ~1 1 Annual crops Gallery Ungrazed 1 year 20 4 1 11 ~ forest - ~ /~lgh Med1um -- - -- ---- 10 o F1gure 23 Total b1omasss of so1l macrofauna under d1fferent types of land use al Canmagua • Others CiJ Ants liiJ Term1tes O Earthworms N -N lnd1v1duals m 2 (thousands) 5 forest 4 3 2 1 Savanna /~ Ungrazed Grazed ~-...._ Bumed Stockmg rate /1 '1 year Med1~m 6m- -\ ¡-,..... ~HGh lntroduced pastures 1 B decumbensl kudzu Annual crops ~l.. R1ce o u 1 1 11 11 11 1 1 11 1 1 11 1 1 IQ 1 1 1 1 ' 1 1 1 1 1 1 1 1 1 F1gure 24 Total dens1ty of so1l macrofauna under d1fferent types of land use at Can magua • Others liiil Ants El Term1tes O Earthworms These three deta1led stud1es w1ll defme the 1mpact of earthworms on total ecosystem funct1on They w1ll also explam effect of land management on the dens1ty and vanat1ons 1n d1fferent spec1es of earthworms RESULTS a) Taxonomy The b1od1vers1ty of the study s1te 1s very 1mportant So far 1 O to 1 5 spec1es have been found the doubt ansmg because not all speCies have been pos1t1vely 1dent1f1ed (Table 211 The taxonom1c study mcludes natural and managed savannas as well as the gallery forest The nat1ve savanna has e1ght spec1es whlle the legume grass assoc1at1on has seven Although the latter system 1s a d1sturbed one the b10d1vers1ty seems not to be affected much These pasturas are now 15 years old and whlle 1t 1s hkely that there was a drast1c decrease of so1l fauna (espec1ally earthworms) when they were estabhshed the b1omass of earthworms m them 1s now the h1ghest recordad m Canmagua pnnCipally due to a larga populat1on of Marttodnlus sp 1 The most 1mportant spec1es under study are Marttodrtlus sp 1 (sp nov 7) A large (up to 30 cm m lengthl dark grey spec1es Approx1mately 1 O to 15 md1v1duals m 2 and max1mum we1ght for an adult 1s 25 g (fresh we1ghtl Andtorrhmus sp 1 (sp nov7) Up to 30 cm 1n length also but thmner than Marttodrtlus L1ght grey 1n color less than 1 O md1v1duals m 2 and max1mum adult we1ght of 15 g Andtodrtlus sp (sp nov?) No more than 14 cm m length Wlthout p1gmentat1on Up to 1 O md1v1duals m 2 no more than 6 g max1mum we1ght for an adult Small polyhum1c (sp nov 7) 5 6 cm m length w1thout p1gmentat1on More than 100 mdiVIduals m 2 max1mum adult we1ght about O 20 g Ep1ge1c 1 (sp nov ?) 5 6 cm m length dark red p1gmentat1on on the dorsal part of the body No more than 20 1nd1v1duals m 2 max1mum adult we1gh of O 1 5 g Ocnerodnhdae (sp nov ?l 2 3 cm 1n length w1thout p1gmentat1on 20 30 md1v1duals m 2 max1mum adult we1ght O 015 g Th1s spec1es hves and feeds on the casts that f1ll the long and large gallenes of anec1c spec1es such as Marttodnlus where there 1s much h1gher quant1ty orgamc matter than m the surroundmg s01l b) Populat1on dynam1cs Smce October 1993 the 1mproved pasture system has been sampled to determme the populat1on dynam1cs of the earthworm commumty (Tabla 221 S1m1lar stud1es were started m the nat1ve savanna m March 1994 The expenmental des1gn cons1sts of excavatmg f1ve 1 m2 quadrats to 50 cm depth m both systems and handsortmg the macrofauna 1n the f1eld The sample 1s subd1v1ded mto f1ve strata m 1 O cm mcrements All earthworms m the samples are 213 placed m 10% formaldehyde and later separated m the laboratory mto speCies end the adults JUVemles and cocoons counted and we1ghed Table 22 Dens1ty and b1omass of earthworms m nat1ve savanna and a f1fteen year old mtroduced pastura between October 1993 and May 1994 (n = 5 samples/system/month) Nat1ve savanna lntroduced pastura Month Dens1ty B1omass Dens1ty B1omass md1v1duals m 2 g fresh wt m 2 md1v1duals m 2 g fresh wt m 2 1993 October 135 18 1 November 86 21 8 December 38 16 2 1994 March 25 08 46 53 7 Apnl 107 10 3 102 77 6 M ay 111 94 216 110 o And1odrtlus and And1orrhmus apparently have no strategy or resostance form so theor populat1ons were greatly reduced and the surviVors imtoated reproduct1on early m the wat season (generally on May) The so1l was very hard dunng January and February and regular samphng was abandonad Nevertheless a hm1ted samphng of one 1 m2 quadrat to 11 O cm was done 1n the 1ntroduced pastura 1n February after two months w1thout ra1n to determ1ne the d1fferent strateg1es used by the d1fferent spec1es dunng the dry season The mean so1l water content was about 15 % on a dry so1l bas1s Ali of the small polyhum1c Glossoscolec1dae were 1n cocoons wh1ch allow them to res1st the adverse mo1sture cond1t1ons The whole populat1on of Mart1odnlus was m d1apause apparently obhgatory all the adults and JUVemles rolled up m aest1vatmg chambers as much as 90 cm deep 1n the so11 Th1s spec1al charactenst1c of an anec1c (s01l hvmg but surface feedmg) spec1es allows them to reduce otherw1se h1gh mortahty dunng the unfavorable cond1t1ons of the dry season Larga (25 mm lengthl cocoons from wh1ch the young would subsequently emerge compnsed the greater part of the populat10n These were mamly m the 30 40 cm stratum m contrast to the JUvemles and adults at greater depth The number of md1v1duals 1n the early wet season (March May) was not greatly d1fferent between the nat1ve savanna and the mtroduced pastura but there was more than a tenfold mcrease m b1omass 1n the latter comparad w1th the savanna (Tabla 23) 214 Table 23 Mean dens1ty and b1omass of earthworms m nat1ve savanna and mtroduced pasture over three months March Apnl and May 1994 Nat1ve savanna lntroduced pasture Dens1ty B1omass Dens1ty B1omass md1v1duals m 2 g fresh wt m 2 md1v1duals m 2 g fresh wt m 2 81 68 121 804 Another d1fference between these two systems 1s related to the number of turnculated casts wh•ch are the tower hke sp1ral casts depos1ted over the so1l surface by earthworms At Canmagua these structures are produced by MartJOdfllus and they m ay be up to 1 O cm m he1ght Table 24 shows the number of turnculated casts m the nat1ve savanna and m the 1mproved pasture Table 24 Number of turnculated casts m ten 1 m2 random samphngs 1n the nat1ve savanna and 1ntroduced pasture at the begmmng of the wet season at Canmagua Cond1t1on Dry Fresh Dry base/fresh Nat1ve savanna 91 2 3 lntroduced pasture 268 40 59 Fresh casts md1cate a real presence of an md1v1dual of MartJodfllus and may be taken as ev1dence of act1v1ty of th1s spec1es The 1mproved pasture system presents a product1on of fresh casts about 20 t1mes greater than the nat1ve savanna wh1ch appears to be cons1stent w1th the b1omass data el Laboratory stud1es Cultures of the maJor spec1es have been done under laboratory cond1t1ons to evaluate the•r so1l mgest1on rates (usmg bas1c methodology) and then extrapolate these data to f1eld cond1t1ons Knowledge of the overall annual earthworm commumty populat•on allow us to est1mate the total so1l mgest1on of th1s commun1ty per year and per hectare The two spec1es AndJOdfllus and MartJodnlus have the greatest mtake of so1l at a water content of 34 % on a dry so1l bas1s that 1s near pF 2 5 eqUJvalent to f1eld capac1ty The global consumpt1on on a yearly bas•s reqUJres data on the populat1on dynam1cs throughout the year wh1ch 1s not yet available Over the 6 month penod October to Apnl the global consumpt1on of a populat1on of AndJodfllus 1s about 70 t dry so1l ha 1 and for MartJodfllus 1s nearly 400 t dry s01l ha 1 (Table 25) The numbers could be expected to be more than double when the whole wet season 1s mcluded 215 Table 25 So1lmgest1on rates and global consumpt1ons of two spec1es of earthworms from Canmagua Spec1es And1odnlus Mart1odnlus S01l mgest1on rate' g dry SOII day 1 54 21 3 ' At so1l water content of 34 % on a dry so1l bas1s di Spat1al d1str1but1on of commumt1es (GEOSTATISTICS) Global consumpt1on' t dry s01l ha ' 6 months 1 70 390 A prehmmary samphng has been done m the Brach1ana decumbens/ P phaseolo1des (Kudzul pasture (September 19931 usmg an expenmental des1gn cons1stmg of 64 40x40x 15 cm samples taken every ten meters on a 70x70 gnd Of the ma1or spec1es dlstmgUished only populat1ons of the small polyhum1c Glossoscolec1dae had a spat1al structure These populat1ons had a d1stnbut1on m patches a few tens of meters m d1ameter w1th d1stmct patterns for adults and JUVemles The expenmental des1gn d1d not allow the def1mt1on of structures m other populat1ons poss1bly due to the¡r low dens1t1es By November 1994 there were substant1al changas m the spat1al d1stnbut1on (F1gs 25 271 Geostat1st1cs have also been apphed observat1ons of d1stnbut1on of turncules at the so1l surface as an md1rect measure of populat1ons These data correspond well w1th those of excavated samples and conf1rm the patchy populat1on d1stnbut1on The next step 1s to determme whether env~ronmental factors regulate the patch1ness or 1f demograph1c factors are respons1ble Carbon storage deep 1n the sod by mtroduced pasturas 1n the South Amer1can savannas M F1sher R Vera 1 Rao J Sanz C Lascano R Thomas M A yarza L F Chávez fTLJ The terrestnal b1osphere appears to take up part of the anthropogen1c carbon d1ox1de em1ss1ons about wh1ch there 1s mcreasmg concern and awareness because of 1ts poss1ble consequences Whlle much of the mcrease of atmosphenc C02 caused by agnculture 1n the last few decades 1s w1dely blamed on changas m land use m the trop1cs we report here that mtroduced pasturas based on deep rooted grasses m the South Amencan savannas can accumulate orgamc carbon at depth 1n the s01l These grasses were mtroduced and selected by the Centro Internacional de Agncultura Trop1cal (CIATI emphas1zmg deep rootedness pnnCipally for extract1on of calc1um and other nutnents and for better expl01tat1on of so1l water to mamtam green forage dunng the 3 5 months dry season The amount of carbon sequestered could be as h1gh as 507 m1lhon tons (Mtl carbon a year wh1ch could account for part of the m1ssmg global carbon smk of O 4 4 3 b1lhon tons (Gt) The d1scovery of th1s amount of carbon sequestrat1on 1n tropical savanna so1ls represents an 1mmed1ate and effect1ve alternat1ve m efforts to combat the threat of global warmmg w1th the added advantage of mcreased agncultural product10n on so1ls of low fert1hty 216 F1gure 25 A 1 8 The d1stnbutron of a pohhumrc specres of Glossoscolecrdae earthworm rn a fifteen year-old Brach1ana decumbenslkudzu grazed pastura al Canmagua rn September 1993 A adults B ¡uvenlles The contours hnk srtes wrth the same densrty (figures are rndrvrduals m 2) The surveyed area measured 70 m a srde 217 A 8 Figure 26 The d1Str1but1on earthworms 1n a s1xteen year-old Brach1ana decumbenslkudzu grazed pastura al Canmagua 1n October 1994 A d1stnbut1on of ¡uven1fes of a pohhum1c spec1es of Glossoscolec1dae B dl&trlbUIIon of fresh surface cests of MartJodnlus spp The contours hnk s1tes WJ!h !he sa.me dens1ty (figures are 1nd1V1duals or casts m 2) The surveyed ares measured 70 m a s1de 218 F1gure 27 The -d1slnbut1on of the JUVemles of a pohhum1c spec1es of Glossoscolec1dae earthworm m an ungrazed nallve savanna at Canmagua 1n November 1993 The contours hnk s1tes w1th the same dens1ty (figures are 1nd1v1duals m 2) The surveyed area measured 70 m a s1de 219 Savannas occupy sorne 250 m1llion hectares (M ha) of South Amanea mamly m Braz1l (200 M ha) Colombia (20 M ha) and Venezuela ( 12 M ha) They are u sed for extens1ve cattle ranchmg on the natlve forage although m Braz1l croppmg w1th ma1ze and soybeans (now 12 M ha) and mtroduced pasturas (35 M ha) have become 1mportant dunng the last 30 years The so1ls of the savannas are mamly ox1sols and ult1sols charactenzed by low pH s (4 O 4 8) alummum saturat1on up to 90% and low levels of P and Ca Ramfall1s 1500 3000 mm w1th a ummodal seasonal d1stnbut1on The perenmal grasses Andropogon gayanus and 8rach1ara humldlcola are of Afncan ongm A gayanus 1s tall growmg w1th a tussock hab1t wh1le 8 hum¡d¡co/a forms swards The legumes Arach1s pmto1 and Sty/osanthes cap1tata are from South Amenca A pmto1 1s a v1gorous stolomferous perenmal and S cap1tata 1s a free seedmg b1annual Grass legume assoc1at1ons of these forages produce cattle livewe1ght gams up to 500 kg ha 1 yr 1 compared w1th 7 20 kg ha 1 on well managed savanna They produce green forage for several months mto the dry season and regrow v1gorously soon after the opemng rams of the wet season Smce 1980 all four have been released as cult1vars 1n one or more countnes 1n South Amenca There are now 7 M ha planted toA gayanus 1n Braz1l alone (M A Ayarza personal commumcat1on) Although deep rootedness 1s cons1dered a ma¡or factor 1n adaptat1on to low fert11ity so1ls espec1ally 1n A gayanus the role of the roots m the dynam1cs of soll carbon has largely been 1gnored We measured soll carbon 1n an A gayanus pasture and m two 8 hum¡d¡co/a pastures at two s1tes on the eastern plams (llanos) of Colombia sorne 200 km apart (Table 26) In each case we obtamed correspondmg measurements for the nat1ve savanna The pasturas had d1ffenng h1stones (Table 26) but all had carned grass basad pasturas for three to nme years The pastura at Matazul farm was not fert11ized after the nce that was used as a p1oneer crop to establish 1t whlle at Canmagua fert11izer was applied at pastura establishment and each second yaar thereaftar Thay were all grazed by cattle at normal stockmg rates for 1mprovad pasturas 1n the reg1on All the grass based pasturas made a stnkmg contnbut1on to so1l C comparad w1th tha nat1ve savanna espec1ally when grown w1th a legume (Table 27) Data for other pasturas based on another grass of Afncan ongm 8rachlaf/a d1ctyonaura cv Llanero at Matazul farm show lower but stlll S1gmf1cant f1xat1on of about 30 t C ha 1 1n 3 ~ years (data not presentad) lf thase data are representativa of tha araas sown to pasturas of A gayanus and 8rach1af/a spec1es m South Amenca conservat1valy est1matad at 35 M ha then from 100 to as much as 507 Mt C IS bamg saquasterad each year The contnbut1on of the leguma 1n the s1x years smce 1ts establishment at Canmagua may be est1mated by tak1ng the d1fference between the grass alone and the assoc1at1on The d1fference 1s 44 7 t C ha 1 so that although legumes contnbuta only about 1 5% to root b1omass the assoc1at1on w1th A pmto1 mcreased carbon f1xat1on by 7 8 t ha 1 yr 1 comparad w1th the pure grass Comparad w1th the savanna the grass based pasturas sequester most of the C 1n the deeper part of the so1l proflle well below the plough layer (normally 1 O 15 cm) Th1s C should therefore be less prone to ox1dat1on and hence loss durmg any croppmg phase that m1ght be undertakan 1n mtegratad crop and pastura systams lndaad such systems should be able to accommodata rotat1ons Wlth annual crops and st1ll contnbuta to C saquestrat1on Jones et 220 al drew attention to the role of fire 1n determ1n1ng the b¡¡lance between the vegetat10n of native savannas as either a net s1nk or source of carbon 1n the tropics lntroduced pasturas are rarely burned except by accident 1n contrast to the native savannas which are usually burned as often as each year Table 26 Location and charactenstics of the two Sites on the savannas of the eastern pla1ns of Colombia Site Matazul farm Canmagua Research Station location Eastern pla1ns (llanos) Eastern pla1ns (llanos) 200 km Puerto lopez Colombia ENE of Puerto lopez Colombia latitud e 4°9 N 72°39 w 4°37 N 71°19 w longitud e 160 175 Altltude (m) Mean annual 2 700 2 240 ra1nfall (mm yr 11 Soil ÜXiSOI ÜXiSOI Texture Clay loam Clay loam pH (1 1 water) 44 4 1 P (0 20cm) 1 8 1 5 Bray 11 (ppm) Pasture details 1989 Cropped from 1984 Sown to 8 humldJco/a savanna With upland nce cv Humidicola from savanna undersown With mixed With the legume DesmodJUm A gayanus cv Canmagua 1 ova/¡foflum which falled and S cap1tata cv Capica 1987 Resown to 8 humJdJco/a pasture cv Humidicola alone or With A 1989 93 Rotationally grazed pmto1 cv Mam ForraJero With cattle at 2 head ha 1 1988 93 Rotationally grazed with cattle at 3 head ha 1 Date soil December 1992 Apnl 1993 sampled 221 Table 27 Y1eld and net gam of carbon 1n grass based pasturas comparad w1th nat1ve savanna on two s1tes on the savannas of the eastern plams of Colombia Methods Each sample compnsed e1ght cores at Matazul and four at Canmagua taken at random by so1l auger to the depths 1nd1cated In the grazed A gayanus/S cap1tata pasturas at Matazul Farm samples were taken from each quarter of each of three 1 ha plots 1n a random1zed complete block expenment covenng 9 ha At the same t1me four samples were taken from the nat1ve savanna 1mmed1ately adJacent to the expenment At Canmagua the grass basad pasturas wera O 5 ha plots 1n a random1zad complete block grazmg expenment F1va samplas were takan from aach of two rephcates of the plots and three from the 1mmed1ately adjacent savanna The samples were dned and m1lled to pass a 1 mm s1eve Fme roots were not removed befare milhng Sub samples were d1gested 1n sulphunc ac1d potass1um d1chromate heated (150 °C) for 30 mm on a temperatura controlled hotplate Carbon concentrat10n 1n the d1gest was datermmed colonmetncally agamst cahbrated standards made up of carbon free s01l to wh1ch measured amounts of glucosa were added S01l bulk dens1ty used to convert the grav1metnc f1gures of so1l C to volumetnc data was determmed by standard methods25 The data for each depth of each pastura were treated as mdependent samples and the standard error of each mean and the standard errors of the d1fference between the means were calculated D1fferences between the means were testad for stat1st1cal S1gmf1cance w1th Student s t test Srte MalazL! Fann Cmumgua Research Slaboo Pastue Agay¡n¡s/ Blll.mdcola B h1111Cfcda 1 Savama S-""" Savama alone A pmta Corbon In Csrbon Dl1lerence Csrbon Csrbon Dl1lerence Csrbon Dl1lerence Deplh layer mlayer fran m !ayer mlayer fran '" layer fran savama savama savama an lha lha 1 ha" :tSE lha lha 1 ha :tSE lila" lha±SE ().20 640 711 71±20- 703 760 57±43ns 881 178±42- 20-40 427 519 93±28- 524 576 53±32 ns 712 186±60- 4().100 f 798 1142 343±93- 743 892 149±82 1084 34 0±100"' Tolal 1865 2372 &l7:t114- 1971 222.8 257±7r 2675 704±15s- ' Al Cmumgua Research Slatton ll"e deeper layer was 4Cl-80 an ns P>O 05 PMien ll"e means (n=14 lar MalazL! Fam, n=12 lar Conniagua) There are few data of so1l C below 15 20 cm m the trop1cs although there are occas1onal f1gures to 40 cm Long et al and the1r collaborators camed out careful stud1es on nat1ve trop1cal grasslands to document total pnmary product1v1ty mcludmg roots 1n the surface 15 cm They showed that product1v1ty of trop1cal grasslands above ground and below ground to 15 cm was up to f1ve t1mes h1gher than prev10usly reportad mamly because losses due to 222 senescence were 1gnored On the bas1s of our data even they may have substant1ally underest1mated root product1on They d1d not present data of soll C Although CIA T s emphas1s on deep rooted grasses was for reasons other than sequestrat1on of C we have shown that soma mtroduced grasses 1n the trop1cal South Amanean savannas do sequester C deep 1n the soll There must be a phys1olog1cal cost 1n growmg such a mass1ve root system but strangely we do not see much ev1dence of 1t 8oth A gayanus and 8 hum¡d¡co/a grow at least as v1gorously as other prom1s1ng mtroduced grasses and much more v1gorously than the savanna speCies We suggest that the sequestrat1on of C 1n South Amanean savanna s01ls 1s of global 1mportance We suggest that deep rootedness can be expl01ted from the p01nt of v1ew of both the Individual farmer and the commumty at larga for the mutual benef1t of both Can the ab11ity of tropical grasses to sequester C at depth espec1ally when grown 1n mixtures w1th legumes be used 1n select1on and breedmg? What are the 1mplicat1ons for the generat1on of technolog1es to recuperate degradad pasturas 1n ele arad areas of the hum1d forests? The latter 1s an emot1onal 1ssue but there 1s considerable capac1ty for these so1ls too to sequester carbon 1f they are managed correctly The combmat10n of a deep rooted grass w1th a mtrogen f1x1ng legume can mcrease nutnent cycling greatly 1mprove ammal product1on and markedly mcrease soll b1olog1cal act1v1ty These effects occur mamly at the so1l surface whlle C storage takes place below the plough layer Thus far from bemg env1ronmentally degradmg 1mproved pasturas can fulflll the restorat1ve role 1n tropical systems that was recogmzed 1n pre Roman t1mes for Med1terranean systems and may play a v1tal part 1n stab11iz1ng the global C cycle and mm1m1Z1ng the greenhouse effect of atmosphenc carbon d1ox1de Crop nutr1t1onal requ1rements and 1nput use eff1ciency on llanos ox1sols under crop rotat1ons D K Fnesen (IFDCJ J 1 Sanz D Molma and M RIVera fTLJ The predommant s01ls of the Colomb1an Llanos are h1ghly ac1dlc Ox1sols and Ult1sols whose mmeralogy 1s dommated by kaolimte and the ox1des and hydrous ox1des of 1ron and alummum These mmerals 1mpart to so1ls propert1es wh1ch are not conduc1ve to crop product1on and the eff1c1ent use of nutnent mputs The cat1on exchange capac1ty 1s very low (typ1cally 2 5 meq/1 00 g) largely pH dependent and h1ghly saturated w1th exchangeable alummum (Al) Levels of exchangeable calc1um (Ca) magnes1um (Mg) and potass1um (K) are consequently also low so that the applicat1on of fertilizar needs to be carefully balanced to avo1d unfavorable mteract1ons wh1ch would lead to nutnent deflc1enc1es 1n the crop meff1C1ent use of mputs and exacerbated losses through leachmg The h1gh ox1de contents of Ox1sols and Ult1sols also result 1n a h1gh clay mmeral surface area wh1ch strongly sorbs phosphate (P) and g1ves nse to what 1s commonly referred to as a h1gh P f1xat1on capac1ty However there are reports of substant1al residual effects of applied P on so callad h1gh P f1x1ng so1ls To max1m1ze P fert11izer use eff1c1ency 1t 1s necessary to quant1fy the res1dual value of prev1ous P fertilizar applicat1ons and to understand the fate of applied P and 1ts rate of movement between 1mportant P pools 1n so1l mcludmg so1l orgamc P pools wh1ch play a p1votal role 1n P cycling 223 Because of the mherent 1nfert1hty of savanna so1ls and the need to rat1onahze fert1hzer mputs and max1m1ze nutnent use eff1c1ency 1n the croppmg systems bemg developed by CIA T for the Llanos expenments were estabhshed 1n 1993 to ( 11 determme opt1mal levels of soluble phosphate fert1hzer and opt1mum ratas of hme potass1um and magnes1um for selected annual crops on contrastmg solls of the eastern savannas (21 est1mate res1dual effect1veness of P hme K and Mg apphcat1ons and requ1s1te mamtenance ratas for opt1mal crop growth (3) charactenze the fate of P apphcat1ons and parametenze a modal of P res1dues 1n h1ghly weathered solls and (41 momtor the dynam1cs of apphed cat1ons and Al and pH 1n the soll and the mteract1on of amendments on nutnent fluxes and fate These expenments have now completad 1 Y:z to 2 annual rotat1on cycles Th1s report summanzes both results on 1n1t1al crop nutnt1onal reqUirements on Llanos Ox1sols and ac1d1c lncept1sols and prehmmary results on the residual effect1veness of fert1hzer mputs Matenals and Methods Dunng 1993 expenments were estabhshed at three s1tes on the eastern plams of Colombia on loamy clay Ox1sols at the Corpo1ca/CIA T Research Stat1on at Canmagua and the Matazul Farm approx1mately 70 km east of Puerto Lopez and on a sandy lncept1sol at La Flonda Farm on the Seranra approx1mately 50 km south of Matazul Due to log1St1cal problems the Seranra s1te was abandonad 1n 1994 So1l charactenst1cs at each s1te are summanzed 1n Tabla 28 The expenments are umformly two crops per year rotat1ons of ma1ze and soybeans or nce and cowpea The formar rotat1on reqUires h1gher levels of hme mputs due to the greater sens1t1v1ty of the germplasm to ac1d soll cond1t1ons than the latter rotat1on Two types of expenments were Implementad as descnbed below Tabla 28 Chem1cal charactenst1cs of so1ls at the three expenmental s1tes Org e Total Bray P pH Exchangeable Cattons (meq/1 OOg) Al St Textura N IH201 Satn i%1 C%1 Cppml Cppml Al H Ca Mg K ECE e Ganmagua l.oamy 2.41 1249 14 426 228 023 009 007 022 289 790 day Matazul Sandy 150 682 19 440 193 023 012 006 005 239 807 day La Ronda Sand 080 400 31 401 071 008 004 003 014 100 709 L1me potasslum magnes1um balance experlment Th1s expenment 1s an mcomplete faetona! of 5 or 6 levels of calcltlc hme (0 150 300 600 1200 kg/ha on nce O 400 800 1600 3200 6400 kg/ha on ma1ze) 4 of potass1um (0 40 80 120 kg K/ha) and 3 of magnes1um (0 30 60 kg Mg/hal g1v1ng a total of 32 treatments set out 1n a spht plot des1gn w1th three random1zed blocks L1me treatments (mam plotsl were apphed once at the 1mt1at1on of the expenments whereas K and Mg treatments are apphed to each crop ( Y:z ratas on the legumes) So1l proflle samples toa depth of 100 cm (0 10 10 20 20 40 40 60 60 80 80 100 cm) are taken from selected treatments annually pnor to sowmg the cereal crop surface soll samples are taken from all plots pnor to each crop All samples are analyzed for pH 224 exchangeable cat1ons (Al H Ca Mg K) Orgamc C total N and avallable P are also momtored 1n the surface so1l Phosphorus res1dual value expenment Th1s expenment compnses a total of 1 6 treatments set out 1n a random1zed block des1gn w1th 4 rephcat1ons The treatments cons1st of 1 O rates of P (as TSP broadcast and mcorporated at plantmg) apphed once only at the begmmng of the expenment and 6 rates apphed annually to the cereal crop 1n the rotat1ons Taken together these treatments fully charactenze the P response funct1on 1n the f1rst year and enable est1mat1on of the res1dual effect1veness of the 1n1t1al apphcat1on 1n subsequent years Crop observat1ons mclude response to P fert1hzer apphcat1ons and res1dual value d1agnost1c t1ssue analys1s and component analys1s to determme P removals and returns v1a crop res1dues Soll samples taken pnor to each crop wlll be used to momtor changas 1n lab1le P pools as mfluenced by P fert1hzer apphcat1ons and t1me Measurements w1ll 1nclude lab1le P by class1cal so1l test methods (Bray and Olsen avallable P) 1sotop1C exchange and determmat1on of orgamc and morgamc fract1ons followmg the Hedley procedure In all expenments nutnents not mcluded as treatments (mcludmg N S and Zn) are apphed at adequate levels All fert1hzers (treatments and basal) are broadcast and mcorporated pnor to plantmg although N and K apphcat1ons are spht whereby a second and th1rd dressmg are apphed at 30 and 60 days after plantmg Results S01ls at all three expenment s1tes are h1ghly ac1d1c havmg a pH (H20) 1n the ranga of 4 4 5 and an exchangeable Al saturat1on of 70 80% (Table 28) S01l textura ranges from sandy through to loamy clay w1th correspondmg d1fferences 1n so1l orgamc C and total N among s1tes All s1tes are umformly very low 1n ava1lable P as measured w1th the convent1onal Bray2 extractant L1me potass1um magnes1um balance/1nteractions L1me rates chosen were basad on the relat1ve reqUirements of nce (0 to 1200 kg/ha) and ma1ze (O to 6400 kg/ha) and the des1re to observe the full response curve throughout the ant1c1pated 4 year durat1on of the expenments Measured sorne 5 to 8 months after apphcat1on calc1t1c hme had very httle effect on soll pH at rates of about 1500 kg/ha or less (F1gure 28A and 280) Even the apphcat1on of 6 4 t/ha to the loamy clay Ox1sol at Canmagua d1d not mcrease the so1l pH above 5 5 an 1nd1cat1on of the h1gh ac1d buffenng capac1ty of th1s s01l Nevertheless hme had 1mmed1ate and 1mportant effects on exchangeable Al and Ca on both so1ls (Figure 28) lnterpolated from the relat1onsh1p between Al saturat1on and hme apphed (not shown) the standard 300 kg/ha apphcat1on at Matazul reduced Al saturat1on to about 70% wh1le 500 kg/ha reduced 1t to about 60% At Canmagua Al saturat10n was reduced to about 35% w1th an apphcat1on of 2 t/ha of calc1te Two sequent1al soll samphngs at Canmagua one season and two seasons after hmmg showed no not1ceable temporal changas 1n so1l pH exchangeable Al or Ca (F1gure 28) Clearly a longar t1me mterval 1s necessary to d1scern losses of Ca from the surface so11 and reduced res1dual effect1veness of the apphed hme Two nce cult1vars (the ac1d s01l tolerant Sabanas 6 hne and the susceptible Llanos 5) were sown at Matazul 1n 1994 on plots wh1ch had been hmed one year earher As 1n the precedmg year (see CIAT Savannas Program 81enmal Report 1992 1993) Sabanas 6 d1d not respond to hme (calcita) when adequate levels of K and Mg were also apphed and relat1vely httle K (about 40 kg/ha) was reqUired for max1mum y1eld (F1gure 29) As ev1dent 1n the negat1ve mteract1on Ca (as calc1te) and Mg compensated for each other 1n mcreasmg nce 225 product1on Thus m the absence of calc1te 30 kg Mg/ha resultad m max1mum nce gram y1elds wh1le 600 kg/ha of calc1te (about 150 kg Ca/hal gave max1mum y1elds Wlthout Mg apphcat1on (Figure 2981 Taken together these results suggest that the dommant constra1nt to Sabanas 6 growth m these so1ls 1s Ca def1c1ency (for wh1ch Mg can subst1tutel rather than Al tox1c1ty In contrast to Sabanas 6 nce Llanos 5 had a s1gmf1cant hme reqUirement for wh1ch Mg was unable to compensate (F1gure 2901 Moreover y1elds of the susceptible Llanos 5 were generally much poorer although lt had a h19her apparent K reqUirement than Sabanas 6 (Figure 29CI Cowpeas were sown after nce on the sandy lncept1sol at La Flonda m September 1993 Y1elds were max1mal at about 1 t/ha w1th as httle as 600 kg/ha of hme apphed to the nce (Figure 30AI Cowpea also responded markedly to K apphcat1ons of up to 40 kg/ha wh1le h1gher rates tended to reduce y1elds espec1ally at lower hme rates S1m1lar observat1ons have been made w1th cowpea m N1gena and are attnbuted to Al tox1c1ty 1nduced when mcreasmg concentrat1ons of K 1ons exchange Wlth and displace Al mto soll solut1on As w1th the precedmg nce crop a very s1gmf1cant def1c1ency m Mg was also observad at La Flonda IF1gure 3081 The recently released ac1d so1l tolerant ma1ze hne (CIMMYT S1kuam 1 was plantad at Canmagua dunng the second season of 1993 lt produced a max1mum y1eld of about 3 5 t/ha w1th about 1 5 t/ha of caiCitiC hme (F1gure 31 Al Th1s amount of hme reduced the exchangeable Al saturat1on to about 45% More s1gmf1cant however 1s the observat1on that th1s matenal produced 2 6 t/ha or 77% of the max1mum y1eld even w1thout hme Response to K was observad up to 80 kg/ha IF1gure 31 81 but m contrast to observat1ons at the other two s1tes there was no s1gmf1cant response to Mg at Canmagua (data not shownl Phosphorus fertthzer response and restdual effects Response of nce and cowpea at Matazul and ma1ze at Canmagua to P fert1hzer reflect the very low levels of ava1lable P m these Ox1sols m general v1rtually no y1eld was obtamed w1thout apphed P (Figure 32 and 331 Nevertheless apphcat1on of P fert1hzer produced sharp responses w1th max1mum y1elds bemg obta1ned w1th 40 60 kg P/ha on nce and 60 80 kg P/ha on ma1ze Cowpea grown on residual P apphed to the prev1ous nce crop m 1993 also produced max1mum y1elds at 40 60 kg P/ha (Figure 33AI Results from the f~rst semester of 1994 at Canmagua and Matazul prov1de d~rect compansons between res1dual P apphed 8 to 11 months earher to ma1ze or nce and freshly apphed P Although companson of md1v1dual treatments wh1ch rece1ved all P m the prev1ous year w1th those wh1ch rece1ved the same amount spht between two years suggests a d1m1mshmg effect1veness of the older P response to both the res1dual and annual P apphcat1on strateg1es are descnbed by the same response funct1ons These results md1cate f1rstly that P fert1hzer reqUirements of crops on Llanos Ox1sols are not excess1ve and secondly that P apphcat1ons have potent1ally substant1al res1dual value Conclus1ons Llanos smls are def1c1ent m most of the majar nutnents necessary for crop product1on and P 1s clearly among the most hm1tmg nutnents Fortunately germplasm now ex1sts wh1Ch 1s reasonably tolerant of the other majar agncultural constramt of these so1ls sml ac1d1ty and Al tox1c1ty Relat1vely modest rates of hme were found to be suff1C1ent to produce max1mum y1elds of nce cowpeas and ma1ze usmg the advanced germplasm Very prehmmary data also 226 ( suggest relat1vely small reduct1ons of the effects of hme after a 1 year penod Res1dual effect1veness of P fert1hzer on these h1gh P f1x1ng solls also appears to be apprec1able However for both hme and P 1t w1ll be necessary to momtor these effects over a penod of several years to quant1fy the mput use eff1c1ency and to determme the1r tate For phosphorus mformat1on of th1s type wlll enable the parametenzat1on of models descnbmg fert1hzer P res1dual value and cychng mcludmg the CENTURY model wh1ch s1mulates P cychng 1n agroecosystems lt w1ll also ass1st 1n the development of the P sub model of the CERES crop S1mulat10n models and enable the extrapolat1on of results to other env1ronments 227 60 60 A O Uname-.. aall (Jul831 D 0 Unamondodaaii(Aprl83( • B monillo aller llmlng (Apr IMI - 55 e 1 monthoatlerllmlng (Bep 831 55 o N ::r:: - ::r:: 50 50 Q. - -o 45 45 m 40 40 o 300 600 1200 o 1600 3200 6400 30 30 - B C) o o o o o ... e- 20 20 cu e - -e( 10 10 .e u >< w 00 00 o 300 600 1200 o 1600 3200 6400 20 60 - e C) o o 15 ... - 40 C" cu e 1 o -ca o 20 .e 05 u >< w 00 00 o 300 600 1200 o 1600 3200 6400 Rate of lime (kg/ha) Rate of hme (kg/ha) Fcgure 28 lnfluence of calcllic hme on so1l pH exchangeable Al and Ca 1n Ox1sols at Matazul (A B C) and Canmagua (O E F) and changas 1n t1me after apphcat1on 228 A e 3000 3000 - __ ., ftll .e - C) ..111: - 2000 2000 :!:! Cl) o > o o e i! 1000 o o O kg-K/ha 1000 (!) • 40 kg-Kiha ., 80 kg Klha • 120 k -Kiha o o o 300 600 1200 o 300 600 1200 B 3000 A 3000 - ~~---.,. D ftll .e - C) ..111: - 2000 2000 .,.,.,.... "C Q; / > o _..,.,.,... -e e! 1000 1000 o O kg-Mglha (!) o • 30 ., 60 .o o o o 300 600 1200 o 300 600 1200 Rate of Ltme (kg/ha) Rate of Lime (kg/ha) F1gure 29 Response of two nce cult1vars (Sabanas-6 [A B] and Llanos 5 [C D]) to and 1nteract1ons of lime (calc1te) potass1um and magnes1um on an Ox1sol at Matazul farm (1994) 229 -111 .e -DI .111: -'t:l Gi > 5 1! (!) 1200 A B 900 ~----A / "" 1' 600 1 1 1 • K60 + Mg o 300 ... K40 + Mg ... K40 o Kso Mg • Kso !::,. K40 Mg o o 300 600 1200 o 300 600 1200 Rate of Lime (kg/ha) Rate of L1me (kg/ha) Figure 30 lnteract1ons of (A) hme and K at adequate Mg and (Et hme K and Mg on cowpea y1elds on a sandy Ox1sol at La Flonda (1993) 230 4000 4000 A B 3500 3500 -ftS .e -C) .X 3000 - :!':! CP > 2500 2500 e l! (!) 2000 2000 0i' {( 0{, ,{ o 1600 3200 6400 o 40 80 120 Rate of L1me (kg/ha) Rate of K Apphed (kg/ha} F1gure 31 Response of ma1ze (CIMMYT S1kuan1 ) to (A) calc1bc hme at opt1mallevels of potass1um and magnes1um and (B) to potass1um at opt1mallevels of hme and Mg 231 -111 .e ~ - 'ti Qi > e i! (!) 4000 A Sem 2 1993 B Sem 1 1994 • o • • • • 3000 • • • 2000 • Residual P (applled Aug 83) 1000 o Annual P (appllad 1883 & Apr 84) Y= 2948 (1 0 938 eO 0273"P) y= 3170 (1 0 921 e .O 03BO"P) 2 r =O 921 2 r =O 549 o o 20 40 80 80 120 180 200 o 20 40 60 80 120 160 200 P Applled (kg/ha) Cumul P Apphed (kg/ha) F1gure 32 Response of ma1ze (CIMMYT S1kuam) to [A] freshly apphed TSP and [8] residual and re-apphed TSP on a loamy clay Ox1sol at Canmagua 232 -1'11 i - "C a¡ ~ S:: i!! (!) 1600 4000 A Cowpea B R1ce (Lmea-3) (Sem 2 1993) (Sem 1 1994) • 1200 3000 • • 800 2000 Residual P (applled May 831 400 1000 Annual P (applled 1883 & Apr 841 Y= 993(1 O 871 e .o OMI"P) Y= 3255 (1 O 711 e .o 02&2"P) 2 r =O 978 2 r =0897 o o o 20 40 60 80 120 160 200 o 20 40 60 80 100120 160 200 P Applied (kglha) Cumul P Apphed (kglha) F1gure 33 Response of [A] cowpea to res1dual TSP and (8] nce to res1dual and re apphed TSP on a loamy-day OXIsol at Matazul Fann 233 TI02 INTER PROGRAM PROJECT WITH FORAGE PROGRAM NITRO !,lEN FIXA TI ON ANO NITROGEN TRANSFER IN FORAGE LEGUMES R1chard J Thomas (TL) 1 N1trogen f1xat1on by forage legumes In grass/legume pasturas Rates of mtrogen f1xat10n by forage legumes are normally measured dunng the f1rst or second year of pastura establishment only and there have been few reports of f1xat1on ratas over longer t1me penods as 1t 1s assumed that the legume contmues to f1x mtrogen at s1m1lar rates to those measured durmg establishment Th1s assumpt1on needs exammmg however as so1l fert1llty m ay be decllmng after the 1n1t1al fert1llzer appllcat1on g1ven at establishment and there 1s ev1dence that decreased s01l fert1llty reduces the % of the legume s N denved from f1xat1on (Trop1cal Pasturas 1987 1991 Report 1991) N1trogen f1xat1on 1n the forage legumes Arach1s pmto1 Centrosema acutlfolwm and Stylosanthes cap1tata grown m assoc1at1on w1th the grass Brach1ana d1ctyoneura was subsequently measured for three success1ve years m the satelllte s1tes to the core expenment m Canmagua (sea B1annual Report Savanna Program 1992 93 for deta1ls) The amounts of mtrogen f1xat1on par umt legume b1omass and % of legume N denved from f1xat1on (%Ndfa) are shown 1n Table 29 for sandy loam and clay loam soll types and the two levels of 1n1t1al fert1llzat1on used 1n the core expenment Tabla 30 shows the data 1n terms of absoluta amounts f1xed and % legume m the pastura lmt1al fert1llzat10n had llttle or no effect on %Ndfa at e1ther s1te ovar the three years S1m1larly the %Ndfa was mamtamed over t1me even though so1l fert1llty was supposedly decllmng The %Ndfa m all three legumes remamed at h1gh levels ( > 80% m general) Conclus1ons As reportad prev1ously (B1annual Report Savanna Program 1992 93) the mam effect of fert1llty was on legume populat1on (Tabla 30) and not on f1xat10n per umt b10mass and therefore an est1mate of the amounts of mtrogen f1xed can be made from a measurement of legume b1omass m the pastura Thus such est1mates can be done very s1mply m farmer s f1elds Non f1x1ng controls In the above expenment the grasses 8 dJctyoneura Pamcum max1mum ecotype KK 16 and the nat1ve savanna were used as non f1xmg controls m the 15N 1sotope d!lut1on techmque To check 1f these were appropnate non f1xmg controls a glasshouse expenment was done w1th soll from lntro 11 (clay loam) to determme the N uptake patterns and N d1stnbut1on m two grasses non f1xmg mutants of Stylosanthes cap1tata and the normal N2 f1xmg S cap1tata cv Cap1ca The results (Table 31) 1nd1cate that m general the amounts of N taken up and 1ts d1stnbut1on were s1m1lar for the non f1xmg plants but they all d1ffered from N2 f1x1ng S cap1tata Conclus1ons The results 1mply that caut1on should be exerc1sed m the est1mates of mtrogen f1xed as there 1s no guarantee that f1xmg and non f1xmg plants are takmg N up from the same so1l N pools over t1me and th1s may mvalldate the 1sotope dllut1on techmque However the amount of legume N denved from f1xat1on d1d not d1ffer greatly usmg any of the non f1xmg controls m th1s expenment nor m the f1eld expenments reportad prev1ously and 1n the glasshouse expenment a mean value of 83% ± 2 8% was obtamed In add1t1on 1f the %NDFA 1s as h1gh as 60% or more then d1screpanc1es m patterns of N uptake w1ll not affect est1mates of N2 f1xat1on greatly The ' 6N ennchment of the non f1x1ng S cap1tata was s1m1lar 234 to the grasses (Table 31 1 and 1nd1cates that 1t can be used as a vahd non f1x1ng control prov1ded N fert1hzer 1s apphed to obtam s1mllar growth rates and N accumulat1on Tabla 29 Amounts of N2 flxad kg N/t laguma DM and Ndfa A pmtot e scuttfoflum S CBPI!BtB S1te Fert1hty 1991 1992 1993 1991 1992 1993 1991 1992 1993 Sandy Low 18 7 25 3 23 6 27 1 17 9 75 (81 5) (88 91 (64 5) (85 61 (87 51 (74 2) Loa m H1gh 24 4 33 1 204 92 31 o 17 4 93 (87 1) (91 7) (96 31 (94 5) (90 2) (88 01 (89 3) Clay Low 19 7 15 3 29 1 20 3 36 o 21 6 16 6 26 6 (71 7) (81 4) (91 4) (92 1) (92 9) (79 7) (87 91 (90 71 Loa m H1gh 22 8 179 23 3 29 1 25 5 (38 O) 22 4 16 o 28 o (85 6) (68 31 (81 7) (92 9) (95 4) (95 9) (89 1) (92 2) (89 O) Number 1n brackets are % Ndfa Tabla 30 Amounts of N2 f1xed ovar 12 weaks kg N/ha and % lagume content '" pastura A ptntOI e scutdollum S cap1tsta S1te Fert1hty 1991 1992 1993 1991 1992 1993 1991 1992 1993 Sandy Low 08 1 7 03 21 o 44 03 (4 0) (5 01 (1 4) (46 3) (6 31 (1 4) Loa m H1gh 74 25 08 02 40 o 22 04 (17 7) (4 51 (2 3) (0 8) (44 2) (16 6) (1 01 Clay Low 09 3 6 24 o 50 14 8 47 08 (1 9) (4 61 (6 61 (1 4) (27 0) (24 O) (2 4) 68 o 70 52 46 (2 6) 31 o 60 26 Loa m H1gh (10 0) (2 4) (54) (9 4) (6 2) (32 31 (21 6) (7 2) Number 1n brackets are % legume contents 10 the pasturas 235 Table 31 Companson of non flxmg controla 1n the ostomato of N2 foxatoon by S CBpttBtB % boomass %N m Specoos DM (g/plantl abo ve Total N abovo % atom ground parta mg/plant ground parts 16N excess S cspttata foxong 1 75 ab 70 5 a 41 4 a 75 7 a O 048 a S capttata non O 75 e 69 6 a 70b 57 4 b o 312 b foxong p 8 1 32 b 65 2 a 69b 54 2 b o 266 b maxtmum 1 93 a 53 3 b 86b 56 8 b o 275 b dtctyoneura Glasshouse grown planta harvested 30 daya after plantong Means of 5 ropototoons Numbor followed by somolar letters on each column do not doffor sognofocantly (P < O 05) 236 CHAPTER VI INTERPROGRAM PROJECT TI 01 PROTOTYPE SYSTEMS FOR ECOLOGICALLY SOUND INTENSIFICATION OF PRODUCTION IN THE HILLSIDES Preparad by José lgnac1o Sanz Collaborat1ng CIA T Sc1ent1sts (In alphabet1cal arder) E AmézqUita (TL) J Ashby (HSJ R 8est (eassava) D K Fnesen (IFDe/TLJ S Fupsaka (LM SRGJ A G1Jsman (TLJ e lgles1as (eassava) 8 Knapp (HSJ J Kornegay (8ean) e Lascano (TFJ 8 Maass (TFJ K Muller reassava HSJ J 1 Sanz (Pro¡ect Leader) J Sm1th (TLJ R Thomas (TLJ e A OUiroz (HSJ A Meléndez (HSJ A lntroduct1on Just1f1cat1on The proJect a1ms to contnbute to the Program s goal about the development of knowledge on how to combme conservat1on and product1on technolog1es Wlth the a1m of regeneratmg the natural resource base wh1le bemg econom1cally VIable to farmers The purpose of the project 1s to develop sustamable agrosllvopastoral systems that 1mprove so1l quahty water management and eff1c1ency and product1v1ty of labor After consultmg w1th researchers and farmers mvolved w1th h1lls1de act1v1t1es as well as a bread local and mternat1onal hterature search 1t 1s clear that most of the ex1stmg h111s1des research cons1sts of s1te or s1ngle problem spec1f1c expenmental work (eg Magohs et al 1991 Garc1a and Marcano 1990 Raros 1985) or soc1oeconom1c stud1es (eg Browman 1 984) normally not followed on to salve the encountered needs and even m1sconce1ved strateg1es as reportad also by Browman 1987 There IS a general def1c1ency on attempts to denve pnnc1ples w1th strateg1c vahd1ty e1ther at the s1te level as well as at a more coordmated mternat1onal macrolevel eg Andean h1IIS1des w1thm a range of predefmed charactenst1cs lt 1s 1mportant for CIA T s H1IIS1des Program to f1ll the ex1stmg gap at an mternat1onal level We need to bnng m and take out the generated data and expenence takmg place m d1fferent parts of the Andean reg1on Th1s would make poss1ble extrapolat1on across s1m1lar env1ronments a1mmg for a large scale 1mpact More spec1f1cally research on 1ntegrated crop hvestock systems mcludmg stud1es on nutnent cychng and nutnent use eff1c1ency IS weak m the reg10n Also research mvolvmg trees 1s almost all concentrated on agroforestry but not on agropastoral or agros1lvopastoral systems The project a1ms to mcrease the use of perenmal plants (grass and legume forages as well as trees) m croppmg systems and 1n varymg landscapes In domg so 1t attempts to develop pnnc1ples on mult1farm systems that can 1mprove the eff1c1ency of the use of the land both m t1me and space whlle mcreasmg the ablllty to preserve the env1ronment The project m the f1rst few years 1s startmg m the watershed of the Rfo Ovejas m Cauca Colombia Representat1v1ty of the expenmental s1tes IS bemg chosen w1th the help of nat1onal 1nst1tut10ns local orgamzat10ns as well as GIS and secondary mformat10n on so1ls chmate vegetat1on actual land use etc ava1lable for the area Extrapolab1hty to other h1IIS1de areas 1n Latm Amenca of the denved pnnc1ples from th1s pro1ect 1n 1ts f1rst few years 1s expected also w1th the help of Similar 1nStltut1ons GIS as well as modelhng tools The Rlo Ovejas watershed covers 106 000 hectares rangmg between 1100 and 3000 m a s 1 w1th 67% of farmers w1th less than 3 ha of land The populat1on IS 85 000 w1th 41% ilhterate The natural forest vegetat1on cover 1s on average 4% for the watershed W1th the few f1gures quoted above 1t can be seen that the Rlo Ovejas watershed represents the typ1cal h1lls1de s1tuat10n for Latm Amenca today h1gh dens1ty of a populat1on w1th hm1ted formal educat1on rap1d deplet1on of forests so1l losses assoc1ated w1th monocroppmg very few technolog1cal and pohcy alternativas The consequences of th1s eros1on extend beyond the watershed affectmg downstream water quahty and produc1ng sed1mentat1on of nvers and dams At the sama t1me low mcomes 1mply httle access to mputs leadmg to deplet1on of the so1l for food product1on Thus typ1fymg the cycle of poverty degradat1on further poverty Many orgamzat1ons both government and non government have been try1ng to 1mplement sustamable land use pract1ces 1n our study area and 1n general m the Lat1n Amencan H1lls1des Unfortunately successful w1despread adopt1on of stud1ed pract1ces 1s not common due to several reasons such as lack of v1gorous research on the ecolog1cal processes that the mduced pract1ces a1m to susta1n a very fragmentad process of tnal and error focused on product1v1ty and 1gnonng long term development and conservat1on needs and weak and unsustamed research capab1hty B Problem solvmg strategy and expected outputs The group of collaboratmg CIAT sc1ent1sts m the project through a process of Planmng by Object1ves 1dent1f1ed a set of major problems to be dealt w1th by our research agenda These problems are 1 Detenorat10n of so1l quahty so1l eros1on fert1hty decrease decreased nutnent regenerat1on (shorter fallows} 2 Unreahzed product1on potent1al of ex1stmg farmmg systems lack of complementanly of farm act1v1t1es mcreased pest1c1de use poor hnkage between farmmg systems and markets 3 Poor water management hm1ted 1rngat10n for small farmers water pollut1on hm1tat1on of water management due to topography 4 Low labor product1v1ty 5 Poor mformat1on 6 Def1c1ent mfrastructure Based on the above problems and 1n arder to comply w1th the stated purpose of the project the expected outputs are 1 Sustamable agros1lvopastoral systems 2 Stable or 1mproved so1l quahty 238 3 lmproved water management 4 lmproved labor product1v1ty and eff1c1ency In th1s sect1on only the f1rst three outputs w1ll be dealt w1th and the soe~oeconom1c aspects spec1f1cally market opportumt1es and s01l eros1on are dealt Wlth separately m the lnterprogram Pro¡ect m Hllls1des Subpro¡ect market opportumt1es hnked to eros1on control pract1ces a key to adopt1on' Recent surveys carned out by the Program show that 1mportant monocrops 1n the area 1n order of decreasmg area are cassava ( 16 18 months cycle labor mtens1ve) coffee sugar cane beans (becommg more 1mportantl ma1ze tomato s1sal plantam frUits lntercroppmg 1s common m the area eg cassava beans and cassava ma1ze beans m the low altltude cassava ma1ze and plantam coffee s1sal m m1d altltude The strategy cons1sts of covenng the range from short cycle shallow rooted monocrops to deep rooted perenmal and more d1verse agros1lvopastoral systems lt Wlllmclude ex1stmg as well as proposed systems 1dent1f1ed e1ther by farmers nat1onal mst1tut1ons and/or CIA T s SCientiStS The s1te select10n for the lnltial phase was at 1500 m 1n order to f1t m the ranga of 1400 m a s 1 to about 1800 m a s 1 to ensure adaptat1on of CIA T s germplasm The s1te select1on m terms of representat1v1ty w1thm the watershed and the extrapolab1hty outs1de the watershed of the encountered prmc1ples of th1s research was d1scussed m p 1 abo ve In the 1mt1al phase the expenments are s1mple and as large as poss1ble mcludmg areas for cattle grazmg m sorne mstances They are planned for long term evaluat1on and m th1s way (larga and s1mple) they w1ll allow for future add1t1ons and mod1f1cat10ns along as progress IS made m the evaluat1ons and contnbut1ons from the collaborators (nat1onal mst1tut1ons farmers and CIAT sc1ent1stsl come mto the process C Selected systems and hypotheses The crops and pasturas are v1ewed w1thm the systems as plant types beans short term shallow rooted ma1ze med1um term med1um rooted cassava long term deep rooted pasturas perenmal deep rooted herbaceous traes perenmal deep rooted woody In the selected farm s1tes the followmg systems are bemg evaluated 1 Monocrops cassava ma1ze beans 2 Ma1ze + legumes cocktail (green manure or forage) m two contrastmg fallows (cropped land and 2 years old fallow) 3 Ma1ze + Brach1arta d1ctyoneura + legumes cocktail (for grazmg) (to compare w1th 1 above and 4 and 8 below) 239 4 Naturalizad grass (Melm1s mmutlflora) for grazmg (to compare w1th 3 above and 8 below) 5 Cassava + legumes cocktail (green manure or forage) 1n two contrastmg fallows (as 1n 2 above) 6 Tnple crop assoc1at1on beans + cassava + ma1ze 7 Barners and f1eld(s) penmeter(s) Wlth grasses grasses + legumes and grasses + legumes + trees 8 Cut and carry spec1es on steep slopes grasses and legumes trees (to compare w1th 3 and 4 above) 9 Uncropped plots subd1v1ded 1n bare and mulch cover for soll phys1cs stud1es Systems 2 and 5 above are 1n two s1tes w1th varymg cond1t1ons and all systems w1th three replicat10ns except systems 3 and 4 wh1ch are respect1vely 1 and O 7 ha for grazmg System 1 1s trad1t10nal monocroppmg Wlthout conservat1on pract1ces and System 4 1s for trad1t10nal grazmg of naturalizad pasturas Systems 2 and 5 are the same as System 1 but they already 1nclude a cover crop (legumes) for soll protect1on/fert11ity enhancement System 6 mcludes the three crops from System 1 but all together 1n the f1eld a1m1ng for contmUity of product1on/mcome System 3 has only one season w1th a crop and 1s contmued Wlthout further t1llage under 1mproved grass/legumes cover for grazmg System 8 on steep slopes a1ms to have perenmal speCies for cut and carry w1thout further tlllage after the 1mt1al one System 7 a1ms so1l protect1on wh1le bemg product1ve In these systems there 1s a grad1ent from shallow rooted short cycle systems to perenmal both herbaceous and woody deep rooted systems System 9 uncropped plots w1th bare and mulch covered plots represent the two extremes of the grad1ent The general hypotheses for the pro1ect are through perenmal deep rooted systems 1t 1S poss1ble to 1 lmprove quant1ty and quality of so1l orgamc matter 2 Hold so1l 1n place agamst eros10n 3 lmprove m1crob1al act1v1ty 4 Stab11ize/lncrease s01l fert11ity 240 5 Have less vanab1hty more secunty of mcome cash flow The germplasm has so far been selected as follows Crops the most common 1n the area and 1n CIAT and CIMMYT commod1t1es cassava (Algodona) beans (Caucayá) ma1ze (S1kuam 1) Grasses and legumes the outstandmg matenals 1n both CIAT s tnals and farmers evaluat1on Grasses Brachtana dtctyoneura Penmsetum purpureum (elefante enano) Axonopus scopanus (lmpenal 60) and Melm1s mmut1flora Legumes Centrosema acutlfoltum Arachts pmto1 Stylosanthes cap1tata Chamaecnsta rotund1folta and Desmod1um macrocarpum Bamers German proJect results and farmers op1mon Elefante enano lmpenal 60 sugar cane pmeapple (local vanety) and several speCJes of legume trees Trees CIAT s farmers and CVC s recommendat1ons unt1l now There 1s need to keep lookmg for adaptad useful spec1es and systems 1n th1s area of research Ca/ftandra ca/othyrsus CIPAV 2 Ca¡anus ca¡an CIAT 18700 Codaryocaftx gyrotdes CIAT 3001 and 23748 Cratylta argentea CIAT 18516 Sesbama sesban CIAT 21250 lndtgofera sp Leucaena leucocephala CIA T 17 4 7 4 Desmodtum velutmum CIA T 1 3421 8 Erythrma edults Dendrolobtum sp Desmodtum ca¡amfoltum CIAT 3124 Tadehag1 tnquetrum CIAT 13276 Leucaena dtversJfolta D Satelhte expenments These are for testmg a number of hnes of the selected crops 1n order to evaluate the1r behav1or 1n companson to the vanet1es used 1n the systems tnals 1) Beans 11 hnes (3 reps ) 2) Cassava 5 lmes (4 reps ) 3) R1ce 2 lmes ( 1 rep ) 4) Legume trees several spec1es (3 reps ) E The way ahead lt 1s expected that the tnals themselves wlll serve the purpose of demonstrat1on tnals 1n farmers f1elds where our collaboratmg 1nst1tUt1ons and farmers have an open access for trammg act1v1t1es In the futura 1f resources allow the work should scale up to other areas w1th1n Colombia and other countnes 1n the reg1on w1th s1m1lar h1lls1de ecosystems and start networkmg and shanng expenences as 1t 1s already happemng w1th our work 1n the savannas The purpose would be to standard1ze methodolog1es 1dent1fy common treatments to fac1htate compansons across s1tes to benef1t from the expenences ava1lable 1n the reg1on and to exchange mformat1on w1th the vanous part1c1pants F References Apart from the Input by the vanous collaborators sorne ex1stmg documents were quoted or new ones were preparad for the elaborat10n of th1s Pro)ect 1 Browman D L 1984 Pastorahsm and development m h1gh Andean and lands Journal of And Env1ronments 7 (4) 313 328 241 2 Browman D L (Ed ) 1987 Pastorahsm m h1ghland Peru and Bohv1a 1n And land use strateg1es and nsk management m the Andes A reg1onal anthropolog1cal perspectiva Westv1ew Press USA pp 121 149 3 CIAT (1994) D1agnóst1co tecnológico prehmmar en el Departamento del Cauca Colombia 4 CIAT ( 1993) Forages for so1l 1mprovement and a mm al feed supply A proposal for Bundesmm1stenum fur W1rtschafthche Zusammenarbe1t (BMZ) 5 CIAT (1994) Fundmg request for 1995 6 CIAT ( 1993) lmprovmg agncultural sustamab1hty and hvehhoods m the Trop1cal Amanean h1lls1des A proposal for lnternat1onal Development Research Centre (IDRC) Sw1ss Development Cooperat1on (SDC) 7 CIA T ( 1994) Planmng by obJeCtlves Senes of meetmgs of CIAT Sc1ent1sts mvolved m the Project 8 CIA T ( 1994) Resultados da la evaluación en grupo con agncultores en un v1vero de legummosas para propósitos de forraJe o abono verde Vereda Melcho P1endamó Cauca Colombia 9 CIPASLA (Consorcio lntennst1tuc1onal para Agncultura Sostemble en Laderas) ( 1994) Reumón de d1scus1ón de las propuestas técmcas para el área de la m1crocuenca Cabuyal Cauca Colombia 1 O Garc1a R and M arcano J 1990 Evaluación de SIStemas de cultiVOS en las pendientes erosionadas del Valle Med1o Yaracuy Jn FONAIAPd1vulga Fondo Nac1onal de lnvest1gac1ones Agropecuanas FONAIAP Venezuela 11 Magohs E Gahndo 1 C de L and Mello Netto A V de 1991 Comportamento de Sistemas de cult1vo da mand1oca em relacao a producao e as pardas por erosao Rev1sta Bras1le1ra de C1enc1a do Solo 15(3) 357 362 12 Raros R S 1985 Evaluat1on and development of h1lls1de farmmg techmques for root crop product10n Termmal Report Baybay Leyte V1sayas State College of Agnculture Department of Forestry 29p 242 r 1/ :< L "' "" "' //Uifllllli 1 JI !"A. , 0.0!!!- :::-- - _......, 11 / / n l./111111111 ' "-../e .. .. • E PL A lfO ........ '" - ~-• 00 FINCA SAN ISIDRO MUN CIPO DE CALOONO E EDA PESCADOR ( KM) PROPIETARIO PARROQUIA DE PESCADOR AREA TOTAL EN ALQUILER 4 HAS 447970M ESCALA 000 lnterprogram PrOJ&Ct 1n Hllls1des TI O 1 subproject Market opportumt1es hnked to eros1on control pract1ces a key to adopt1on? Joyotee Sm1th (TLJ lntroduct1on Just1f1cat1on 1n the reg1onal context A number of authors have linked the soc1al and resource degradat10n problems of trop1cal Amenca to 1m balances between rural and urban development poliCies .nequality 1n access to resources and the promot1on of cap1tal .ntens1ve modes of product1on both 1n agnculture and mdustry (Gutman 1988 Leonard 19871 Expans1on of agnculture 1n the frag1le ecosystems of the m1dalt1tude h1IIS1des was largely a result of the appropnat1on of the best lands by the conqUistadors dunng the colomal penod and the1r mcorporat1on .nto large farms devoted to cattle ranch.ng or the cult1vat1on of commerc1al crops (Carter 19911 As the 1nd1genous populat10n .ncreased restnct1ons 1n he access to land led to populat1on movements to the h1lls1des wh1ch are where today most of the poor farmers res1de Ma.ntenance of the natural resource base '"the h111s1des 1s thus of v1tal1mportance not only to ensure the future livelihood of resource poor farmers but also to prevent the1r futura m1grat1on to urban centers thus augmentmg the soc1al problems wh1ch are endem1c among resource poor m1grants 1n urban are as Just1f1cat1on 1n the context of the program The goal of the Hllls1des Program 1s to develop knowledge about how to comb1ne conservat1on and product10n technolog1es 1n ways that susta.n and regenerate the natural resource base and are also econom1cally v1able and acceptable to farmers The program has 1dent1f1ed s01l eros10n as one of the most pressmg resource management problems The literatura on eros10n control technolog1es 1n tropical countnes 1s however charactenzed by one dom.nant theme d1sappo.nt.ng levels of adopt10n (la.ng and Ashby 1992 Ka1mow1tz 19921 The literature also shows that g1ven the farmer s c1rcumstances so1l deplet1on '" many cases 1s rat10nal from the farmers po.nt of v1ew (Ashby 1985 Anderson and Thampap1lla1 19901 As so1l degenerates over t1me y1eld and .ncome losses bUIId up At early stages of s01l deplet10n the net returns w1thout soll conservat1on exceed the net returns w1th conservat1on Over t1me as so1l degenerates further the gap declines unt1l eventually net returns w1th conservat1on are h1gher than those w1thout Adopt1on 1s unlikely to occur unt1l th1s po.nt wh1ch one study calculates to be at least 40 to 60 years after degenerat1on beg.ns depend1ng on the d1scount rate used (SeltZ and others 19791 Thus there 1s a conflict between the farmers econom1c log1c and ecolog1cal cons1derat1ons (Gutman 19881 The literatura also po.nts out that wh1le farmers cons1der the monetary benef1ts of eros1on control such as y1eld .ncreases they are unlikely to cons1der non monetary benef1ts such as soll res11ience or downstream benef1ts wh1ch accrue to others Thus the extent to wh1ch s01l conservat10n pract1ces are voluntanly adopted by farmers w1ll be subopt1mal from soc1ety s p01nt of v1ew (lzac 19941 The problems w1th adopt1on ra1sed above 1mply that farmers wlll have to be offered .ncent1ves to .nduce the t1mely adopt10n of soll conservat10n pract1ces Incentives have commonly taken the form of subs1d1es or regulat1ons The former however IS costly and 1n many cases .nduces d1stort1ons 1n other sectors of the economy The latter 1s extremely 244 d1ff1cult to 1mplement The research reportad here explores a d1fferent type of mcent1ve The ob¡ect1ve 1s to 1dent1fy mcome earmng opportumt1es hnked to so1l conservat1on pract1ces Adopt1on 1s expected to occur because of the opportumty to mcrease mcome Wlth so1l conservat1on occurnng as a byproduct Th1s approach denves support from the fact that m the few cases of successful adopt10n that have occurred so1l conservat1on pract1ces perm1tted the mtroduct1on of h1gh value crops or supported the mtroduct1on of hvestock or generated mcome by bemg assoc1ated w1th value added processes (T1ffen 1993 N1mlos and Savage 19911 Lmkmg the market opportumty to conservat1on pract1ces 1s however v1tal as the hterature IS replete w1th cases where the mtroduct1on of mcome generatmg opportumt1es w1thout any hnks to conservat1on have exacerbated resource degradat1on (Thrupp 1993) The research reportad here 1s part of a larger pro¡ect wh1ch seeks to develop prototype systems for ecolog1cally sound mtens1f1cat1on of product1on m the H1lls1des From the ecolog1cal pomt of v1ew the strategy 1s to ach1eve a trans1t1on from short cycle shallow rooted monocrops to deep rooted perenmal more d1verse agros1lvopastoral systems The pro¡ect mcludes soc1ologlcal stud1es of farmer typolog1es (Helle) bas1c stud1es on the phys1cal aspects of soll eros1on (Jesus Castillo) stud1es of the relat1onsh1p between so1l eros1on and product1v1ty of ma¡or crops m the area (Felicitas) the development of technolog1es for usmg barners for eros1on control (Karl Muller) a consort1um of vanous types of 1nstltUt1ons mcludmg NGOs who wlll be pnnc1pally respons1ble for 1mplementat10n of technolog1cal and mst1tUt1onal fmdmgs The research reportad here complements th1s work by a) lnvest1gatmg the market potent1al of matenals bemg used m the barners currently under development b) ldent1fymg alternat1ve market opportumt1es whose potent1al for eros1on control and compat1b1hty Wlth the strategy of the pro¡ect w1ll be mvest1gated e) ldent1fymg mst1tut10nal arrangements requ1red to enable farmers to explo1t market opportumt1es d) lncorporatmg results of farmer typolog1es m the analys1s of the adoptab1hty of alternat1ve eros1on control technolog1es Problem Solv1ng Strategy The bas1c hypothes1s 1s that technology and mst1tut1onal hnkages are key factors m prov1dmg resource poor farmers w1th market opportumt1es hnked to conservat1on pract1ces Sorne of the opt1ons currently be1ng cons1dered by b1ophys1cal sc1ent1sts mclude perenmal and annual forages for the1r potent1al as feed for da1ry cattle cereals such as m1llet to be used as poultry feed for egg product1on perenmal barners such as c1tronella w1th potent1al for mdustnal use and vetiver grass w1th very mm1mal econom1c value The current pro¡ect wlll help b1ophys1cal sc1ent1sts narrow down opt1ons for mclus1on m prototype systems At the sama t1me 1t w1ll prov1de 1deas for new opt1ons by 1dent1fymg new market opportumt1es whose eros1on control potent1al w111 be mvest1gated by b1ophys1cal sc1ent1sts The1r results w1ll m turn feed mto the ex ante econom1c evaluat1on of potent1al technolog1es The research compnses the followmg steps a) ldentlf1cat1on of market opportumt1es b) Evaluat1on of eros1on control potent1al of 1dent1f1ed market opportumt1es 245 e) ldent1f1cat1on of mst1tut1onal mechamsms reqUired to support market opportumt1es w1th eros1on control potent1al d) Ex ante evaluat10n of adoptab1hty e) lmplementat10n of mst1tut1onal mechamsms on a pllot bas1s f) Ex post evaluat1on g) Extrapolat10n to other s1tes a) ldentd1cat1on of market opportunn1es (Carlos Ostertag) Th1s w1ll help b1o phys1cal sc1ent1sts narrow down opt1ons for mclus1on m prototype systems The search for market opportumt1es w1ll be based on Ansoff s (1957) framework for strateg1es to ach1eve growth m markets Current Products New Products Current Market Product Markets penetrat1on development New Market D1vers1f1cat1on Markets development Market penetrat1on uses marketing tools such as pnce reduct1on or quahty 1mprovement to ach1eve mcreases m market share m current markets Returns and nsk are usually low Market development 1dent1f1es new users (such as mstltUtlonal users) or new geograph1c areas (such as export markets) for selhng the same products as befare Returns and nskmess are usually moderate Product development offers new or mod1f1ed products such as orgamcally produced products 1n current markets Th1s 1s also a strategy of moderate returns and nskmess D1verslf1cat10n offers new products m new markets eg exportmg flowers These are usually charactenzed as h1gh nsk h1gh returns strateg1es A demand dnven approach w1ll be taken w1th the 1dentlflcat1on of opportumt1es bemg based on the needs and wants of target markets such as supermarkets manufacturers of processed foods mtermed1anes etc b) Eros1on control potent1al of 1dent1f1ed market opportumt1es Th1s act1v1ty wlll be carned out mamly by the b1ophys1cal sc1ent1Sts m the projeCt (Karl Muller J 1 Sanz Jesus Castillo) and 1s expected to suggest new opt1ons wh1ch pnor to the 1dentlf1cat1on of the1r market value may have been rejected by b1ophys1cal sc1ent1sts on the bas1s of the1r eros1on control ab1htms e) ldent1f1cat1on of mst1tut10nal mechamsms to support market opportunn1es Carlos Ostertag (HSI Th1s w1ll be based on an analys1s comparmg current and h1stoncal marketmg arrangements 246 wrth potentral new rnstrtutronal arrangements proposed by target purchasers of farm products The analysrs wrll assume that productron costs rnclude both transformatron costs and transactrons costs such as search costs rnformatron costs and enforcement costs (North 1990) Based on Coa se s ( 1960) proposrtron that the effrcrent competrtrve equrhbrrum of neoclassrcal economrcs obtarns only when transactrons costs are zero and that rnstrtutrons are mechanrsms for reducrng transactrons costs the analysrs wrllrdentrfy the most rmportant transactrons costs rn current rnstrtutronal arrangements and rdentrfy new rnstrtutronal arrangements whrch mrnrmrze these costs The conceptual framework of Kohls and Ulh ( 1992) wrll be u sed to classrfy current hrstorrcal and potentral new arrangements Three types of marketrng functrons wrll be consrdered (a) exchange functrons or actrvrtres rnvolved rn the transfer of trtle of goods such as buyrng and selhng (b) Physrcal functrons re actrvrtres rnvolvrng handling movrng storrng and physrcally transformrng the commodrty (e) facrhtatrng functrons or actrvrtres rnvolved rn lowerrng the transactrons costs of the other two functrons Examples are Standardrzatron of goods whrch reduces rnformatron costs and rnsurance whrch reduces the cost of uncertarnty Whenever possrble quantrtatrve measures of the extent to whrch each rnstrtutronal arrangement lowers transactrons costs wrll be obtarned Ouahtatrve assessments wrll be used where quantrfrcatron rs problematrc d) Ex ante evaluatron of adoptabrhty Ja1ro Castano (HSJ Thrs wrll rnclude two actrvrtres (r) A study of the extent to whrch conservatron practrces have been adopted (u) the estrmatron of a multrple goal lrnear programmrng model (McGregor and Dent 1993 Cockhn et al 1998 lgnrzro 1985) Thrs rs a normatrve model whrch can be used to rdentrfy potentral confhcts between multrple goals eg farmer s goals whrch could lead to adoptron vs mrnrmrzatron of sorl erosron The analysrs rndrcates to what extent underachrevement of one goal rs requrred rn order to achreve satrsfactron of another For example vetrver grass barrrers may achreve the sorl conservatron goal but may result rn substantral underachrevement of farmers goals and therefore have httle success of adoptron wrthout subsrdres whrch would help farmers achreve therr goals An alternatrve barrrer such as a perennral forage such as elephant grass may underachreve both the sorl conservatron goal and farmer goals wrthout rnstrtutronal arrangements related to mrlk marketrng Approprrate rnstrtutronal arrangements may however enable rt to meet farmer goals and thus lead to voluntary adoptron Thus tradeoffs between the goals become evrdent Thus the modelrs a tool to ard decrsron makers rnvolved rn rural development The specrfrcatron of farmer goals wrll be basad on Helle s research on farmer typologres Ecologrcal goals and ecologrcal rmpacts of technologrcal alternatrves wrll be provrded by Muller Sanz and Castrllo Only on srte effects wrll be consrdered Carlos Ostertag s work wrll be used to rdentrfy new optrons for rnclusron rn the model and to srmulate the rmpact of rnstrtutronal marketrng arrangements e) lmplementatron of institutronal arrangements on a pilot basrs Thrs wrll be the responsrbrhty of consortrum members partrcularly the NGOs 247 f) Expost evaluatron Thrs wlil consrst of a conventronal adoptron study g) Extrapolatron G1ven the heterogene1ty of h1lls1des m tropical Amenca the ob¡ect1ve of a study such as th1s m a particular s1te 1s to denve pnnc1ples about how to proceed w1th a s1m1lar problem 1n other s1tes Cntena w1ll be developed to charactenze 1) lnst1tut1onal arrangements Th1s w1ll be done by charactenzmg each mst1tut1on accordmg to the types of transact1ons costs 1t reduces 2) sml conservat1on technolog1es accordmg to resources reqUired for adopt1on and accordmg to effect1veness m controlhng eros1on 3) farmers soc1oeconom1c and b1ophys1cal env1ronment m terms of transact1ons costs resource ava1lab1hty and the extent of soli deplet1on An expert system w1ll then be developed wh1ch matches charactenst1cs of the farmers env1ronment to charactenst1cs of technolog1es and mst1tUt1ons The mnovat1veness of the research hes m the followmg 1 l The focus of the CG system so far has been on developmg b10phys1cal technolog1es w1th soc1oeconom1c constramts taken as g1ven The ob¡ect1ve of th1s research 1s to overcome a soc1oeconom1c constramt lack of market opportumt1es by mtroducmg a soc1al technology v1z mst1tut1onal marketmg arrangements 2) lncent1ves for adopt1on of conservat1on technolog1es have usually been 1n the form of subs1d1es or regulat1ons the d1ff1cult1es assoc1ated w1th wh1ch are w1dely known Th1s study tnes to develop a method for voluntanly mducmg adopt10n for mcome generat1on reasons w1th conservatlon occurrmg as a byproduct 3) The emphasrs on transactrons costs 1s m hne w1th new trends m econom1cs research onented towards relaxmg the restnct1ve assumpt10ns of neoclass1cal econom1cs (sea Pacer 1993) 4) The use of mult1ple goal programmmg to examme tradeoffs between mcome and resource conservat10n 1s rare m tropical agnculture Progress Report a) Ouahtat1ve mterv1ews on 7 h1stoncal mst1tut1onal arrangements have been completad The results show that mst1tut1ons are unable to reduce the transact1ons costs of selhng output as compared to mtermed1anes lnst1tut1ons have been more successful 1n lowenng mformat1on costs related to quahty reqUirements of target markets and access to 1mproved technolog1es bl Informal mterv1ews w1th a small sample of farmers shows that da1rymg 1s regarded as h1ghly prof1table but very cap1tal mtens1ve One opt1on m ay be mst1tut1onal marketmg arrangements for m1lk or m1lk products w1th better off farmers who would purchase forage from poorer farmers A s1mliar arrangement may be poss1ble 1n the case of panela product1on w1th poorer farmers prov1dmg sugarcane for processmg by better off farmers Technolog1cal ass1stance v1a mst1tUt10nal arrangements w1ll be reqUired 248 as farmers cons1der upland cana product1on unprof1table Current Panela processmg technolog1es are also regarded as only margmally prof1table Much mterest m horticultura was also revealed el Past mst1tut1onal expenences h1ghhghted the cychcal natura of cash crop successes Th1s 1nd1cates the need for a d1vers1f1ed portfoho Under the apertura 1t may also be less nsky to focus on products wh1ch are not tradeable because of bulk or penshab1hty Examples are fresh mllk or forages d) lnterv1ews carned out w1th f1eld staff of the H/S program resultad 1n a prehmmary farmer typology md1genous mhab1tants trad1t1onal Caucanos progress1ve caucanos and nanñenses Ob¡ect1ves att1tude to nsk t1me honzon and att1tude to s01l conservat1on were ascertamed lnformat1on was also obtamed on seasonahty m labor reqwrements cash and toad avallab1hty land tenure and ava1lab1hty of cred1t changas m crop compos1t1on The dommant form of output d1sposal IS to mtermed1anes References Anderson J R and Thampapllla1 J Pro¡ect and Pohcy lntervent1on Washmgton D C Ansoff 1957 1990 Soll Conservat1on m Developmg Countnes Pohcy and Research Senes 8 The World Bank Ashby J A 1985 The Soc1al Ecology of So1l Eros1on 1n a Columb1an Farmmg System Rural Soctology 50(31 377 396 Carter S E 1991 Anáhs1s Geográfico del Uso de la T1erra en Centroaménca In Vartaman Alarcon D and Alarcon Mlllan E (edsl Agrtcultura Sostemble en las Laderas Centroamertcanas Oportumdades de Colaboract6n lntermstttuctonal Memortas del Taller Coronado Costa Rtca 13 16 de Agosto de 1991 Instituto lnteramencano de Cooperac16n para la Agncultura San José C R 19 73 Coase R H 1960 The Problem of Soc1al Cost Joumal of Law and Economtcs 3 1 44 Cockhn C Lonergan S and Sm1th B 1988 Assessmg Confhcts m the Use of Rural Resources an Analyt1cal Framework Joumal of Rural Studtes 4( 11 9 20 Gutman P 1 988 Desarrollo Rural y Med1o Ambiente en Aménca Latma Bibliotecas Umvers1tanas Buenos A1res lgmz1o J P 1991 lntroduct1on to Lmear Goal Programmmg Pennsylvama State Umvers1ty lzac A M N 1994 Ecolog1cal econom1c assessment of so1l management pract1ces for sustamable land use m tropical countnes In D J Greenland and 1 Szaboles (eds) Sotl Restltence and Sustamable Land Use CAB lnternat1onal Walhngford 77 96 249 Ka1mow1tz D 1992 La Expenenc1a de Centro Aménca y la Repubhca Dom1mcana con Proyectos de Inversión que Buscan Sostemb1hdad en las Laderas Documento para el Semmano sobre Agncultura Sostemble en Aménca Lat1na del BID y del IICA Washington D C 9 1 O Septiembre Kohls R L and Uhl J L 1990 Marketmg of Agncu/tural Products 7th ed1t1on MacM1IIan Press Lamg D L and Ashby J A 1992 Adopt1on of Natural Resource Management Pract1ces by Resource poor Farmers In Zob1sch M A (edl Acceptance of So1/ and Water Conservat1on Strateg1es and Technolog1es DITSL W1tzenhausen Leonard H J 1987 Natural Resources and Econom1c Deve/opment m Central Amenca A Reg10nal Proflle Transact1on Books New Brunsw1ck McGregor M J and Dent J B 1993 An Apphcat10n of Lex1cograph1c Goal Programmmg to Resolve the Allocat1on of Water from the Raka1a R1ver (New Zealandl Agricultura/ systems 41 349 367 N1mlos T and Savage R F 1991 Successful Conservat1on 1n the Equatonal H1ghlands Journal of So!l and Water Conservat!On 46(5) 341 343 North D C 1990 lnst1tut1ons lnstltUtlonal Change and Econom1c Performance Cambndge Umvers1ty Press Cambndge Pacer E C Jr 1993 The Work Agenda m Agncultural Econom1cs Why a Sh1ft 1n Emphas1s 1s Warranted Amencan Journal of Agncultural Econom1cs 75 60 68 Se1tz W D Taylor C R Sp1tze R G F Osteen C and Nelson M C 1979 lmpacts of So1l Eros10n Control Land Econom1cs 55( 11 28 42 Econom1c Thrupp L A 1992 Challenges m Latm Amenca s Recent Agroexport Boom Development WRI New York Tlffen M 1993 250 PUBLICATIONS 1994 Amézqwta E 1994 Eros1on and Farm1ng Pract1ces 1n the Colomb1an Andean Reg1on Poster presentad at the lnternat1onal Workshop on Conservat1on farm1ng for slop~ng uplands 1n Southeast As1a Challenges Opportumt1es and Prospects Orgamzed by IBSRAM 20 26 November 1994 Mamla Ph1hpp1nes Correa F J L E Datnoff M D W~nslow K Okada D K Fnesen J 1 Sanz and G H Snyder 1994 S1hcon def1c1ency of upland nce on h1ghly weathered savanna s01ls 1n Colombia 11 D1seases and gra1n quahty Poster presentad to the IX lnternat1onal R1ce Conference for Latm Amenca and the Canbbean and V Reumáo Nac1onal de Pesqwsa de Arroz (RENAPAJ G01ílma G01ás Brazll March 21 25 1994 (In press) F1sher M J 1 M Rao M A Ayarza C E Lascano J 1 Sanz R J Thomas and R R Vera 1994 Carbon storage by ~ntroduced deep rooted grasses 1n the South Amencan savannas In Nature 15 September 1994 Vol 371 p 236 238 Fnesen D K J 1 Sanz F J Correa M D W1nslow K Okada 1994 S1hcon def1c1ency of upland nce on h1ghly weathered savanna solls 1n Colomb1a 1 Ev1dence of a ma¡or y1eld constralnt Poster presentad to the IX lnternat1onal R1ce Conference for Latm Amenca and the Canbbean and V Reumáo Nac1onal de Pesqwsa de Arroz (RENAPAJ G01ílma Go1ás Braz1l March 21 25 1994 (In press) G1¡sman A J 1994 Simulación de la d1nám1ca de la maten a orgámca del suelo en SIStemas suelo planta ammal Proceed1ngs of the XI Encuentro Nac1onal de Informática Umvers1tana 13 16 Septiembre 1994 Cah Colomb1a G1rard C M G R1ppste1n 1994 Ut1hsat10n de données SPOT HRV pour la cartograph1e de savanes et pílturages dans les Llanos de Colomb1e Bullet1n Soc1été Franc;:a1se de Photogramétne et de T élédétectlon 131 p Goldman A and Sm1th J 1994 Agncultural Transformat1ons 1n lnd1a and Northern N1gena Explonng the Natura of Green Revolut1ons World Development (In Press) M1les J W R J Thomas C E Lascano M J F1sher R R Vera and J 1 Sanz 1994 Evaluat1on of Stylosanthes for selected farm1ng systems of tropical Amanea In de Leeuw P N Mohamed Saleem M A and Nyamu A M (eds ) 1994 Sty/osanthes as a Forage and Fallow Crop Proceed~ngs of the Reg1onal Workshop on the Use of Stylosanthes 1n West Afnca Kaduna N1gena 26 31 October 1992 ILCA Add1s Ababa Eth1op1a p 25 33 Rao 1 M M A Ayarza and R J Thomas 1994 The use of carbon 1sotope rat10s to evaluate legume contnbut1on to so1l enhancement 1n tropical pasturas In Plant and So1l O 1 22 (In press) Reátegw K R R Vera W L Loker and M Vásquez 1994 On farm grass legume pastura performance 1n the Peruv1an ra~nforest Expenmental Agnculture (In press) Rrppstem G and C M Grrard 1994 Utrhzacrón de datos radrométncos espacral y terrestre para el estudro e rnventano de la vegetacrón de los llanos Onentales de Colombra Pasturas Troptca/es Vol 16(2) 2 9 Rondón M A and R J Thomas 1994 A prston actron ball mrll for the raprd preparatron of plant and soll samples for the automated analysrs of mtrogen ('"N) and carbon (13C) In Commun S01l Ser Plant Anal 25(3&4) 435 445 Smrth J 1994 New challenges confrontmg socral screntrsts m IARCS wrth ecoreg10nal mandates In Socral Scrence m the CGCIAR Proceedmgs of a Meetrng of CGIAR Socral Screntrsts held at the lnternatronal Servrce for Natronal Agncultural Research (ISNAR) The Hague the Netherlands August 1992 M P Colhnson and Kerrr Wnght Platars (Eds ) The World Bank Washmgton D C p 11 12 Smrth J A D Barau A Goldman A J H Mareck 1994 The Role of Technology rn Agncultural lntensrfrcatron the Evolutron of Marze Productron m the Northern Gumea Savanna of Nrgena Economrc Development and Cultural Changa 42 (3) 537 554 Smrth J G Weber 1994 Strategrc Research m Heterogeneous Mandate Areas An Example from the West Afncan Savanna In Agncultural Technology Pohcy lssues for the lnternatronal Commumty (ed) J R Anderson CAB lnternatronal Walhngford p 540 560 Smrth J and A M Mandac 1994 SubJectrve versus Objectrve Yreld Drstnbutrons as Measures of Productron Rrsk Amencan Journal of Agncultural Economrcs (In press) Smrth J A D Barau A Goldman and J H Mareck 1994 The Role of Technology rn Agncultural lntensrfrcatron The Evolutron of Marze Productron m the Northern Gurnea Savanna of Nrgena In Economrc Development and Cultural Changa p 537 554 Thomas R J and C E Lascano 1994 The benefrts of forage legumes for hvestock productron and nutnent cychng rn pasture and agropastoral systems of acrd sorl savannas of Latrn Amenca In Lrvestock and sustarnable nutnent cycles rn mrxed farmmg systems of sub Sahara Afnca ILCA Ethropra (In press) Thomas R J 1994 Role of legumes m provrdrng N for sustamable troprcal pasture systems In Management of BNF for the development of more productrve and sustamable agncultural systems Ed J H Ladha M Peoples Kluwer Academrc Pubs (In press) Toutam B B Peyre de Fabregués G Robergé A Brgot and G Rrppstem 1994 Revue des travaux de recherche de vulgansatron et d utrhsatron effectués sur Stylosanthes en Afnque de 1 Ouest (A revrew of research and use of Stylosanthes rn West Afnca) IN P N de Leeuw M A Mohamed Saleem and A M Nyamu (eds ) Stylosanthes a Forage and Fallow Crop Proceedrngs of the Regronal Workshop on the Use of Stylosanthes m West Afnca held rn Kaduna Nrgena 26 31 October 1992 ILCA (lnternatronal Lrvestock Centre for Afnca) Addrs Ababa Ethropra p 7 16 Ullnch C R R Vera y J H Wemger 1994 Produccrón de leche con vacas de doble propósrto en pasturas solas y asocradas con legumrnosas In Pasturas Troprcales Vol 16 No 3 Drcrembre 1994 p 29 30 252 Vera R R J 1 Sanz P Hoyos O L Mohna M R1vera and M C Moya 1994 Pasture establishment and recuperat1on w1th undersown nce on the ac1d so1l savannas of South Amenca In HUJsman E A J W M Osse O van der He1de S Tamm1nga B J Tolkamp W G P Schouten C E Holhngworth and G L van Wmkel (eds 1 B1olog1cal Bas1s of Sustamable Ammal Product1on Proceedmgs of the Fourth Zod1ac Sympos1um Wagemngen The Netherlands Apnl 13 15 1993 EAAP Pubhcat1on No 67 1994 Wagemngen Pers Wagemngen The Netherlands Yonkeu S E Teodenken Pamo and G R1ppstem 1994 An evaluat1on of sorne access1ons and vanet1es of Sty/osanthes mtroduced m Adamawa Plateau Cameroon IN P N de Leeuw M A Mohamed Saleem and A M Nyamu (eds 1 Sty/osanthes a Forage and Fallow Crop Proceedmgs of the Reg1onal Workshop on the Use of Stylosanthes 1n West Afnca held 1n Kaduna N1gena 26 31 October 1992 ILCA (lnternatlonal L1vestock Centre for Afncal Add1s Ababa Eth1op1a pp 97 1 02 253 MEMBERS OF THE TROPICAL LOWLANDS PROGRAM PrinCipal Staff Raul R Vera Ph D Ammal Sctenttst Program Leader M1gueiA Ayarza Ph D S01l Sctenttst Crop Pastura Systems Cerrados (Stattoned 1n Bras1ha Braztl) My/es J F1sher Ph D Ecophystologtst Ecophystology Denms K Fflesen Ph D Sotl Sctenttst Nutnent Cychng (IFDC) Georges R1ppstem Dr agr Grassland Ecolog1st lmproved Grasslands Research (EMVT CIRADl José IgnaciO Sanz Ph D S01l Sctenttst Crop Ltvestock Systems Llanos Joyotee Sm1th Ph D Agncultural Econom1st Economtcs R1chard J Thomas Ph D Plant/Mtcrobtal Phystologtst N Cychng and Nttrogen Ftxatton M1chael Thung Ph D Agncultural Sctenttst Mult1 spec1es Productton/Forest Marg1ns Edgar AmézqUita Ph D Semor Research Fellow S01l Phystcs/Piant Nutnt1on N Cychng and Nttrogen F1xat1on Ar¡an J G1¡sman Ph D Semor Research Fellow Soll Sctence/Piant Nutnt1on N Cychng and Nttrogen F•xat1on Astrld Oberson Dr se agr Postdoctoral Fellow Nutnent Cychng Research Assoc1ates Neuza M Asakawa M Se B~alog1st N Cychng and Nttrogen Ftxatton Phanor Hoyos M Se Antmal Sctenttst L1vestock Productton Systems (Vtllavtcene~o) Research Ass1stants Héctor Fab1o Alarcón Agronom1st N Cychng and Nttrogen F1xat1on Gonzalo Antomo Borrero Tamayo Btochemtst Nutnent Cychng José V1cente Cadawd Agncultural Econom1st Economtcs LUIS Fernando Chfwez Btologtst Ecophystology/Soll Phys1cs Germán Escobar Berón Btologtst lmproved Grasslands Research Canmagua Cynth1a Gom1de Agronomtst Crop Pastura Systems Cerrados (Brastha Braztll D1ego LUis Molma López Agronomtst Crop Ltvestock Systems Llanos (Vtllavtcencto) Glofla Isabel Ocampo Bactenologtst N Cychng and Nttrogen Ftxatton Maflela R1vera Pena Agronomtst Nutnent Cychng Be/1saf/o Volverás Agronomtst Ecophystology (Canmagua) 255 Post and pre graduate students Ph1flpp Aeby Eng post..graduate student 1n Env1ronmental Se~ence at the E1dgenoss1sche Techmsche Hochschule CETHJ Sw1ss Federallnst1tute of Technology Zunch Sw1tzerland Jalfo Alexander Cepeda Umvers1dad Nac1onal de Colombia Facultad de Agronomfa Palm1ra (Agronomy) Marganta Escobar Berón Umvers1dad Nac1onal de Colombia Facultad de Agronomfa Palm1ra (Agronomy) (Untll March 31 1994) Céci/e Grol/1er lnst1tut Supéneur Techmque d Outre Mer IISTOMJ Centre Polytechmque Samt LoUis France Juan José J1ménez post graduate student m B1ology at the Umvers1dad Complutense Madnd Spam Secretan es Carmen Cervantes de Tch1ra Amparo J1ménez A C1efo Nunez de Rodrfguez Magda Soares Lambert (Brasilia Braz11) 256 COLLABORATING INSTITUTIONS Bayreuth Umvers1ty Germany CIAT Centro de Jnvest1gac1ón Agrfcola Tropical Santa Cruz de la S1erra Bollv1a CIRAD EMVT Centre de Coopérat1on Jnternat1onale en Recherche Agronom1que pour le Développement Département d Elevage et de Médecme Véténnaue des Pays Trop1caux Ma1sons Alfort Franca CNJ JCA CJAT Centro Nac1onal de lnvest1gac1ones Jnst1tuto Colombiano Agropecuano Centro Internacional de Agncultura Tropical Canmagua Colombia CNPAF Centro Nac10nal de Pesqu1sa de Arroz e Fel]ilo Go1i!ma GO Brazll CNPGC Centro Nacional de PesqUisa de Gado de Corte Campo Grande MG Brazll CNPMS Centro Nac1onal de PesqUisa de Ma1z e SoJa Brazll Colorado State Umvers1ty Natural Resource Ecology Laboratory Fort Colhns Co USA Cornell Umvers1ty USA CORPOJCA Corporación Colombiana de Jnvest1gac1ón Agropecuana Colombia CPAC EMBRAPA Centro de PesqUisa Agropecuána dos Cerrados Planaltma DF Empresa Bras1le1ra de Pesqu1sa Agropecuána Brazll CPAF/EMBRAPA Centro de PesqUisa Agroforestal do Acre Rfo Branco AC Empresa Bras1le1ra de PesqUisa Agropecuána Brazll FEDEARROZ Federación Nac1onal de Arroceros de Colombia Vlllav1cenc10 Colombia FONAIAP Fondo Nac1onal de lnvest1gac1ones Agropecuanas Venezuela Funda~i!o Getuho Vargas R10 de Jane1ro Brazll ICA Instituto Colomb1ano Agropecuano Colombia IFDC lnternat1onal Fert1hzer Development Center Muscle Shoals AL Umted States IGAC Instituto Geográfico Agustfn Codazz1 Santafé de Bogotá Colombia INA PG lnstltut Nat1onal Agronom1que de Pans Gngnon Franca 257 INRA lnstitut Nationale de Recherche Agronomique Oi¡on France MARNR Mtnisteno del Ambiente y de los Recursos Naturales Renovables Caracas Venezuela ORSTOM lnstitut Francais de Recherche Scientifique pour le Développement en Coopération Bondy Cedex Franca SPOT IMAGE France TSBF Tropical Soil Biology and Fertihty Nairobi Kenya UC Umversidad Complutense Department of Biology Madnd Spain UN Umversidad Nacional Department of Biology Santafé de Bogotá Colombia UN Umversidad Nacional Faculty of Agronomy Palm1ra Colombia UNELLEZ Umversidad NaCional Expenmental de los Llanos Occidentales Ezequtel Zamora Guanare Venezuela Umversidad Federal de Uberlindia Brazll Umvers1té Pans XII Val de Marne Franca Umvers1ty of Hohenhe1m Germany Umvers1ty of Saskatchewan Saskatoon Saskatchewan Canada UT Umvers1dad Tecnológ1ca de los Llanos Onentales Vlilavicencio Colombia Wagenmgen Agricultura! Umvers1ty Department of Agronomy The Netherlands 258