5 Sources of Future Growth in Indian Irrigated Agriculture Mark Svendsen Modern Indian irrigation development goes back in time at least to the construction of the Western Yamuna Canal near Delhi in 1355 by Ferozshah Tughlaq. Much earlier irrigation development in the subcontinent was undertaken by the Harappa and Mohen-jo-daro civilizations of 2500 B.C. and the builders of irrigation tanks in South India and Sri Lanka (Rogers 1983). By 1900, British India^ had about 13.2 million ha of total irrigated area, including 7.5 million ha of public works (4.5 million ha from large-scale public works and 3.0 million ha from minor public works) and 5.7 million ha of private works (4.0 million ha from private wells and 1.7 million ha from other private works) (India, Ministry of Irrigation and Power 1972). Around this time in the United States, interest in using irrigation development as a means for opening the American West was beginning to swell. In 1890-91 an engineer named Herbert Wilson of the United States Geological Survey was sent to India to learn about its large-scale canal irrigation. This visit was followed in 1902 by the creation of the U.S. Bureau of Reclamation, which was established to "reclaim the arid West." In the 85 years that followed, the Bureau constructed irrigation schemes that supplied water to 4 million ha of western farmland. In October 1987, the Bureau of Reclamation made the startling announcement that it was transforming itself "from a construction company to a resource management organization" and would henceforth concentrate on managing existing projects, conserving water, ensuring water quality, and protecting the environment. In doing this it would cut its staff by half and transfer its headquarters from Washington, D.C., to Denver (Shabecoff 1987). Although new construction that had already been authorized will continue in the United States for some years, this shift marks the end of an era in which capture and control of water resources were major federal and state development efforts. No direct parallel with the Indian experience is being suggested, but most of the factors that led the Bureau of Reclamation to this decision-- shrinking opportunities for new construction, rising costs, agricul- Mr. Jorge Ramirez, a research associate at IFPRI, made important contributions in col lee ting and analyz ing data for this report. This includes the areas of present-day Pakistan and Bangla­ desh and is therefore not strictly comparable with subsequent figures for modern India. 43 tural surpluses, large federal budget deficits, and negative environmental impacts and environmentalist opposition to new construction—are present in India. This chapter, and in a sense the workshop itself, assumes that in India these kinds of forces will gather strength in the next ten to twenty years until the expansion of irrigated area becomes less important as a source of growth in agricultural output than other kinds of changes within the irrigation sector. It also assumes that the subsector of irrigated agriculture will necessarily continue to shoulder a major share of the burden of increasing agricultural production. The latter assumption is supported by the static nature of the agricultural land base—around 143 million ha of net sown area (India, Planning Commission 1985b)--and requires that growth in output come almost entirely from increased productivity. To explore the implications of this notion, this chapter first reviews past sources of irrigation-related growth in Indian ir­ rigated agriculture and the nature of remaining potential for expansion. It then speculates on potential sources of future growth. OVERVIEW OF GROWTH PATTERNS SINCE 1950 Agriculture Foodgrain production in India has grown at a compound rate of about 2.7 percent per year since 1950. This has more than kept pace with the population growth rate (about 2 percent) and has reversed the long trend of declining per capita food production that pre­ vailed from the 1920s through the 1940s. There seems to be general agreement that these gains, particularly those attributed to the green revolution, are closely tied to the pace of irrigation development. Estimates of irrigation's contribution to this growth in production vary, ranging downward from Seckler and Sampath's (1985) estimate of 60 percent, but few would deny the importance of irrigation. Daines and Pawar (1987, 2) assert that although "attribution is difficult to assign. . . . few analysts would give irrigation less than half the credit for the progress agriculture has made in India during the last three decades." I do not treat this issue further here. For the purpose of the general case being made, it is not necessary to know precise details of the connection between irrigation and agricultural productivity-- only that it exists, that it is driven largely by expansion of the area under irrigation, and that it is reasonably strong. Irrigation From 1951, when central planning began, until 1983, net ir­ rigated area expanded at a compound rate of 2.2 percent per year (Table 5.1). This overall figure masks, however, some interesting shifts in the composition of this growth. Over the period 1951 to 44 Table 5.1--Percentage growth and average statistics for net irrigated area, by source, 1951-83 1951 -83 1951 -65 1968- -83 Variable Growth Average Growth Average Growth Average (percent)(1 ,000 ha) (percent)(l,000 ha) (percen t)(l,000 ha) Total net irrigated area 2.2 28,881 1.7 23,321 2.4 34,352 Net area irrigated by Private canals -2.2 1,037 -0.5 1,241 -3.4 842 Government canals 2.4 10,677 2.3 8,465 2.3 12,835 All canals 2.0 11,714 2.0 9,706 2.0 13,677 Wells 3.9 10,667 1.6 6,941 4.4 14,412 Tanks -0.5 4,095 2.4 4,312 -1.7 3,837 Other 0.2 2,394 -0.0 2,361 0.4 2,425 Dry land 0.2 123,000 1.1 121,000 -0.3 125,000 1983, the area irrigated by government canals increased at a steady compound annual rate of 2.4 percent, while tank-irrigated area decreased at 0.5 percent per year, and well irrigation grew at a strong 3.9 percent. This growth performance led well irrigation to surpass canal irrigation in net area served for the first time in 1973/74 (Table 5.2). Since then, the gap between area irrigated by wells and that irrigated by canals has continued to widen. Disaggregating the growth rate into two periods (1951-65 and 1968-83, s,ee Table 5. l)--essentially before and after the green revolution --shows that the total net area increased slightly more rapidly (2.4 percent per year) during the second period than during the first (1.7 percent). Among different types of irrigation, creating canal irrigation command proceeded at an even pace--about 2.3 percent per year--during both periods. Tank area grew at a similar rate of 2.4 percent during the first period and then declined at a rate of 1.7 percent during the second. Well irriga­ tion, on the other hand, shows the opposite trend, growing at a modest 1.6 percent during the first period and accelerating to 4.4 percent during the second. Given this pace of growth and the large base that was built, well irrigation will increasingly dominate the irrigation picture in India. Irrigation from private canals The years 1966 and 1967 were periods of severe drought across India, and their omission in time series such as this is common. 45 Table 5.2--Net irrigated area, by source, 1950-83 Canals Other Year Government Private Total Tanks Wells Sources Total (1,000 ha) 1950-51 7,158 1 ,137 8,295 3,613 5,978 2,967 20,853 1951-52 7,534 1 ,194 8,728 3,444 6.517 2,360 21,049 1952-53 7,599 1 ,352 8.951 3,214 6.485 2,427 21,077 1953-54 7,559 1 ,314 8,873 4,187 6,640 2,087 21,788 1954-55 7,833 1 ,161 8,994 4,002 6,702 2,261 21,959 1955-56 8,025 1 ,360 9,385 4,423 6,739 2,211 22,758 1956-57 7,916 1 ,357 9,273 4,492 6,566 2,202 22,533 1957-58 8,303 1 ,349 9,652 4.536 6,818 2,150 23.156 1958-59 8,391 1 ,279 9,670 4,759 6,686 2,286 23,401 1959-60 8,752 1 ,305 10,057 4,648 7,083 2,208 23,966 1960-61 9.170 1 ,208 10,378 4,561 7,290 2,440 24,661 1961-62 9,338 1 ,162 10,500 4,613 7,352 2,420 24,885 1962-63 9,686 1 ,146 10,832 4,781 7,650 2,403 25,666 1963-64 9,848 1 ,158 11,006 4,597 7,786 2,484 25,871 1964-65 9,861 1 ,136 10,997 4,815 7,824 2,520 26,156 1965-66 9.827 1 ,133 10,960 4,441 8,445 2,595 26,441 1966-67 10,200 1 ,000 11,200 4,600 9,200 2,200 27,100 1967-68 10,279 1 ,025 11,304 4,599 9,264 2,356 27,523 1968-69 10,900 1 ,000 11,900 4,000 10,800 2,400 29,000 1969-70 11,300 1 ,000 12.300 4,400 11,100 2,500 30,400 1970-71 11,972 866 12.838 4,112 11,887 2,265 31,103 1971-72 11,949 901 12,850 4,140 12,235 2,607 31,891 1972-73 12,192 863 13,055 3,621 13,024 2,249 31,949 1973-74 12,200 900 13,100 3,900 13,300 2,300 32,600 1974-75 12,664 861 13,525 3,548 14,214 2,423 33,730 1975-76 12,933 858 13,791 3,972 14,444 2,386 34,593 1976-77 13,016 845 13.861 3,901 15,087 2,300 35.149 1977-78 13,727 843 14,570 3,899 15,603 2,479 36,551 1978-79 14,289 839 15,128 3,936 16,427 2,569 38,060 1979-80 13,914 837 14,751 3.482 17,817 2,418 38,478 1980-81 14,456 836 15,292 3,198 17,734 2,585 38,806 1981-82 14,701 496 15,197 3,581 18,549 2,597 39,924 1982-83 14.875 495 15,370 3,112 19,112 2,375 39,969 Sources: The Ford Foundation, Data on the Indian Economy. 1951-1969 (New Delhi: Ford Foundation, 1970); India, Department of Agriculture, Indian Agriculture in Brief. 21st ed. (New Delhi, 1987); India, Ministry of Irrigation and Power, Report of the Irrigation Commission, vol.2 (New Delhi, 1972); India, Central Statistical Organization, Stat ist ical Pocket Book of India (New Delhi: Department of Statistics, Ministry of Planning, 1980); TATA Services Limited, Statistical Outline of India (New Delhi: Department of Economics and Statistics, 1987). 46 declined throughout and at an increasing pace after 1965. Changes in total, total canal, tank, and well irrigation are shown graphi­ cally in Figure 5.1. Regionally, expansion of net irrigated area has been uneven, as seen in Tables 5.3 and 5.4. Regional figures, with the regions defined in Figure 5.2, were developed by estimating compound growth rates between decade endpoints (Table 5.4), based on the average of three years of data for each point (Table 5.3). The years were selected to achieve a reasonable match in rainfall conditions for each period. Growth was strongest in a band stretching transversely across the western peninsula of India up through Uttar Pradesh. For all of India, growth was roughly twice as rapid in the areas irrigated by wells as in the canal command, except in the western region, where canal irrigation grew most rapidly. In the south, expansion was virtually nil as shrinking tank and "other" irrigation was compen­ sated for by increased well irrigation. In eastern India also, strong growth in the area irrigated by wells was partially offset by the contraction of the area under tank and "other" sources. It would be interesting to know the extent of actual geographic overlap in these areas of contraction and expansion. As of 1982-83, net irrigated area (Table 5.5) was relatively evenly distributed among these regions. This is, of course, partly an artifact of the way in which regional boundaries are drawn. Gross irrigated area shows, however, a slightly different pattern, with Uttar Pradesh and eastern India assuming much greater promi­ nence; they contain some 46.0 percent of the nation's gross ir­ rigated area, but only 39.5 percent of its net irrigated area. Presumably this is due to the higher rainfall that prevails over much of this combined area, which produces higher cropping inten­ sity, and the extensive alluvial aquifer that lies below it. The Gangetic basin holds the most gross irrigated area in the nation. Unfortunately, these figures show only the nominal area of irri­ gated coverage and do not address the critical problem of the quality of the irrigation service provided--that is, the ability of the service to produce agricultural output. An area receiving a single irrigation delivery is indistinguishable from one receiving unlimited water on demand. Typically, studies deal with this issue by ascribing a special quality to well irrigation, based on the greater measure of reliability or of farmer control it is felt to have. This, although clearly an inadequate proxy, is about all that the generally available secondary data sources can support. Increasingly, this issue will have to be addressed directly by encouraging the generation and analysis of data on the quantity and timing of irrigation water deliveries over geographic space. Progress in both analysis and in practice depends on it. IRRIGATION POTENTIAL The figures given above show the steady growth in irrigated area as a result of both public investment in canals and largely private (though heavily subsidized) investment in well irrigation. They 47 Table 5.3--Average net irrigated area, by region and source Region and years Canals averaged Government Private Total Tanks Wells Other Total (1,000 ha) Average of 1969/70, 1971/72, 1973/74 Southern 2,628 11 2,639 1 ,955 1,391 289 6,273 Northern 2,367 201 2,569 6 2,173 119 4,867 Uttar Pradesh 2,462 1 2,453 339 3,949 256 7,016 Central 1,518 1 1,519 353 1,772 133 3,776 Eastern 1,860 670 2,530 881 686 1,306 5,403 Western 955 23 978 617 2.218 217 4,029 All India 11,782 910 12,692 4 ,160 12,161 2,480 31,512 Average of 1978/79, 1980/81, 1982/83 Southern 2,660 6 2.666 1 .671 1.826 193 6,356 Northern 2,723 179 2.901 3 2.911 119 5,934 Uttar Pradesh 3,206 1 3,207 188 5,698 316 9,410 Central 2,058 1 2,059 303 2,885 220 5,467 Eastern 2,532 512 3,044 608 1,360 1,101 6,113 Western 1,332 25 1,357 634 3,020 283 5,293 All India 14,534 723 15,257 3 ,407 17,753 2,494 38,912 Source: India, Department of Agricult ure, I ndian Aq r iculti ure in Brief (New Delhi , 1987) Table 5.4--Compound growth rates of irrigated area, by region, 1971/72-1980/81 Region Canals Tanks Wells Other Total Southern 0.12 -1.73 3.07 -4.39 0.14 Northern 1.36 -7.97 3.30 0.00 2.23 Uttar Pradesh 3.02 -6.31 4.16 2.39 3.31 Central 3.44 -1.67 5.57 5.72 4.20 Eastern 2.07 -4.03 7.90 -1.88 1.38 Western 3.71 0.31 3.49 3.01 3.08 All India 2.07 -2.19 4.29 0.06 2.37 Note: The compound growth rate was estimated using the three-year averages presented in Table 5.3. 48 Figure 5.1--Area Irrigated, by source, 1950-82 fhou» X a> 2 «S (D f_ «J • - +» 3 ~ - v « in * » c S c oi X J o <-> C O H - " o 3 c= o O O O O O O O C- CD u S.' CD °-«t o ~ __ o i j a O l m . a t o c i - •O- t - . e t - o OI 01 t- >> O O O O O O C M O O O O O O CO o CM o o r^ »** r— -vr m c\j m co -**- c» . xi J" - « * * X> R 3 +» CD * ; in m +» ^ CD ^2? = - i " o " ^ s s VI X I 01 CD • • - » C» E 2 -M VI CD a •> S 4-> CD •!-> v> o O CD r<> U1 V d) XJ CD « « « w e n + J r _ ; « ~ » : » £ g * •- « - C « ai o m o o o o o o n o i / x o u i o O N ( O N ( O r H ( O l T ) O - - - - - - - i n C M i n i n K (O en «—I *—* CM * - • CM *—• *—< 0 « * ^ (- |— u — - « • tu * * +> E j •—-»-»•— m 0 « O ttf -r- « GO j - ^ s - c o - ^ - o - ^ ^ - T 3 T 3 " » - O -M f— +» •«* i n C C « C Q . C C O C O Z! •- ' « M - CD X 0 0 ) 0 > O +J S- »M CD -*-*«-• «-« f- C «S 3 O C O C D - H " 0 en o - = 3 a . o o (- t- CD m t~ s~ c - C .£= (- J- 0) CD t - t ZJ I - -*-» C W CO •— O O -M CD «J 0 •— < 7 ) Z 3 U U J 3 < 53 Table 5.7--Estimates of ultimate irrigation potential for surface water and ground water, various years, 1972/88 Surface Ground Year Source Water Water Total (mi 11 ion ha) 1972 Irrigation Commission 59.0 22.0 81.0 1975 Fifth Five- Year Plan 72.0 35.0 107.0 1985 Seventh Five- Year Plan 73.5 40.0 113.5 1988 Planning Commission 98.0 50.0 148.0a d Assumes optimal use of available water resources by allowing interbasin water transfer and international cooperation for joint river development and improved management (Seventh Five-Year Plan). very expensive and from transnational schemes that require interna­ tional agreements and cooperation, which have proved elusive in the past. Third, the cost of surface water development is rising as the easily exploitable sites are exhausted and the objective of develop­ ment shifts from protective to productive, or more intensive, irrigation. The overall real cost of building major- and medium- scale schemes has more than doubled in the thirty years between 1950 m and 1980. At 1970-71 prices, Sawant (1986) reports that the per hectare expenditure for major- and medium-scale irrigation con­ struction was Rs 2,770 in the First Five-Year Plan, and Rs 5,880 in 1979-80. The anticipated expenditure for the Sixth Plan was Rs 6,696. Since undiscovered sources are likely to be more expensive to develop than known ones, exploitable potential probably will not continue to increase significantly. In the groundwater sector, the situation is not so clear-cut. Here, where the resource is hidden from view and assessment is inherently more difficult and imprecise, exploration and quantifi­ cation did not begin until the early 1970s. The interaction between surface water and groundwater resources makes this assessment even more difficult, especially when surface irrigation itself con­ tributes significantly to groundwater recharge. Since 1972, the estimate of ultimate groundwater potential has nearly doubled (Table 5.7) and it will probably continue to expand. In some states, like Uttar Pradesh, groundwater development has already reached the ultimate potential targeted earlier, but ex­ perience indicates that significant potential remains to be devel- 54 oped (Desai, chapter 4 ) . In Tamil Nadu, on the other hand, ground­ water development in some districts is already constrained, and the overall rate of irrigation expansion in the southern region as a whole is negative. Most authorities agree that water resources in the state are fully allocated at present levels of use efficiency (Kandaswamy 1987). In addition, irrigation is not the only sector to claim the nation's water resources. In chapter 4, Desai indicates that nonirrigation uses of water (domestic, industrial, and cooling) are expected to increase significantly in the years ahead. Not only is the absolute use increasing, but the share is as well. By the turn of the century, just a decade from now, nonirrigat ion uses will require nearly one-sixth of the nation's tapped water resources. To be sure, not all of these uses will be consumptive, so there will be some scope for reuse. As the total relative share of other uses increases, however, the share consumptively used will probably grow as well and water quality considerations will become increasingly important. This means that in some states, such as Tamil Nadu, irrigation will have to run in place just to stand still. There has long been a gap between potential created and poten­ tial used in the figures developed by the state irrigation depart­ ments. Persistent efforts to close this gap, notably through command area development programs, have met with only limited success. One may, therefore, speculate on the remaining scope for expanding irrigated area if one internalizes this gap into the values of ultimate potential. To make this adjustment, ultimate potential figures for each region were reduced by the percentage of the existing gap in use as given in C.G. Desai (1988). This adjustment means that the ef­ ficiency estimates used in computing the ultimate potential are higher than can be justified by actual experience and adjusts them downward. The results are shown in Table 5.8. As can be seen, the ultimate irrigation potential drops to 101.1 million ha, and unexploited potential is reduced from 40.2 percent to 32.9 percent. The effect on surface irrigation potential, which falls 17 percent, is even more profound than that on groundwater potential, which drops only 6 percent, based on the gap levels pre­ vailing in 1984/85. The revised values of unexploited potential by region are shown in Table 5.8. Until the potential gap can be closed, these values are the most appropriate ones to use. COSTS OF IRRIGATION DEVELOPMENT The marginal cost of developing a hectare of irrigated land integrates a number of the factors that influence the feasibility of developing surface irrigation further. These factors include the separation distance between arable land and water source, the diffi­ culty of exploiting the site, the extent of displacement of existing settlements, and the level and cost of the available technology. The all-India expenditure on constructing major and medium irrigation projects between 1950 and 1990 (projected) is shown in Table 5.9. Figures indicate that the expenditure per hectare of potential created (in 1980 Rs) rose from Rs 6,780 per ha in the 55 en CM co -̂ - ^- > c « . — » O C fO J3 01 3 © - ^ - ^ 3 E 00 ••-• +» CL f ) CD 0 U 01 0 O 0 ^ 0 « £ •— 1 , - t - . u +J 01 * - «- +J C > » »_ «- -Q E a> 01 " - c " 2 o +• a - ( , CO j _ o O J " " •»- a . t - ™ Q « X I ai t f (DTI ai +> ™ s 2 ** °> • • - » 0 ™ 5 ^ «+> 0 3 3 3 E I I e 3 3 E I I Gr ou nd tl ma te d ri ef , nt la l f r o wa s e s t wa s es ti Ol c nt la l f r o wa s e s t wa s es ti -M . m c 01 W ^ - * oi n - ^ +*"Z * 3 - D u « O « 01 """» 01 ^ 3 * : •» « 0 I - m ° > » " X I > . - ^ © - ^ T_ * : •» « 0 I - m ° > » " X I > . - ^ * - 01 ^ -a °> •> _ O P r c c= T~ *-> • » +; o o 3 O J- 0 "•" _S a-< o - - *- E , :* « - + » " • " .^_ -M «= c cn 0 +" ? • I - O „ Dl«— w - C o • z - e n - - 0 ai * » o o o -a ai x ai o c v i N i n o * H t o ^ H O B O P I I i l N p i i n o ^ i o ^ N r*-> co oo co c\j co co : ^ oi is oi « ai 0 ai Z O " J) O TJ -T r+^ * ^ " £ •< J 3 °>, QJ CD 2 E - K — o i £ Ej o ai +* •— o > n- ai o JZ 0) Cl O l «r 0 • • - » c ai st ed pe ro T3 ai a . x i n co o i •*!- oo o oo m N c n c s i t - i r v N ^ - o i t n t o t - » C M »-t •4- ^ H CM co T4- cn <-i Ol 3 " ° ,_ 01 c 01 I : y * - • J J I - - a • ai 0 0 _ ,. ai -- i - o T ) " * - o + J - a ~ 0 C C -• 0 <*- 01 ZJ -l-» I - -•-» O - 0 3 O l~ - ! "D Crt Q- D l 0 Q. -- 0 • +* a l i o c oo ai CD —. - +»»•**- i n o 3 oo oo o . cn o> ai «-<«—• o +J o o 0 1 - ( - Q - r - C C TJ ai ai -M -f-> 3 %- +> c: co « •— O O -M 01 0 CD •— ( O Z 3 0 l i J 3 < CO cn c- t - 01 a i «4- ••-ai 01 a : o f 56 C « i- o a •w- a> TJ >> C I a . > x -•- iu o c ai o 3 •*-• i— 3 - - -M C Q_ -M c a> u co I H co co a) •-• oo a i ^* a> u O l C (_ «S C t - u i co co co r>>. ce co co co -•- ai a> -*-» oi ai • — i f- 0_0_ O 0_ a . ID t - . — «- « « X I 3 +J •»- - co a . « a -M -M i- c c a c cu c a> at o > • < • • - » f - -- «*: o <_> a> CO 3 0 « f 0>«— -M CO » « « •*- o i a> (- 3 T J « - < O CU C C — > C CU "•"» »- - ^ I— CU J- Q.-M Q_ X « CO CO O N c o c o o o c n m c o c o c o o o c n N O c o c o i n O i N O O ) H O t n N N ( 0 ( 0 0 ) 0 ) ( O O C 3 H ^ ^ - -»- O O O O O —* —• O O o < H N c o < - « m i n o m o> CO '—• . •— C O - «-CO «- a . CO O f— •a 1 - O -Q - ¥ — oo a X I X I O I 1 — c CD «-H • — • - M CD O + • CD 1= CD C CO o C- B I - »--•-» -< 4 - O CO SO O C X ! --- - « X I » - . — c e o a CD I Q - O «- x i n o u i en i i - 57 First Five-Year Plan to an estimated Rs 15,347 per ha in the Seventh, a compound annual rate of 2.2 percent. These figures must, however, be treated with caution. Since the construction of an irrigation project sometimes takes jnore than one, or even two, five-year plan periods to complete, expenditures shown in the table do not necessarily correspond to the potential created as a result of that expenditure. If the investment level is relatively constant from year to year, this difference does not matter a great deal. When the level of investment is growing, however, as it was in this case, this procedure will seriously overestimate the cost of a hectare of potential created. Table 5.10 presents a better estimate of the real cost of creating an irrigated hectare, although it too possesses certain deficiencies. To estimate these figures, the cost stream associated with each project completed during a plan period was summed to estimate the cost of developing that project. The cost streams for all projects completed during the plan period were then aggregated and divided by the potential created during that period to obtain the area-weighted unit cost. This establishes a direct relationship between the costs incurred and the area actually developed by those expenditures. Because the data were already aggregated by project, no correc­ tion for inflation could be applied within each project cost stream. To compare expenditures among periods, a price index for the middle year of each period was applied to the aggregated cost of the projects completed during the period. Thus to the extent that investment in a particular project also took place during preceding plans, these values underestimate the real per hectare cost of development. Nevertheless, the figures do represent the relative values of this parameter across regions for a given plan. The western and southern regions have the highest cost of irrigation development, while the eastern region has the highest growth of the cost per irrigated hectare. Table 5.11 shows the estimates of expenditure and cost per plan­ ning period derived from Tables 5.9 and 5.10, which should bracket the true cost of development. As can be seen, the first estimate increased by a factor of 2.16, in real terms, in the thirty years between the First and the Sixth Five-Year Plans, while the second increased by a factor of 1.79. For the Sixth Plan, the two show Pant (personal communication) indicates that the actual duration of project construction typically ranges between twelve and twenty years rather than the five to ten years usually shown in project planning documents. This assumes that the duration of project construction is similar in different regions. 59 Table 5.11--Costs and expenditures per irrigated hectare in major and medium systems, by five-year plan Expenditure per Cost per Period Covered Irrigated Hectare3 Irrigated Hectare (1986 Rs) First plan 11,160 6,780 Second plan 14,817 6,706 Third plan 16,305 6,974c Annual plans 13,490 n.a. Fourth plan 16,588 7 - 8 3 8 J Fifth plan 13,866 6,719d 1978/79 19,398 n.a. 1979/80 23,518 n.a. Sixth plan 24,123 12,124 Seventh plan 25,261 n.a. Source: India, Planning Commission, and Tables 5.9 and 5.10. Note: Costs and expenditures are in 1986 Rs. The 1986 exchange rate was U.S.$1 = Rs 12.61. n.a. Not available fFrom Table 5.9 bFrom Table 5.10 cAverage for the period covering the Third Plan and the annual plans. Average for the period covering the Fifth Plan and the two follow ing annual plans. unit costs of Rs 24,123 and Rs 12,124 (1986) per hectare, respec­ tively.8 The estimated real cost per hectare given in the second column of Table 5.11 remained constant for almost twenty-five years and then virtually doubled between the Fifth and Sixth Plans (from US$1,100 to $1,913). This suggests that the economics of building medium- and large-scale systems will become increasingly less favor­ able as the ultimate potential ceiling is approached. Estimating the elasticity of cost per irrigated area relative to average unexploited potential using data from Table 5.10 suggests that a o This measure of cost increase, represented in constant 1986 rupees, largely eliminates the cost escalation attributed to extended periods of project construction, where inflation is to blame. Extended construction periods can still lead to higher, but usually unspecified, costs per hectare due to the inefficient nature of stop and go construction activity. 60 decrease of 1 percent of unexploited potential produces an increase of approximately 2.8 percent in the development cost of an irrigated hectare. Estimating the cost of developing groundwater is more difficult. Groundwater development is the dominant component of the minor irrigation sector in India, and private wells account for the bulk of minor irrigation development. Table 5.12 shows the expenditures in each five-year plan and the total institutional lending for minor irrigation. If groundwater development costs are a propor­ tionate share of total government expenditures for minor irriga­ tion, the state investment required to create one hectare of land irrigated with well water was approximately 3,000 (1980) rupees during the Seventh Plan. The unit cost had declined significantly from its peak of 10,700 (1980) rupees per hectare during the Second Plan. This could indicate decreased reliance on institutional sources of credit for private groundwater development, but the magnitude of the drop, expanded use of less expensive electrically driven pumps, and improved pump and motor technology suggest that real reductions occurred as well. FUTURE SOURCES OF GROWTH Continued growth in agricultural output will be assumed neces­ sary in the indefinite future, and, as indicated earlier, the irrigated sector will be required to bear a major share of this burden. Given the current population growth rates, the continuing need to generate new employment in rural areas, and the traditional emphasis placed on self-reliance in the production of food, espe­ cially foodgrains, the first part of this assumption seems self- evident. Professor B. D. Dhawan (1988a) makes a concise and convincing argument for the second point. The days in which expansion of irrigated area can drive in­ creases in agricultural production seem to be drawing to a close. The prospective portion of the Seventh Plan anticipates that the currently assessed ultimate potential will be fully exploited by 2010. Only about one-third of India's currently assessed ultimate potential remains to be exploited through raw expansion of area, as does only one-fourth of the higher-quality groundwater potential. Moreover, the real per hectare costs of developing the more abundant unexploited surface potential will probably continue to rise as wel 1. The value of the increased production resulting from this expansion may rise commensurately with the costs of exploiting new water sources, but this is by no means assured. More likely, During the Sixth Plan, 88 percent of the minor irrigation program was devoted to groundwater development. Investment from personal savings and informal credit sources also occurs, although because of subsidized interest rates, much of the borrowing is probably institutional. 61 • — a u fl) Q) I — CD CD - U (_ > I C O O 0 ) O i n i n h . « H o r s « - • c o o o c n c M C N J t n i n c M e n N C O o o o m c M N i f ) e n m o N o o m i n ^ - p ) C M « T 3 0 > i — • * - 01 0 > « - M - M « 3 C ( d U C 0 ) 0 ) < t O O < C Q_ O O O O O O " - * ^ o • M CD . — C o i mt CD « 3 E J - C - M GU £= f » > < : CD i n C M ^ - r ^ ^ - i n o o o ^ r o N C D N c n n t M r - . c D ^ - o j o > a > p ^ c M ^ H c o i n ^ j - r n m m • — CD %- i a E o _ O 0> - M I — CD C > CD C S- o o o o o o o o CO «—I C*3 »—• ^ " «—( I— CO CD ( D ^ * ( O r » i - l O O ^ " «—i -^ - i r> • - • ^ - o > o i CD O • — E ai O > f - ai t - "n O ) • — c a> o X c 3 « J o -»-> CD - - -^ - o t- • P C O . co « j c c - M •—• - ^ c : Li_ CD •— J - «s t - -•-» =3 o o • O O O O O O O O • o » m m « H O O ^ o • I H H r o ( O o o a > • f o r - i N c o r v ^ i n P* * - - j - - a CD CD CD e J- o C 3 • » - CD fl) C a . CD r— X S- d i i j ( - CD - Q " D CD O f- •r- CD O O O O O O O O O ( O C M C O ( D C O r - i | s . C M ITi c o ^ r c v j t N j - — i c r > o > o o « • — « f— i — « * . — «d o_*—s t — Q - t ~ Q . Q - » — Q - r - + » C L Q - Q - O - C CD • M C T 3 I d -*-» - C * J - C C t - 0 1 O f - 3 f - - l - > 3 - M C D « l s - o - - c : 3 « # - c x > - i - » - r - C D - C C O - • - C * • - C D - — • L . I / ) h < I L L . < W W «J t o o i —̂ c i _ CD -a > c CD C •"• e > " O 0> o - - CD CD t - L O U M - •a i- C0 o c ex -•-» + J i - c « • - CO CD CO 1 CO E C q - O ) 3 • O f — - - * U O CD o • — 09 + » o >> en «j i— £Z -M -r- -r- 09 .. S- T3 C CD CD « 3 O c U E.— o o (- -r- O 3 S- X C 62 improvements in the quality of irrigation service, especially from surface sources, will be required to induce increases in purchased input application and a shift to higher value crops. Future gains must increasingly come from improving the quality of irrigation service, using water on existing cultivable command area more efficiently, and recycling water not beneficially used for crop evapotranspiration or leaching. Ironically, most of these changes will increase the assessed level of ultimate potential itself. The proposed increase of 34.5 million ha is based in part on improved water management. Thus progress on this front will increase the ultimate potential as well as the intensity of irrigation, the production per unit of water, and other measures of specific productivity. These efficiency-based increases in potential will not, however, occur automatically, and they should be reflected in figures only when there is some reasona­ ble assurance that they can be realized. The foregoing leads logically to a review of potential sources of continued growth. In addition to continued but decelerating expansion of irrigated area from newly developed water sources, these alternatives include (1) the conjunctive use of groundwater and surface water; (2) the improved performance of existing surface systems; (3) the improved use of existing groundwater extraction machinery; (4) the interaction of irrigation service with other factors, such as fertilizer use and choice of crop; and (5) the improvement of irrigation technology. This last category is dependent on the others and acts largely through them. In the last sections of this chapter, I discuss briefly the first two of these sources--conjunctive use and improved efficiency of surface systems. Conjunctive Use of Groundwater and Surface Water I leave the task of laying out the principal case for conjunc­ tive use to other chapters in this volume. Professor Dhawan, in particular, has examined conjunctive use extensively in recent years. Instead, I illustrate the importance of shallow groundwater pumping for reusing water that is lost from surface irrigation systems. This is undeniably an attractive notion, but one that is not always fully understood or appreciated. Conjunctive use can be thought of as a mechanism for increasing the efficiency of the surface system that serves as the original source of water. At the same time, it is an important part of the solution to the problems of waterlogging which are felt increasingly in many areas in India. The following model illustrates the importance of the interac­ tion between surface water and groundwater and the way in which they complement each other. It represents the fraction of canal water supplied to a given area that is eventually used to benefit agricultural production and how this fraction responds to changes in the technical efficiency of the surface irrigation system and the Intensity of Irrigation is defined by the 1976 National Commission on Agriculture as the gross irrigated area in an agricul­ tural year, expressed as a percentage of the project's cultivable command area. 63 portion of groundwater that is pumped. For simplicity, groundwater that occurs naturally, that is, not attributable to losses from canal irrigation, is ignored. The terms used are defined as follows. Values of the parameters used are shown in parentheses following the appropriate definition. Bg = groundwater beneficially used by crops, Bs = surface water beneficially used by crops, Eg = overall efficiency of groundwater irrigation (0.7), Es = overall efficiency of surface irrigation (0, 0.25, 0.35, 0.45), Qg = groundwater available for extraction, Qs = surface water delivered to area, U = fraction of percolating water that is unrecoverable (0.20), and X = percentage of reusable groundwater that is extracted. The dependent variable is R, the ratio of canal-derived water beneficially used to the supply delivered by the canal. R = (Bs + Bg)/Qs. (5.1) Beneficially used water is a function of the supply available, the respective overall efficiency, and, in the case of groundwater, the amount of water pumped. Bs = Qs Es, and (5.2) Bg = Qg Eg X. (5.3) The quantity of groundwater (derived from canal sources) is inverse­ ly related to the efficiency of surface irrigation and the fraction that is unrecoverable once it reaches the groundwater aquifer. Qg = (1 - U) (1 - Es) Qs. (5.4) Substituting equation (5.4) into equation (5.3) and then equations (5.2) and (5.3) into equation (5.l), which defines the overall use ratio, results in the following expression. R = Es + (1 - U) (1 - Es) X. The efficiency of groundwater use, Eg, was eliminated because the groundwater not used by the crop was assumed to be lost to deep percolation and then returned to the groundwater aquifer. If several iterations of reuse are allowed, the groundwater use effi­ ciency term approaches unity. This simple model was then used to plot the curves shown in Figure 5.3. As can be seen, when the efficiency of surface water use is zero, the fraction of the available surface water used is a function of the amount pumped and peaks at a maximum level of 80 percent of the supply delivered. This example represents a water spreading operation with no crop being grown and as such is somewhat unrealistic. Since it also ignores evaporation during infiltration, which is unproductive in this case, the peak R value is reduced 64 Figure 5.3--Contribution of conjunctive use to crop water availability 1.0 R 0.9 0.8 0.7 a 6 0.5 0.4 0.3 0.2 0.1 + 0.0 Efficiency = 0.45 Efficiency = 0.35 -7 Efficiency = 0.25 \ \ EfTiderKy = 0.00 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 X Note: R = total fraction of canal water beneficially used; X = fraction of reusable groundwater extracted. somewhat. Enterprising farmers probably would not, however, allow extensive water spreading to take place without taking advantage of the opportunity to produce a crop. At three more realistic (and fairly typical) values of surface Irrigation efficiency, 0.25, 0.35, and 0.45. the fraction of delivered water that is used productively begins at a higher value. rises less rapidly, and peaks at between 85 and 87 percent of the amount delivered by the canal. The peak use attainable rises as surface irrigation efficiencies rise because it is assumed that water is ultimately lost from the system only through subsurface flow of groundwater out of the region. Obviously far more sophisticated, site-specific models are needed to represent this interaction for predictive purposes. This simple presentation illustrates how conjunctive use, or reuse, of canal water and groundwater can raise the technical efficiency of very Inefficient canal irrigation systems to levels equivalent to those of modern well-managed trickle and drip systems. In that, it is quite realistic. The other interesting feature shown on this graph is that the use ratio (R) at any given pumping fraction (X) depends on the 65 efficiency with which the surface systems ultimately supplying the water is operated. This suggests that conjunctive use and efficient operation of surface systems are complementary innovations that partially offset each other and therefore must be considered together. It might be reasonable to regard conjunctive (re)use as a short- to mid-range solution strategy, while in the longer run taking steps to carry out the difficult task of improving the operational efficiency of the surface systems themselves. Improving the efficiency of surface irrigation may be a more efficient approach given the reduced exposure to unrecoverable losses and the energy economies of each approach. Substitutabi1ity is not, however, complete, and some losses from surface system operations are inevitable. Although the scope for efficient conjunctive reuse may first increase and then decrease over time, some scope will always remain, especially in paddy areas. There is another sense, however, in which conjunctive use can be seen as augmenting storage within a particular basin for use on a subsequent crop. In this case, high losses in the surface irriga­ tion system would not only be tolerated, but actively encouraged. This is the gist of an idea that Roger Revelle (1975) proposed in "The Ganges Water Machine." The major attractions of such a scheme are that the associated conjunctive reuse increases cropping intensity and thereby increases and spreads both production and labor demand. Water would probably have a higher value in this case than it would if used during the wetter crop season. Formidable problems are, however, involved in planning, organizing, and implementing such an effort. Improved Performance of Surface Systems Although conjunctive use provides an extremely promising and at­ tractive option for gains in production and efficiency over the short to medium run, intrinsic inefficiencies are associated with it. When water moves below the ground surface it loses potential, energy, and raising it to the surface inevitably incurs costs. Since these costs are paid in energy as well as money, and since energy is also a scarce and constraining resource in many parts of India, we should also look at energy-efficient alternatives to improving performance. Alternatives for improving the performance of existing systems can operate through measures that increase the area served by a given supply of water, increase yields, increase the value of the crop mix grown, or increase cropping intensity. The first three of these can be the result of improvements in the temporal and spatial This is strictly true only if the water is used in the same area from which it was lost. Some of this water can be recovered by standard gravity diversions from natural watercourses downslope of the point at which subsurface flows are intercepted by these channels. 66 pattern of water distribution or its predictability within a given season. The latter requires some form of storage. Most observers estimate the overall technical efficiency of medium and major surface irrigation systems to be around 25 or 30 percent. This means that some 70 to 75 percent of the water diverted from the river or released from the reservoir is not used beneficially, that is, it does not contribute to filling the crop evapotranspiration requirements within the command area of the system. Some of these losses, such as percolation in puddled rice fields or seepage from unlined channels, are unavoidable. These losses are legitimate candidates for recovery through shallow groundwater pumping. On the other hand, uneven spatial allocation of water across large systems and inordinately variable deliveries are important and preventable sources of technical inefficiency in many cases. Questions of equity are obviously involved as well. Achieving these improvements can combine physical rehabilita­ tion, changes in administrative or managerial practices, or institu­ tional improvements. Moreover, they can be targeted below the outlet or above it. A variety of measures including one or more of these components has been developed, experimented with, and applied over a period of many years. These include programs of land leveling, water course improvement, and canal lining; rotational irrigation scheduling based on the North Indian warabandhi frame­ work; command area development authorities and programs; and expanded government responsibility and control. The record, I think it fair to say, is mixed but not predominantly successful. Part of the explanation for this is the tremendous range and number of circumstances found in irrigation systems across India, which make standardized approaches likely to fail in most of the circumstances in which they are applied. Another major reason may be the failure in many cases to address physical, managerial, and institutional issues together. Moreover, the larger context in which improvement programs are undertaken simply may not be con­ ducive to their growth. This leads to a fundamental issue that Sundar (1984) has called the "commitment to manage," or what might be described as the perceived need to manage. Aircraft operation and maintenance, for example, would rank very high on this scale. It possesses a well- established system of accountability, and the penalties of system failure are dramatic and highly visible. In the case of irrigation system operation and maintenance, on the other hand, the consequence of failure can be just as serious, but responsibility is diffused and operations and maintenance personnel are rarely evaluated on the basis of the failure of a system to irrigate. The feeling is inescapable that the larger sociopolitical system does not attach a high priority to effective irrigation management. 1 ̂ Increased cropping intensity can also be achieved, to some extent, by changing the irrigation and cropping calendars without storage and by shortening the duration of the first system-wide cropping season and that of between-crop turnaround to allow two crops to be taken during periods of high rainfall and river discharge. 67 Improving irrigation management at the system level requires, in addition to physical control facilities, knowledge, analytic tools, resources (personnel and operating expenses), and the ability to establish appropriate goals and act to achieve them. Beyond that, however, it requires that a value be placed on successful perfor­ mance and that a system of accountability and a climate of incen­ tives exist to reward success. Just how successful efforts to improve system management can be without addressing the questions of what constitutes successful performance and how an accountability system should tie system performance to managerial performance is a critical and unresolved challenge. The forces hypothesized earlier for enhancing the importance of improving the agricultural performance of existing irrigation systems may also provide the pressure necessary for these kinds of changes to take place. In the meantime, it is important to begin to search for and experiment with organizational and procedural models that can be applied when the pressures for improved performance become sufficiently powerful to ensure that they are installed. CONCLUSIONS India has an enviable record of engineering accomplishments in the field of irrigation, and those accomplishments deserve much of the credit for the nation's impressive growth in foodgrain produc­ tion since its independence. Since 1965, well irrigation has been increasingly responsible for this growth in agricultural output, while tank irrigation has declined significantly and net area served by canals has expanded at a steady pace. Compared with current estimates of ultimate irrigation poten­ tial, the area remaining to be brought under irrigation is dwin­ dling. As of 1984-85, about three-fifths of India's ultimate potential were already being exploited. If the estimate of ultimate potential is adjusted downward by the extent of the "utilization gap" currently prevailing in completed projects, then the ultimate potential that remains falls from 40 to 33 percent. The real marginal costs of creating new surface irrigation capacity are increasing, and economic justification for exploiting the remaining surface water resources is becoming increasingly difficult. The marginal costs of developing groundwater, on the other hand, are falling, though water resource constraints loom here. India is approaching a crossroads where it must choose a new path to sustaining growth in irrigated production. This path leads toward improving the operational performance of existing surface systems by encouraging conjunctive use of surface water and ground­ water (particularly the groundwater that derives from surface system losses), by improving directly the operational efficiency of surface systems themselves, and by improving the use of existing groundwater pumping capacity. Somewhat paradoxically, such programs, once they have demonstrated their effectiveness, will increase the computed level of India's ultimate irrigation potential and reduce the share classed as exploited. 68 Reusing water lost from canal irrigation can result in very high rates of use of the surface water diverted. However, significant additional investment and operating costs are involved in exercising this option. Thus a long-term strategy might rely on conjunctive use over the short run, coupled with gradual improvement in canal operating efficiencies over the long run. A variety of measures can be used to improve canal performance. Ultimately, the sustained success of a significant number of these may depend on creating incentives for irrigation officials and farmers that include a system of accountability for actual perfor­ mance measured against a set of mutually agreed goals. This may sound simple and straightforward, but in practice it is exceedingly difficult to implement and affects the character of irrigation departments and their relationship with both farmers and the larger administrative and political structures. Other areas in which continued growth can be sought include improving the use of existing groundwater extraction machinery through, for example, privatization and development of water markets and pump irrigation societies; applying new technology to the irrigation process; and improving further the availability of complementary inputs and the coordination of input services.