Kenya: Climate variability and climate change and their impacts on the agricultural sector Mario Herrero1, Claudia Ringler2, Jeannette van de Steeg1, Philip Thornton1, Tingju Zhu2, Elizabeth Bryan2, Abisalom Omolo1, Jawoo Koo2, An Notenbaert1 1International Livestock Research Institute, PO Box 30709, Nairobi, Kenya 2International Food Policy Research Institute, 2033 K Street, Washington DC, US Revised Draft 2010 3 Contents Executive summary --------------------------------------------------------------------------------------- 7 Main observations ----------------------------------------------------------------------------------------- 8 1. Background -------------------------------------------------------------------------------------------- 11 2 Current climate characteristics ---------------------------------------------------------------------- 13 Climate variability ------------------------------------------------------------------------------------ 19 3 Projected climate change ----------------------------------------------------------------------------- 23 Projected changes in extreme events --------------------------------------------------------------- 25 4 Impacts of changes on agricultural production ---------------------------------------------------- 30 Length of growing period ---------------------------------------------------------------------------- 38 Agro-ecological zones of Kenya -------------------------------------------------------------------- 33 Crop yields --------------------------------------------------------------------------------------------- 41 Wider Effects on the Economy ---------------------------------------------------------------------- 45 5 Variability, vulnerability and livelihoods ---------------------------------------------------------- 53 Impact of increased climate variability on livestock assets of pastoralists ------------------- 53 Croppers to livestock keepers: Possible livelihood transitions due to climate change ------ 56 Conclusions ----------------------------------------------------------------------------------------------- 58 References ------------------------------------------------------------------------------------------------ 61 4 Figures Figure 1. Current conditions for temperature (2000). From left to right: the mean average of monthly data on temperature, maximum temperature of warmest month, and minimum temperature of coldest month (Hijmans et al., 2005). ....... 14 Figure 2. Current conditions for rainfall (2000). Left, mean annual rainfall (Hijmans et al., 2005). Right, the coefficient of variation of annual rainfall (Thornton and Jones, 2008). .......................................................................................................... 14 Figure 2a. Flood prone regions of Kenya (Otiende 2009). ..................................................... 18 Figure 3. Time series of rainfall departures for individual seasons (solid lines) compared with the annual rainfall departure series. Data are representing eastern Africa as a whole, and are expressed as a percent standard departure (Nicholson, 1996). ..................................................................................................................... 20 Figure 4. Linkage between the Palmer Drought Severity Index (PDSI) and GDP growth , Kenya 1975-1995.( IFPRI 2006). ........................................................................ 22 Figure 5. Seasonal rainfall totals for the short rainy season (October, November, December) at Makindu, Kenya (1959?2004). ..................................................... 227 Figure 6. Percentage changes in the amount of rainfall around 2100 in short rains high rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006). ..................................................................................................................... 28 Figure 7. Percentage Changes in the amount of rainfall around 2100 in short-rains lowest rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006). ....................................................................................................... 28 Figure 8. Changes in the amount of rainfall around 2100 in long rains hight rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006). ................................................................................................................... 298 Figure 9. Changes in the amount of rainfall around 2100 in long rains lowest rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006).. .................................................................................................................... 29 5 Figure 10. The production of main agricultural commodities in Kenya over time (FAOSTAT, 2009). ............................................................................................... 32 Figure 11. The yields of main agricultural comoodities in Kenya over time( FAOSTAT 2009. ...................................................................................................................... 39 Figure 12. The agro-climate zone map of Kenya (KSS, 1982). ............................................. 34 Figure 13. The agro-climate zone map of Kenya, based on LGP classes. ............................. 36 Figure 14.The in length in growing period (in days) for 2000 (Thornton et al., 2006)???39 Figure 15. Percentage change in length in growing period to 2030 and 2050 in Kenya (Thornton et al 2006) ............................................................................................. 40 Figure 15a.Climate change impacts on yields of key commodities in Kenya to 2050 as projected by 6 different models X scenario combinations??????? 43 Figure 16. Models used and Flow of the analysis of the impacts of climate change on crop yields and the wider impacts on the economy of Kenya??... ???.. ...... 46 Figure 17. Maize yield, historic climate and alternative climate change scenarios (kg/Ha).. ................................................................................................................ 49 Figure 18. Kenya:Change in net cereal and maize trade, Kenya, alternative climate change scenarios, (thousand metric tonnes). ........................................................ 51 Figure 19. Per capita calorie availability per day, alternative climate change scenarios, Kenya. (IFPRI Impact Simulations, 2009). ........................................................... 51 Figure 20. Kenya share of Malnourished children, historic climate and alternative climate change scenarios, 2025( %). IFPRI 2009. ............................................................. 52 Figure 21. Kenya: share of Malnourished children, historic climate and alternative climate change scenarios, 2050( %). IFPRI Impact simulation (2009). ............... 53 Figure 22. Evolution of Total Herd size and the number of adult females under two scenarios of climate variable: (1) a drought every five years, and (2) every three years?????????????????????????. 55 Figure 23.Transition zone in the mixed rain fed arid semi arid system, in which the Reliable Crop Growing Days (RCDG) falls below 90 between 2000 and 2050, as projected using the Herd CM32 model and the A1F1 scenario (Jones and Thornton, 2009)???????????????????????...58 6 Tables Table 1. Regional predictions for climate change in Africa by the end of the 21st century (IPCC, 2007). ........................................................................................... 16 Table 1a: Areas affected and number of people affected by floods (Otiende 2009)??? 17 Table 2. Kenya?the total production, average price and value of production for main agricultural commodities, average values for years 2004 to 2006. ....................... 23 Table 3. Descriptions of different moisture availability and temperate zones, used for the agro-climate zone map of Kenya (KSS, 1982). ............................................... 31 Table 4. Kenya: The area under cultivation for different crops over different LGP classes for 2000, and the predicted average difference and standard deviation in area under cultivation over different LGP classes for the 2050, based on averages of the ECHam4 and the HadCM3 GCM for different scenarios. ........... 34 Table 5a. Human population and livesock numbers in different agro ecological zone of Kenya .................................................................................................................... 37 Table 5b. Production of Key agricultural commodites by agro ecological zone?????37 Table 6. Agricultural commodity prices, alternative climate change scenarios. .................. 48 Table 7. Economic impacts of increased drought frequencies in pastoral and agropastoral systems in arid and semi-arid regions of Kenya ................................................ .56 7 Executive summary This document was produced as an output of the project ?Climate Change Adaptation for Smallholder Agriculture in Kenya? funded by The World Bank and executed by the International Food Policy Research Institute, the International Livestock Research Institute, the Kenyan Agricultural Research Institute and the University of Georgia. The objective of this report is to provide an assessment of the impact of climate change and variability on the agriculture sector and economy of Kenya as an initial task to devise adaptation strategies for smallholders in selected agroecological zones of the country. The following tasks were carried out: 1) a review of the historic performance of the Kenyan agricultural sector under varying climate 2) climate variability and climate change impact analyses with special reference to Kenya 3) assessments of the impacts of climate change on crop yield, production, and livestock yield and numbers using crop and livestock simulation models 4) assessment of the wider effects on the economy using IFPRI?s IMPACT model. 8 Main observations 1. Kenya might get wetter. In Kenya, as in most of East Africa, there are very few places where rainfall means are likely to decrease. The increase in rainfall in East Africa, extending into the Horn of Africa, is robust across the ensemble of GCMs, with 18 of 21 models projecting an increase in the core of this region, east of the Great Lakes. 2. The increases in rainfall and temperature will only translate in increased agricultural productivity in specific locations. Increases in rainfall may not lead to increases in agricultural productivity in lowland regions since increases in temperature will also increase evapotranspiration and offset any potential increase in productivity. On the other hand, increases in temperature may remove crop growth constraints in the highlands, thus potentially leading to higher yields. However, to really capitalise on the potential yield increases it will be necessary to invest in these regions in inputs and services. 3. Overall, Kenya will experience country-wide losses in the production of key staples. There seems to be large uncertainty about the magnitudes of the country-wide staple production losses. These vary between minus 10-55% depending on the scenario, crop model and GCM run. Even with modest increases in maize and bean production in the highlands, the country-wide impacts will be a decrease in the production of the major staples due to the large areas where evapotranspiration could increase. 4. However, trade in key staples is likely to offset reductions in crop production due to climate change. Trade in cereal production is likely to increase as a result of climate change to satisfy internal consumption. Under climate change, maize and total cereal imports would be much higher for two out of the three scenarios examined, by between 21 and 44 percent. Under the Hadley scenario, on the other hand, maize imports would be 63 percent below the scenario without climate change. 5. However, prices of key staples are likely to increase and this will dampen demand for food, as affordability of nearly all agricultural commodities?including basic staples and livestock products declines under climate change. As a result, per capita calorie availability in Kenya is likely to decline under all climate change scenarios. 9 6. Lower food accessibility due to increased commodity prices is likely to translate in increases in malnutrition, especially of young children. Climate change is likely to increase the number of malnourished children in both 2025 and 2050. Without climate change, child malnutrition levels are projected to decline from 19 percent in 2000 to 15 percent by 2025 and 11 percent by 2050. Under climate change, child malnutrition levels increase under all alternative climate change scenarios. These effects will probably be exacerbated in areas of high vulnerability, like in the arid and semi-arid areas Kenya. 7. Increased drought frequencies to more than a drought every five years could cause significant, irreversible decreases in livestock numbers in arid and semi-arid areas. Results indicate that a drought once every five years (i.e., representative of current conditions) keeps herd sizes stable in ASALs, and this has in fact been observed in places like Kajiado for a long time. Increased probability of drought to once every three years, could decrease herd sizes as a result of increased mortality and poorer reproductive performance of the animals. This decrease in animal numbers would affect food security and would compromise the sole dependence of pastoralists on livestock and their products, as well as the additional benefits they confer. This simple analysis shows that under increased climate variability, the need for diversification of income, a strategy often (and increasingly) observed in pastoral areas, becomes ever-more important. Climate change and increasingly climate variability will have substantial impacts on environmental security as well, as the conflicts (usually over livestock assets) often observed in these regions are likely to escalate in the future. 8. Kenya will have significant areas in the ASALs where cropping might no longer be possible as a result of climate change and where the role of livestock as a livelihood option is likely to increase. Even under even a moderate greenhouse gas emission scenario for the coming decades, there are likely to be substantial shifts in the patterns of African cropping and livestock keeping to the middle of the century. Potential livelihood transition zones can be identified, and these zones differ in their accessibility, which may have considerable impacts on the type of adaptation options that may be viable. For transition zones that are remote, both market and off-farm employment opportunities may be limited. Substantial changes may be required to people's livelihoods and agricultural systems if food security is to be improved and incomes raised. There will be an increasing need in these areas for highly-targeted schemes that promote livestock ownership and 10 facilitate risk management where this is appropriate, as well as efforts to broaden income- generating opportunities in parts of the continent where this is feasible. 11 1. Background The climate of Africa is warmer than it was 100 years ago and model-based predictions of future GHG-induced climate change for the continent clearly suggest that this warming will continue and, in most scenarios, accelerate (Hulme et al., 2001; Christensen et al., 2007). Observational records show that during the 20th century the continent of Africa has been warming at a rate of about 0.05?C per decade with slightly larger warming in the June? November seasons than in December?May (Hulme et al, 2001). By 2000, the five warmest years in Africa had all occurred since 1988, with 1988 and 1995 being the two warmest years. This rate of warming is not dissimilar to that experienced globally, and the periods of most rapid warming?the 1910s to 1930s and the post-1970s?occur simultaneously in Africa and the rest of the world (IPCC, 2001). The projections for rainfall are less uniform. Hulme et al (2001) illustrated the large regional differences that exist in rainfall variability. East Africa appears to have a relatively stable rainfall regime, although there is some evidence of long-term wetting. There is likely to be an increase in annual mean precipitation in East Africa (Christensen et al, 2007). Many of the impacts of climate change will materialize through changes in extreme events such as droughts and floods. Such extremes result in severe human suffering, and hamper economic development and efforts at poverty reduction. Unfortunately, assessments of climate change are often limited to mean temperature and precipitation. Knowledge of changes in extremes is sparse, particularly for Africa. In some regions, different models project different trends in wet and dry extremes. In other regions, however, models show clear trends such as increasing drought in the Kalahari and increasing floods in East Africa (KNMI, 2006). The challenges climate change poses for development are considerable (Thornton et al. 2006). Despite the uncertainties that exist in long-term climate predictions, it is necessary to explore the sensitivity of the environmental and social systems, and economically valuable assets to climate change (Hulme et al. 2001). High levels of vulnerability and low adaptive capacity in areas of Africa have been linked to factors such as limited ability to adapt financially and institutionally, low per capita gross domestic product (GDP) and high poverty rates, and a lack of safety nets. For example, sub-Saharan Africa is predicted to be particularly hard hit by 12 global warming because it already experiences high temperatures and low (and highly variable) precipitation, the economies are highly dependent on agriculture, and adoption of modern technology is low (Kurukulasuriya et al. 2006). This document gives an overview of available literature on climate variability and climate change in Africa, and specifically in Kenya. First a description is given of the current climate in Kenya, followed by an overview of the range of predictions on climate change. We conclude with an analysis of the agricultural impacts of climate variability and climate change. Section 2 has been adapted from van de Steeg et al. (2009). 13 2 Current climate characteristics In East Africa large water bodies and varied topography give rise to a range of climatic conditions, from a humid tropical climate along the coastal areas to arid low-lying inland elevated plateau regions across Ethiopia, Kenya, Somalia and Tanzania. The presence of the Indian Ocean to the east, and Lake Victoria and Lake Tanganyika, as well as high mountains such as Kilimanjaro and Kenya induce localized climatic patterns in this region (KNMI, 2006). Mean temperature varies with elevation. In Figure 1 the difference between the lowest minimum and maximum temperatures for highland regions is in the order of 8?10?C. Kenya?s climatic conditions vary from a humid tropical climate along the coast to arid areas inlands. While mean temperature varies with elevation, the more remarkable climatic variation is with respect to precipitation (Figure 2). Kenya experiences a bimodal seasonal pattern as it lies astride the equator: the long rains season starts around March and runs through to June, with the peak centred on March to May; the short rains run from September and taper off in November or December (coinciding with the shifting of the Inter-Tropical Convergence Zone). The annual rainfall and the simulated coefficient of variation of annual rainfall (the standard deviation of annual rainfall divided by the mean expressed as a percentage) at a resolution of 10 arc-minutes are shown in Figure 2. Rainfall is correlated to topography; for example the highest elevation regions receive up to 2300 mm per year whilst the low plateau receives only 320 mm. Over two-thirds of the country receives less than 500 mm of rainfall per year, particularly areas around the northern parts of the country (Osbahr and Viner, 2006). The figure shows as well that rainfall is highly variable, especially in the arid and semi-arid regions, and unreliable for rainfed agriculture and livestock production. The rainy seasons can be extremely wet and often either at rapid onset or late, bringing floods and inundation, such as in 2000 (Anyah and Semazzi, 2007). Major floods periodically afflict the Winam Gulf of Lake Victoria, the Lower Tana River basin and the coastal regions (see Figure 2a). Rainfall in this region is strongly linked to the El Ni?o-Southern Oscillation (ENSO) (Ropelewski and Halpert, 1987; Ogallo et al., 1988; Indeje et al., 2000; Mutai et al., 1998). Links between El Ni?o events and climate variability have been suggested, and it is a common perception that high coefficients of variation in rainfall may be attributed to El Ni?o effects (Anyah and Semazzi, 2007). However, currently it is not clear whether a relationship 14 exists between both El Ni?o or La Ni?a events and prolonged drought or particularly wet periods over much of the Greater Horn of Africa (Thornton et al., 2006; Conway et al., 2007). Figure 1. Current conditions for temperature (2000). From left to right: the mean average of monthly data on temperature, maximum temperature of warmest month, and minimum temperature of coldest month (Hijmans et al., 2005). Figure 2. Current conditions for rainfall (2000). Left, mean annual rainfall (Hijmans et al., 2005). Right, the coefficient of variation of annual rainfall (Thornton and Jones, 2008). The arid and semi-arid areas (ASALs) that comprise of 80 percent of total landmass in Kenya are also prone to floods, despite their low levels of rainfall of 300-500 mm annually (WRI et al. 2007). Otiende (2009), citing UNEP (2009) states that flood-related fatalities account for 60 percent of all disaster victims in Kenya (see Table 1a). However, this statement needs to be 15 seen in the light that drought-related damage data are seldom well accounted for (personal communication, EM-DAT disaster database). Kenya experiences major droughts every decade and minor ones every three to four years. In recent years, critical drought periods in the country were experienced in 1984, 1995, 2000 and 2005/2006 (UNEP/GoK 2000). Kenya faced a major drought in 2009 that affected all regions according, leading to hunger and starvation of an approximate 10 million of people countrywide after a poor harvest, crop failure and rising commodity prices (Kenya Red Cross, 2009). The impacts of these droughts on the population are increasing (Table 1) due to high population growth and increasing encroachment of agricultural activities in the arid and semi- arid regions, classified as ASALs. The arid and semi-arid regions are intensifying, and changing from rangeland to mixed systems. This transition from pastoralism to agro- pastoralism is ongoing in many places throughout Africa (Reid et al., 2004; 2008). This is also demonstrated by the reductions in land area in the rangeland based systems towards increases in areas of mixed systems, and the substantial increases in the livestock populations in the mixed systems leading to more intensive types of production systems (Herrero et al., 2008). In Kenya changes from pastoral to mixed systems are projected to occur at rates of 1.2- 2% per yr in terms of area (Herrero et al. 2008). This is not dissimilar to the trends observed up to now. 16 Table 1 Number of people in Kenya requiring relief in the worst flood and drought disasters since 1971 (Osbar and Viner, 2006). Year Type of disaster No. of people affected 2009* Floods 750 000 2009* Drought 3 800 000 2006 Floods 723 000 2006 Drought 3 000 000 2005 Drought 3 500 000 2003 Floods 45 000 2002 Floods 60 000 2001 Drought 3 400 000 2000 Drought 2 740 000 2000 Floods 125 000 1998 Floods 539 000 1997 Floods 212 000 1993 Drought 1 200 000 1992 Drought 2 700 000 1984 Drought 600 000 1979 Drought 40 000 1971 Drought 130 000 130,000 *Data since 2006 based on estimated Kenya Food Security Steering Group The droughts are often followed by periods of intensive rainfall. Torrential rainfall experienced during the wet months often translates into high stream/river flow (runoff) in permanent and intermittent streams/rivers across the country resulting to seasonal floods (Otieno, 2009). Major floods periodically afflict the Winam Gulf of Lake Victoria, the Lower Tana River basin and the coastal regions. Rainfall in this region is strongly linked to the El Ni?o-Southern Oscillation (ENSO) (Ropelewski and Halpert, 1987; Ogallo, 1988; Indeje et al., 2000; Mutai et al., 1998). Osbahr and Viner (2006) indicate that rainfall seasons can be extremely wet and erratic resulting to both large and small river devastating floods like the El Ni?o floods of 1997/98 with significant socio-economic impacts. The 1997/98 El Ni?o flood was associated with one of the largest flood losses in the country in 50 years (Mogaka et al., 2006). The economic and financial losses associated with the El Ni?o flood is in the range of up to US$800 million (Karanja et al., 2002). The 2000 and 2006 droughts were the worst in at least 60 years, and between these two extreme years, several other rainy seasons have failed. Climate change introduces an additional uncertainty into existing vulnerabilities in the ASALs (Osbahr and Viner 2006). At the same time, the number of flood events has increased in frequency and magnitude of people affected. Since 2002, there has been significant flood damage every year in the 17 country. The most significant floods?in terms of number of people affected?occurred in 1997 and 2006. However, compared to droughts, floods continue to affect relatively few people in the country. The increased incidence of floods and droughts might well be a sign of climate change. Table 1a: Areas affected and number of people affected by floods (Otiende 2009). Source: Otiende (2009). 18 Figure 2a: Flood-prone regions, Kenya Source: Otiende (2009) based on WRI (2007). What are the economic impacts of floods and droughts? Mogaka et al. (2006) reports that The 1997/98 El Nino floods and the 1999/2000 La Nina drought cost, on average, at least 14 percent of Kenya's GDP each year of the events; and average annual, long-term costs of extreme events at 2.4 percent of GDP. Another World Bank estimate amounted to 11 percent of GDP for the 1997/98 flood losses and 16 percent of GDP losses attributable to the 1999/2000 drought (Hirji, no date). 19 Climate variability As mentioned, large regional differences exist in rainfall variability. The long rains (March- May) are less variable, so interannual variability is related primarily to fluctuations in the short rains. These are also linked more closely to large-scale, as opposed to local, atmospheric and oceanic factors. Rainfall fluctuations show strong links to ENSO phenomenon, with rainfall tending to be above average during ENSO years (Nicholson, 1996). The importance of short rains for interannual variability is underscored in Figure 3, which compares annual time series of rainfall for the region as a whole with the corresponding time series for four seasons. A visual comparison shows that the similarity is strongest with October-November rainfall. This is confirmed by linear correlation coefficients: the correlation between October-November departures and annual departures is 0.71, compared to 0.53 between April-May and annual rainfall departures (Nicholson, 1996). 20 DJF M = December, January, February and March; JJAS = June, July, August, and September; AM = April, May; ON = October, November. Figure 3. Time series of rainfall departures for individual seasons (solid lines) compared with the annual rainfall departure series. Data are representing eastern Africa as a whole, and are expressed as a percent standard departure (Nicholson, 1996). As Figure 4 shows, climate has been a robust determinant of agricultural sector, and thus general economic performance in Kenya (and elsewhere in rainfed Sub-Saharan Africa). With agriculture accounting for about 26% of the GDP and 75% of the jobs, the Kenyan economy is sensitive to variations in rainfall. Rainfed agriculture is and will remain the dominant source of staple food production and the livelihood foundation of the majority of the rural poor in Kenya. There is a need for the development of the scientific and economic capacity to better understand and cope with existing climate variability (Washington et al., 2006). 21 Figure 4. Linkage between the Palmer Drought Severity Index (PDSI) and GDP growth, Kenya, 1975-1995. (IFPRI 2006) Rainfall amounts and distribution are of paramount importance to rainfed agriculture in Kenya. Figure 5 illustrates season-to-season variability of rainfall totals, for the short rainy season at Makindu (Van de Steeg et al., 2009). As expected, there is great variability in rainfall totals (<150 mm to >800 mm) with a mean of 370 and standard deviation of 180 mm (CV of 49%). Regression lines were fitted to check for evidence of trends in rainfall totals but no statistically significant trend wass extractable from the rainfall data. Rainfall seasonality of this magnitude affects agricultural production and the livelihoods of people, especially in the arid and semi-arid regions, like Makindu (Van de Steeg et al., 2009). Pastoralists have diverse strategies to maintain livestock production. There are several studies that compare how people perceive climate variability, climate change and drought frequency to actual measurements of rainfall variability (Meze-Hausken, 2004; Cooper et al., 2008). In the Makindu example, the general public perception was that local climate had been changing during the last few decades. However, rainfall measurements do not show a downward trend in rainfall (Meze-Hausken, 2004; Cooper et al., 2008). Reasons for the divergence between perceptions and rainfall measurements can be associated with changes in peoples? need for rainfall or be linked to various environmental changes which cause reduced water availability or simply a confusion of the drivers of change in agricultural production and access to reouseces (i.e. increases in population density might have reduced availability of water per family in the region). I N T E R N A T I O N A L F O O D P O L I CY R E SE A R CH I N ST I T U T E So u r ce : I F P R I W HY CL I M ATE M ATTE R S F O R K E N Y A - 2 . 0 0 . 0 2 . 0 4 . 0 6 . 0 8 . 0 1 0 . 0 - 4 . 0 - 3 . 0 - 2 . 0 - 1 . 0 0 . 0 1 . 0 2 . 0 3 . 0 4 . 0 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 D R O U G H T I N D E X G D P G R O W T H 22 Figure 5. Seasonal rainfall totals for the short rainy season (October, November, December) at Makindu, Kenya (1959?2004). Note: Horizontal lines show the mean (370 mm) and ? 1 standard deviation (180 mm) from the mean. Regression lines were fitted to check for evidence of trends in total rainfall. There were no trends that approached statistical significance. The proportion of variation explained by the line was less than 1%. The actual slope was -0.33 mm per year for the rainfall totals. There is a great variety of possible adaptive responses available to deal with climate variability. These include technological options (such as more drought-tolerant crops), behavioural responses (such as changes in dietary choice), managerial changes (such as different livestock feeding practices), and policy options (such as planning regulations and infrastructural development) (Thornton et al., 2009). For example, in the ASALs, livestock herders migrate with their animals in search of pasture and water, with the average distances trekked tripling in drought years. Herding communities typically reserve some pastures back at their homesteads for grazing by vulnerable animals left under the care of women during migration seasons. The herders also ensure that the composition, size and diversity of their animal herds (e.g., a mix of browsers and grazers) suit their variable feed resources and serve to protect them against droughts that could otherwise wipe out their animal stock. 23 3 Projected climate change The climate model simulations under a range of possible emissions scenarios suggest that for Africa in all seasons, the median temperature increase lies between 3?C and 4?C, roughly 1.5 times the global mean response. Half of the models project warming within about 0.5?C of these median values (Christensen et al., 2007). The summary output of 21 Global Circulation Models (GCMs) used by IPCC in their latest report to predict the annual changes in temperature and rainfall that will occur by the end of the 21st century is presented in Table 2. Maximum and minimum predictions of change are given together with the 25, 50 and 75 quartile values from the 21 GCMs (Cooper et al., 2008). Whilst all models agree that it will become warmer, the degree of warming predicted is quite variable. Table 1. Regional predictions for climate change in Africa by the end of the 21st century (IPCC, 2007). Region Season Temperature response (?C) Precipitation response (%) Min 25 50 75 Max Min 25 50 75 Max West Africa DJF 2.3 2.7 3.0 3.5 4.6 -16 -2 6 13 23 MAM 1.7 2.8 3.5 3.6 4.8 -11 -7 -3 5 11 JJA 1.5 2.7 3.3 3.7 4.7 -18 -2 2 7 16 SON 1.9 2.5 3.3 3.7 4.7 -12 0 1 10 15 Annual 1.8 2.7 3.3 3.6 4.7 -9 -2 2 7 13 East Africa DJF 2.0 2.6 3.1 3.4 4.2 -3 6 13 16 33 MAM 1.7 2.7 3.2 3.5 4.5 -9 2 6 9 20 JJA 1.6 2.7 3.4 3.6 4.7 -18 -2 4 7 16 SON 1.9 2.6 3.1 3.6 4.3 -10 3 7 13 38 Annual 1.8 2.5 3.2 3.4 4.3 -3 2 7 11 25 Southern Africa DJF 1.8 2.7 3.1 3.4 4.7 -6 -3 0 5 10 MAM 1.7 2.9 3.1 3.8 4.7 -25 -8 0 4 12 JJA 1.9 3.0 3.4 3.6 4.8 -43 -27 -23 -7 -3 SON 2.1 3.0 3.7 4.0 5.0 -43 -20 -13 -8 3 Annual. 1.9 2.9 3.4 3.7 4.8 -12 -9 -4 2 6 DJF = December, January and February; MAM = March, April, May, JJA = June, July and August; SON = September, October, November. Note: temperature response indicates the projected increase in temperature over current values. 24 For precipitation, the situation is more complicated. Precipitation is highly variable spatially and temporally, and data are limited in some regions (IPCC, 2007). As indicated by Sivakumar et al. (2005) changes in total volume of rainfall in Africa projected by most GCMs are relatively modest, at least in relation to current rainfall variability. Seasonal changes in rainfall are not expected to be large. Great uncertainty exists, however, in relation to regional- scale rainfall changes simulated by GCMs. The problem involves determining the character of the climate change signal on African rainfall against a background of large natural variability compounded by the use of imperfect climate models (Sivakumar et al., 2005). In East Africa there are very few places where rainfall means are likely to decrease (Thornton et al., 2006). The increase in rainfall in East Africa, extending into the Horn of Africa, is robust across the ensemble of GCMs, with 18 of 21 models projecting an increase in the core of this region, east of the Great Lakes (Christensen et al., 2007; Doherty et al., 2009). There is still some uncertainty about this trend, however; as other work suggests that climate models to date have probably under-estimated warming impacts of the Indian Ocean and thus may well be over- estimating rainfall in East Africa during the present century (Funk et al., 2008). If this is correct, then the idea that East Africa will be become wetter in the coming decades may be erroneous. Hulme et al. (2001) discussed two fundamental reasons why there is much less confidence about the magnitude, and even direction, of regional rainfall changes in Africa. Two of these reasons relate to the rather ambiguous representation of climate variability in the tropics in most GCMs via mechanisms such as ENSO, for example, which is a key determinant of African rainfall variability. Another reason is the omission in all current global climate models of any representation of dynamic land cover?atmosphere interactions. Such interactions have been suggested to be important in determining African climate variability during the Holocene and may well have contributed to the more recently observed desiccation of the Sahel (Hulme et al., 2001). Work is now underway, however, to incorporate such links in regional climate models (see, for example, Moore et al., 2009). Limited information on climate change is available for East Africa at country level or local scale. Rainfall projections in Kenya are inconsistent; a range of models and scenarios suggest both increases and decreases in total precipitation (Osbahr and Viner, 2006). Thornton et al. (2006) used changes in aggregate monthly values for temperature and precipitation. For this study possible future long-term monthly climate normals (rainfall, daily temperature and daily 25 temperature diurnal range) were derived by downscaling GCM output to WorldClim v1.3 climate grids at 18 km2 resolution (Hijmans et al., 2005). The outputs from several GCMs and SRES scenarios (Special Report on Emissions Scenarios; IPCC, 2000) were used to derive climate normals for 2000, 2005, 2010, 2015, 2020, 2025 and 2030, using the down-scaling methodology described in Jones and Thornton (2003). These normals were then used with the weather generator MarkSim (Jones and Thornton 2000) to generate daily weather data characteristic of the appropriate climate normals. We used the above-mentioned climate grid data (Thornton et al., 2006) to examine the projected changes in temperature and precipitation for Kenya. While looking at the total annual precipitation projections for Kenya increases in total rainfall in the order of 0.2-0.4% per year were found. These figures for Kenya correspond with findings of long-term wetting by Christensen et al. (2007) and Hulme et al. (2001). However, the regional variations in precipitation are large; the coastal region is likely to become drier, while the Kenyan highlands and Northern Kenya are likely to become wetter. According the UNDP Climate Change Country Profile for Kenya (McSweeney et al., in press) the projections of mean rainfall are consistent in indicating increases in annual rainfall in Kenya. Area average time series show observed climate combined with an ensemble of 15 model simulated recent and future climate under three SRES emissions scenarios (A2, A1B, and B1). The ensemble range spans changes of -1 to +48% by the 2090s. An example of the output is given in Appendix A for the A2 scenario. The projected increases in total rainfall are largest in October-December, but annually these increases are in the order of 20-40 mm per year to 2090 for the arid districts of Kenya. These small increases may be overshadowed if rainfall variability and the frequency of rainfall extreme events increases in the future. Projected changes in extreme events As stated in the Millennium Ecosystem Assessment (2005), natural hazards and disasters are products of both natural variability and human?environment interactions. The extremes of the variability are defined as hazards when they represent threats to people and what they value and defined as disasters when an event overwhelms local capacity to cope. Research on changes in extremes specific to Africa, in either models or observations, is limited. Little can be said yet about changes in climate variability or extreme events in Africa (Sivakumar et al., 26 2005; Christensen et al., 2007). A general increase in the intensity of high-rainfall events, associated in part with the increase in atmospheric water vapour, is expected in Africa, as it is in other regions (Christensen et al., 2007). The increase in the number of extremely wet seasons is increasing to roughly 20% (i.e. 1 in 5 of the seasons are extremely wet, as compared to 1 in 20 in the control period in the late 20th century) (Christensen et al., 2007). Dry extremes are projected to be less severe than they have been during September to December but the GCMs do not show a good agreement in the projected changes of dry extremes during March to May (Thornton et al., 2006; KNMI, 2006). Most climate models simulate drier conditions during the 21st century in eastern Sudan and in Ethiopia. This drying was prevalent during the last decades of the 20th century in these regions. There is little consensus among the models with respect to their simulated changes in extreme rainfall events. A spatially coherent pattern is the increase in 10-year highest rainfall events over northern Somali and the Horn of Africa, and more severe dry events over the same areas. Thus extreme events are likely to become more intense over much of north-eastern East Africa (KNMI, 2006). As noted above, for Kenya there are indications of an upward trend in rainfall under global warming. Wet extremes (defined as high rainfall events occurring once every 10 years) are projected to increase during both the September to December rainy season and the March to May rainy season, locally referred to as the short and long rains respectively. Dry extremes are projected to be less severe in the northern parts of the region during September to December, but the models do not show a good agreement in their projected changes of dry extremes during March to May (Thornton et al., 2006). KNMI (2006) showed the projected variations in wettest events that occur once every 10 years on average. It should be kept in mind that climate models all underestimate the strength of the long rains in the current climate, limiting the confidence of these projections (KNMI, 2006; Thornton et al., 2006). KNMI (2006) used 12 models, on the basis of the realism with which they represent the observed 20th century pattern of African precipitation variation (inter-annual variability and its amplitude). For those models, KNMI investigated the likely changes in precipitation (mean and extremes) using the runs forced with the Special Report Emission Scenario (SRES) A1B scenario. 27 Short-rains (September?December) In the warmer climate around 2100, the GCMs show evidence of an increase in the intensity of extreme rainfall events in much of East Africa, notably in Burundi, Kenya, Rwanda, southern Somali and Uganda. During the short rains, there are indications of the possibility of increases in excess of 50% in 10-year high rainfall events over the north of East Africa. In southern Tanzania the wettest rainfall events are projected to decrease by 0% to 20% (Figure 6) (KNMI, 2006). Figure 6. Percentage changes in the amount of rainfall around 2100 in short rains high rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006). Simulated changes in low-rainfall extremes (Figure 6) show that these events are becoming less severe in Burundi, Rwanda, Uganda, northern Kenya and southern Somali during the September to December season in the most realistic models (with the exception of the Rift Valley in HadGEM1). The simulated increase is far more than 50% in certain parts of the region. Noting that increases in both the wettest and the driest rainfall events have been found over the same areas, this shows an overall shift in the rainfall distribution, with floods becoming more likely than the opposite extreme (KNMI, 2006). 28 Figure 7. Percentage changes in the amount of rainfall around 2100 in short rains lowest rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006). Long-rains (March-May) Even during the long rains, the GCMs continue to simulate an increase in the 10-year highest rainfall events in large parts of East Africa (Figure 7). Over north-eastern Kenya and southern Somali during this season only HadGEM1 does not simulate large increases in the amount of rain in extremely wet seasons. Over southern Tanzania, most models give an indication of an increase in high rainfall events. So while some models show an increase in the severity of extremely low rainfall events in northern Kenya, others simulate a decrease over the same areas. However, these climate models all severely underestimate the strength of the long rains in the current climate, limiting reliability of these projections (KNMI, 2006). Figure 8. Changes in the amount of rainfall around 2100 in long-rains high rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006). However, there is no consensus between the GCMs on the likely changes in the severity of dry events (Figure 8). While some models show an increase in the severity of extremely low rainfall events in northern Kenya, others simulate a decrease over the same areas. Since the model simulations of the 20th century climatology during this season are inaccurate, model projections of future climate during this season are currently unreliable (KNMI, 2006). 29 Figure 9. Changes in the amount of rainfall around 2100 in long rains lowest rainfall events that occur once every 10 years. From left to right, GCM: GFDL CM2.1, MPI ECHAM5, UKMO HadGEM1, and GFDL CM2.0 (KNMI, 2006). Osbahr and Viner (2006) specify that increases in temperatures would have a significant impact on water availability, and are thus expected to exacerbate the drought conditions already regularly experienced and predicted to continue. The unpredictability of Kenya?s rainfall and the tendency for it to fall heavily during short periods are also likely to cause problems by increasing the occurrences of heavy rainfall periods and flooding. Beside the effects of climate change itself, the coastal areas of Kenya should anticipate changes in sea level due to global warming. The projection that sea-level rise could increase flooding, particularly on the coasts of eastern Africa, will probably increase the high socio- economic and physical vulnerability of coastal areas. A rise in sea level in Kenya will have a damaging impact to the production of tree crops situated along the coast (mangoes, cashew nuts and coconuts) and other agriculture based enterprises. A rise in sea level will also affect ecosystems of coastal Kenya, e.g. mangroves and coral reefs with additional consequences for fisheries and tourism (Boko et al., 2007). 30 4. Impacts of climate change on agricultural production in Kenya This chapter examines the impacts of climate change on agricultural production in Kenya. The chapter is divided into three subsections. It first analyses the importance of different crop and livestock products for the Kenyan economy and it discusses the evolution of production trends over time. Secondly, it uses GIS tools to examine the spatial distribution of crops and livestock in different agro-ecological zones of Kenya, while subsection 3 assesses the impacts of potential climate change under a range of scenarios on key Kenyan commodities using the DSSAT crop models (Jones et al. 2001). These data are used as input into the IMPACT model (Rosegrant et al 2007) for examining the effects of climate change impacts on agricultural production changes and on the wider economy (trade and commodity prices) and human well- being outcomes (malnutrition, kilo-calorie availability). Importance of agricultural commodities for Kenya A wide range of commodities are produced in Kenya, the relative importance of these different agricultural commodities varies both spatially as temporally. To assess the relative importance of agricultural commodities, the value of agricultural production was determined (Van de Steeg et al., 2009). A better understanding of the sensitivity of the agricultural sector can be assessed by calculating the value of production (and therefore importance) of different agricultural commodities. This could help target investments and adaptation options for the different commodities and regions (Freeman et al., 2008). The value of production (VOP) was calculated using the formula. The value of production (VOP) was calculated using the formula: VOP i = ?(Prod i * Price i) where: VOP i = Value of production for commodity i (US$) PROD i = Production of commodity i (MT) PRICE i = Price of commodity i (US$/MT). 31 The production data and prices were derived from the FAO statistical database (FAOSTAT), for 2004 to 2006. An average value for these years was used to reduce outliers and large annual fluctuations. Table 3 shows the total production, average price and value of production for main agricultural commodities in Kenya (Van de Steeg et al, 2009). Maize and tea are the most important crops in terms of VOP, contributing up to respectively 17% and 15%. Both milk and meat from cattle contribute to 28.3% to the agricultural VOP. The meat comes mostly from extensive cattle production in pastoral systems and most of the sold milk from semi- intensive mixed systems. Milk it also an essential source of nutrition in the more subsistence- based pastoral systems where it is mostly consumed by the family directly. Other important crops are potatoes, sugarcane and coffee. Table 2. Kenya?the total production, average price and value of production for main agricultural commodities, average values for years 2004 to 2006. Commodity Production (t) Price (US$ / t) VOP (US$) Contribution (%) 1 Milk 2,993,300 221 662,237,692 18.4 2 Maize 2,919,966 203 592,373,502 16.5 3 Tea 321,227 1729 555,412,685 15.4 4 Beef 374,217 948 354,845,973 9.9 5 Potatoes 949,453 369 350,613,881 9.8 6 Sugarcane 4,798,218 25 121,810,761 3.4 7 Coffee 47,310 2365 111,908,336 3.1 Figure 10 shows the evolution of production of the main agricultural commodities over the last 4 decades. The figure shows that maize and milk have dominated the increases in production over time. This has happened as a result of intensifying practices in dairy systems and a conducive policy environment for the production of milk, mostly in highland regions (Staal et al. 2003). Coffee areas have not increased due to market and price instability for this crop. Maize production has increased mostly as a result of area expansion and increased adoption of this crop in both mixed and pastoral areas (Herrero et al. 2010a). Figure 11 shows that overall, technological change and yield increases for the main commodities produced in Kenya has been slow. As with maize, in most areas, agricultural expansion has been the main means to increase production. This is a reflection of the lack of support for agricultural 32 production in the last decades and the lack of inputs, services and market environment to support the intensification of crop production systems. 0 500 1000 1500 2000 2500 3000 3500 1961 1971 1981 1991 2001 Y e a r s P r o d u c t i o n ( T h o u s a n d T o n n e s ) M i l k M a i z e T e a B e e f P o t a t o e s C o f f e e S w e e t p o t a t o e s W h e a t C a ssa ve Figure 10. The production of main agricultural commodities in Kenya over time (FAOSTAT, 2009). 0 20000 40000 60000 80000 100000 120000 140000 1961 1971 1981 1991 2001 Y e a r s Y i e l d ( H g / H a ) M a i z e T e a P o t a t o e s C o f f e e S w e e t p o t a t o e s W h e a t C a ssa va Figure 11. The yield of main agricultural commodities in Kenya over time (FAOSTAT, 2009). 33 Agro-ecological zones of Kenya The potential for agricultural production is determined by physical factors, primarily by soil and climatic conditions, and a complex interaction of socioeconomic, cultural and technological factors, such as farm sizes, level of farming and livestock inputs, management practices including soil conservation and enhancement, veterinary services, economic factors like market prices and access, credit availability, education and extension services (FAO, 1978-81). The climatic resource inventory of Kenya records both temperature and soil moisture conditions. This inventory was carried out as part of the Exploratory Soil Map of Kenya (KSS, 1982), at a scale of 1:1 million. Quantification of moisture conditions was achieved through the concept of reference length of growing period (LGP). The moisture availability zones is divided into 7 classes (Table 4). The quantification of temperature attributes was achieved by defining reference temperature zones. To cater for differences in temperature adaptability of crops, pasture and fuelwood species, nine thermal zones were distinguished (Table 4). The agro-climate zone map resulted into a map with more than 300 mapping units and 40 different combinations of moisture availability and thermal zones classes (Figure 13). 34 Table 3. Descriptions of different moisture availability and temperate zones, used for the agro-climate zone map of Kenya (KSS, 1982). Figure 12. The agro-climate zone map of Kenya (KSS, 1982). 35 The Food and Agriculture Organization of the United Nations (FAO), with the collaboration of the International Institute for Applied Systems Analysis (IIASA), developed a land resources database and a methodological framework to assess food production and population supporting potentials in developing countries, FAO (1971-81, 1976, 1978-80). In the nineties, FAO undertook an AEZ case study of Kenya, with the concurrence of the Kenyan Government and IIASA's participation (FAO/IIASA, 1994). The AEZ within this methodology are mainly based on the Length of Growing Period (LGP). Based on a similar approach, but with more recent data, ILRI derived an LGP map for Kenya (Thornton et al. 2002) that resembled the main agro-ecological zones of Kenya. This map is used in subsequent analyses for determining the magnitude of the expected climate change impacts on agricultural production. The LGP was divided into four classes to resemble the key agro-ecological zones (Figure 12): LGP <90 days: Arid zone 90 ? LGP ? 180 days: Semi-arid zone: 180 ? LGP ? 210 days: Sub-Humid zone LGP > 210 days: Humid zone Note that especially the sub-humid and humid zones encompass a mixture of highland and lowland areas. 36 Figure 13. The agro-climate zone map of Kenya, based on LGP classes. Most people in Kenya live in the humid and sub-humid areas (central highlands, humid lowlands around lake Victoria, coastal zones) (Table 15a) which have a higher potential agricultural productivity, are closer to larger cities and their services, markets and other infrastructure (health centres, schools, etc). These are the areas where most crop production occurs in Kenya (Table 15b). While the area under this agro-ecological zone is small relative to the drylands (arid and semi-arid areas), they have large concentrations of cattle, mostly in mixed systems with some degree of dairying and a significant number of free-ranging sheep and goats. In the past decades, significant human migrations to humid and sub-humid areas has created a significant pressure on natural resources, notably land and soils. For example, farm sizes in places have reduced to the point where farming is no longer viable as a sole activity to support families (Waithaka et al. 2006). In these areas soil fertility problems and land degradation have also been notorious to the point where crops no longer respond to fertiliser applications due to lack of organic matter in some cases (Tittonell et al. 2009). 37 Table 5a. Human population and livestock numbers in different agro-ecological zones of Kenya Human Population and Livestock numbers Agro-climatic zone Human pop('000s) Cattle Goat Sheep Area (Km2) Arid (LGP <90 days) 2,516 2,665,750 4,005,340 2,882,090 351,347 Semi-Arid (LGP 90 - 120 days) 4,377 2,751,580 2,500,690 1,969,670 123,436 Sub-humid (LGP 180 - 210 days) 2,808 1,129,430 722,522 772,678 32,203 Humid (LGP > 210 days) 20,373 7,210,830 2,263,680 2,803,410 83,490 * tropical livestock units (1 TLU = 250 kg bodyweight) In contrast, the vast arid and semi-arid districts are home to about 15% of the population, 40% of the cattle and 60% of the small ruminants of the country (Table 15a). These areas produce most of the dryland crops (sorghum and millet) of the country (Table 15b). These largely neglected areas are characterised by a high degree of poverty and food insecurity, increased conflicts, high rainfall variability and significant production risk, all of which have lead to significant human migrations to cities in higher potential areas (Nairobi, Nakuru, Kisumu, etc) in search of employment. Interestingly, the semi-arid areas present the highest yield gaps for crops, which suggests that with adequate programmes to support agriculture, investment in infrastructure and market development, and adoption of risk management practices, these areas could significantly increase crop and livestock production (Herrero et al. 2010). Table 5b. Production of key agricultural commodities by agro-ecological zone Production in Metric tonnes (?000s), 2000 Agro-ecological zone Cassava Cof fee M aize Pot atoe s Su ga rca ne Swe et pot ato Wh eat Sor gh u m M illet Arid (LGP <90 days) 20 11 171 95 1,441 16 3 29 8 Semi-Arid (LGP 90 - 120 days) 111 45 703 222 1,096 69 4 35 15 Sub-humid (LGP 180 - 210 days) 83 13 292 74 424 29 5 17 4 Humid (LGP > 210 days) 267 25 1,036 223 1,299 421 165 43 21 Total 480 95 2,203 614 4,260 536 177 126 47 38 Maps of the spatial distribution of human population, crop production and livestock densities by agro-ecological zone of Kenya can be found in Appendix A. Understanding climate change impacts on crop and livestock production ? the length of growing period Like Fischer et al. (2002) and Jones and Thornton (2003), we assessed the impact of climate change on agro-ecological characteristics by looking at changes in Length of Growing Period (LGP), as an initial proxy for agricultural impacts. Changes in rainfall patterns, in addition to shifts in thermal regimes, influence local seasonal and annual water balances and in turn affect the distribution of periods during which temperature and moisture conditions permit agricultural crop production. Such characteristics are well reflected by the LGP since Kenya mostly relies on rainfed agriculture (Fischer et al., 2002; Comprehensive Assessment, 2007). The use of this indicator supplemented with crop modelling work provides a framework provides a framework for studying the impacts of climate change on crop yields and production. LGP was calculated as described by Thornton et al. (2006). In this study, for each 10-minute pixel in Kenya climate normals data, monthly values for average daily temperature (?C), average daily diurnal temperature variation (?C), and average monthly rainfall (mm), were read from the appropriate gridded file and interpolated to daily data using the method of Jones (1987). LGP is actually the total number of days in a year when there is enough water to support crop growth. It does not deal well with bimodal rainfall regimes when the two seasons are actually interspersed with a dry period. However, bimodal rainfall patters are less pronounced in Kenya than in the past (Thornton et la 2006). Figure 14 shows the LGP for 2000. 39 Figure 14. The length of growing period (in days) for 2000 (Thornton et al., 2006). Thornton et al. (2006) presented LGP changes for the whole of Africa to 2050 under various model projections, showing few differences in projections under two SRES scenarios (A1F1 and B1). The ?A? scenarios place more emphasis on economic growth, the ?B? scenarios on environmental protection. The ?1? scenarios assume more globalization. For this part of the report revised spatial data layers are utilized (Thornton and Jones, 2008). LGP changes to 2030 and 2050 are projected for Africa using downscaled outputs of coarse-gridded GCM, using methods outlined in Jones and Thornton (2003), using the datasets of WorldCLIM (Hijmans et al., 2005), TYN SC 2.0 dataset (Mitchell et al., 2004), and the outputs from the Hadley Centre Coupled Model version 3 (HadCM3) (Mitchell et al., 1998) and ECHam4 (Roeckner et al., 1996), associated with A1FI and B1 (IPCC, 2001). Figure 14 shows maps of projected changes in LGP from 2000 to 2030 and 2050, from downscaled outputs of the ECHam4 and the HadCM3 GCM for scenarios A1F1 and B1. Following IPCC (2001) map legends, these changes were classified into five: losses in LGP of >20% (?large? losses); of 5? 20% (?moderate? losses); no change (? 5% change); gains of 5?20% (?moderate? gains); and gains of >20% (?large? gains). As discussed by Thornton et al. (2006), various points can be made about these maps. First, it should be noted that some of the large losses and large gains are located in areas with LGP less than 60 days (arid agro-ecozone), i.e. in highly marginal areas for cropping but important 40 for pastoralists. This implies that pastoralism will continue to be a significant livelihood option in these regions vis-?-vis crop expansion in marginal lands under current circumstances but that there is a need to support them with mechanisms to deal with potentially greater variability (risk reduction, insurance based schemes, development of safety nets, etc). Second, there is considerable variability in results arising from the different scenarios, and there is also variability in results arising from the different GCMs used. Third, if anything could be generalized about these different maps, it is that under the range of these SRES scenarios and the GCMs used, many parts of Kenya are likely to experience a decrease in LGP, and in some areas, the decreases may be severe. In other words, projected increases in temperature and projected changes in rainfall patterns and amount (increases in rainfall amounts are projected in many areas) combine to suggest that growing periods will decrease in many places. There are also a few areas, especially in the highlands (humid and sub-humid zones) where the combination of increased temperatures and rainfall changes may lead to an extension of the growing season. Figure 15. The percentage change in length in growing period to 2030 and 2050 in Kenya (Thornton et al., 2006). Differences in projected changes (Figure 13) make it quite challenging to come to a general consensus over climate change trends for Kenya, or certain areas within Kenya. Although the projected increases in rainfall might appear to be good news for arid and semi-arid districts, the increasing temperatures cause a substantial increase evaporation rates, which are likely to 41 balance and exceed any benefit from the increase in precipitation (Osbar and Viner, 2006). This means that the increases in LGP might only translate into very modest, if at all, increases in rangeland or crop productivity in these areas. To elicit the responses of different crops to these changes, the next sections delve further by using crop simulation models to determine plausible impacts of claimte change on agricultural production. Impacts of climate change on crop production The impacts of climate change on crop production for Kenya were studied using the methods described by Rosegrant et al (2009). In summary, statistically downscaled climate data was obtained from the NCAR (NCAR-CCSM3) and CSIRO models under the A2 scenario from the IPCC 4th Assessment report; and also from United Kingdom Meteorological Office Hadley Centre?s Coupled Model, version 3 (HadCM3), using the A2a scenario from IPCC?s Third Assessment Report. Data were used to run the DSSAT suite of crop models for 4 key staple crops (maize, wheat, groundnuts (as a proxy for beans), and irrigated rice to 2050. A single crop variety was chosen for each crop, but management practices were distiguished on the basis of regionally differentiated agricultural practices ands fertiliser inputs. The crop area extent for each crop was determined using the crop layers of You and Wood (2004). Results were summarised for each agro-ecozone and displayed in maps accordingly. Since the differences observed in model X scenario combinations is important, we present individual maps for the different combinations in Figure 14a to aid the interpretation of the responses. All SRES scenarios have higher temperatures in 2050 resulting in higher evaporation of water. When this water vapor eventually returns to the earth as precipitation, it can fall either on land or the oceans. The NCAR model is ?wet? in the sense that average precipitation on land increases by about 7 percent. For Kenya, the NCAR scenario predicts a 45 percent increase in annual rainfall from 2000 to 2050. The CSIRO model (like the Hadley model) envisages a much drier future climate. Globally, rainfall is projected to increases by about 1 percent under the CSIRO scenario from 2000 to 2050 and to decline by 0.4 percent for the HadCM3 scenario. For Kenya, the CSIRO annual rainfall change is an increase by 5 percent and for the HadCM3 scenario an increase of 0.2 percent. Proyected impacts to 2050 results in lower rainfed maize yield for Kenya in 4 out of 6 scenarios. Maize yields are likely to vary modestly according to the CSIRO and NCAR 42 models (+/- 5 to 20 %). In general terms, most modest changes are observed in the humid and sub-humid areas and small gains can be observed in certain highland areas in the humid and sub-humid zone. These results seem to be in line with the data of Thornton et al. (2009). The arid and semi-arid regions show in most cases variable reductions in yields, with the Hadley model estimating the largest decreases (up to 50% in some parts). In four out of six scenarios (NCAR and CSIRO models), large decreases in yields (between 20-50%) are observed for groundnuts and wheat irrespective of agro-ecozone, The DSSAT runs with the HadCM3 downscaled climate data show different effects by agroecozone: potential gains in yields for groundnuts of 5-20% in the humid and sub-humid areas but potential reductions of 20-50% in the arid and semi-arid areas. With the current varieties, these levels of reduction may force farmers to rethink the feasibility of planting this crop in these areas. A similar but reverse story is observers for wheat: modest gains of 5-20% are observed in semi-arid areas, while large losses are predominantly in the humid and sub-humid zones. A similar case is observed for irrigated rice. The variability in the results for groundnut and wheat, for example, shows the difficulties in making generalisations about the impacts of climate change on particular crops in particular places. Current best practice in climate change and crop modelling dictates the use of as many model X scenario combinations to try to reduce the uncertainty in the magnitude of the impacts, the locations and the direction of change in yields, whether they are increasing or decreasing. 43 ? Figure 15a. Climate change impacts on yields of key commodities in Kenya to 2050 as projected by 6 different model X scenario combinations 44 Thornton et al. (2010) analysed the spatial differences in simulated main-season maize and secondary-season Phaseolus bean yields to 2050, and attempt some simple characterisation of crop response. GCMs show an heterogeneous response of crop yield to the changing amounts and patterns of rainfall, and to the generally increasing temperature As shown by Thornton et al. (2010) they may vary by crop type, by location, and through time. Results also indicate that under the four GCM?scenario combinations considered, the aggregate production decreases are projected to be rather modest to 2050. These aggregate production changes, however, hide a large amount of variability, as shown in figure 14a. Several studies indicate the uncertainty of crop models in the response of yield to climate change is comparable in magnitude to the mean simulated yield change (Challinor and Wheeler, 2007; Thornton et al., 2010). However, the results suggest that we need to keep on monitoring the effects of climate change on crop yields. Thornton et al. (2009) demonstrated that maize yields to 2050 are reduced by 20% for the more semiarid areas of Kenya and Tanzania where maize cropping is possible. Most of these losses are in the range 200?700 kg/ha. Production losses of maize to 2050 could be in the order of 8.4% in the mixed rainfed systems in the arid and semiarid areas and 9.8% in the mixed rainfed systems in the humid and semihumid areas of Kenya. By contrast, maize yields are projected to increase in the central and western highlands of Kenya, mostly by between 200 and 700 kg/ha. To 2050 the production of maize is likely to increase by 46.5% in the mixed rainfed systems in the temperate areas of Kenya (Thornton et al, 2010), but total country production will still decline as these areas contribute modestly to the total country production. Scarce information is available on the impacts of climate change on cash crops in Kenya. According to maps provided by UNEP-GRID, a 2-degree Celsius increase in temperature would make much of the current tea area in Kenya unsuitable to tea production, in particular, the tea areas in the Mount Kenya, Aberdares, and Kisumu area (Simonett 1989). In the short term, recent declines in tea production have been directly linked to erratic rainfall patterns and drought. 45 Wider Effects on the Economy Climate change impacts, in the form of yield declines, may be less severe in sub-Saharan Africa than in Asia in terms of reductions in yield. This is primarily because yields in Sub- Saharan-Africa are much lower and their absolute reductions therefore smaller than in Asia. However, sub-Saharan Africa is one of the most vulnerable regions to climate change as a result of its low adaptive capacity (Thornton et al., 2009), linked to high levels of poverty and poor infrastructure, as reflected in the high dependence on rainfed agriculture. Sub-Saharan Africa faces increased net food imports even under historic climate as a result of growing populations, faster economic growth compared to the past, and growing urbanization, coupled with slow improvement in agricultural productivity. According to Nelson et al. (2009), climate change will likely further increase net food import demand in the region. Thus, Kenya will not only be affected by local impacts, but also by climate change impacts in other countries. If climate change impacts are larger in other countries than in Kenya, food imports that might otherwise have been available for sub-Saharan Africa in general and Kenya in particular might be redirected to those countries and regions experiencing even sharper declines in food production as a result of climate change. Kenyan agricultural development strategies need to take into account food price and trading environments under climate change in their assessment of climate change impacts and for the development of appropriate adaptation strategies. To assess these issues for Kenya, we are using an integrated modeling framework. Results on yield changes of different commodities and their spatial distribution are taken from the previous section on crop modelling and used as inputs into a partial equilibrium model of the agricultural sector called IMPACT (International Model for Policy Analysis of Agricultural Commodities and Trade). Model details are presented in Appendix 3. The flow of the analysis is presented in figure 15. 46 Figure 16. Models used and flow of the analysis of the impacts of climate change on crop yields and the wider impacts on the economy of Kenya. Biophysical climate change effects on crop productivity enter into the IMPACT model by affecting both crop area and yield. IMPACT integrates impacts on crop production from altered temperature and precipitation patterns, changes in irrigation water availability and evapotranspiration potential; it also includes the effects of technological change over time, and economic feedback effects through changes in international food prices, which lead to a series of (autonomous) supply and demand responses. Thus, three impacts on crop production from climate change are considered: first, direct effects on rainfed yields through changes in temperature and precipitation; second, indirect effects on irrigated yields from changes in temperature and changes in water availability for irrigation (including from precipitation); and third, autonomous adjustments to area and yield due to price effects and changes in trade flows in the economic model. With comparisons of IMPACT projections with and without climate change scenarios, the ?net? impacts of climate change on agricultural production, demand, trade and prices can be obtained. Downscaled climate data from different GCM/scenarios DSSAT crop growth models IMPACT partial equilibrium model Impacts on: Commodity prices, trade, kilo-calorie availability, number of malnourished children by GCM/scenario combination tested Daily climate data Crop yields 47 World prices are a key indicator of food affordability and security and also of the effects of climate change on agriculture . Table 7 shows the price effects under the three scenarios for 2025 and 2050. Climate change will increase world prices of cereals, grains, and meats compared to a scenario with historic climate. Adverse impacts on food prices are even higher for some crops if the carbon fertilization effect is included, with the exception for rice, soybean and sweet potato for 2025 and rice and soybean for 2050. The carbon fertilization effect has much lower benefits for the African continent as few crops receive adequate fertilization. By 2025, maize prices increase most under the NCAR 369 A2 scenario, followed by the Hadley scenario; by 2050, maize prices are similarly highest under NCAR 369 A2 scenario, followed by the CSIRO 532 A2 scenario. Price increases are somewhat lower for meat and dairy products; however, this analysis does not incorporate the impact of climate change on grazing lands and pastures, nor animal heat stress. If these impacts were included, price effects for these commodities would likely be larger. Table 6. Agricultural commodity prices, alternative climate change scenarios (US$/mt) and percentage change under alternative climate scenarios 2005 Year 2025 Year 2050 No climate change Hadley 369 NCAR369 CSIRO 369 CSIRO 532 No climate change Hadley 369 NCAR 369 CSIRO 369 CSIRO 532 US$/mt (%) US$/mt (%) Beef 2,146 2,336 -9 2 -14 1 2,836 -29 9 -25 7 Pigmeat 911 1,033 6 3 -10 2 1,272 -15 15 -20 13 Sheep and Goat 2,996 3,100 -11 0 -25 0 3,275 -39 6 -39 5 Poultry 1,191 1,396 -7 4 -13 3 1,688 -20 17 -20 14 Rice 211 255 17 19 10 7 310 26 36 12 11 Wheat 134 144 33 48 28 33 162 48 106 43 66 Maize 102 124 27 29 16 23 155 27 52 14 35 Millet 310 324 11 52 -6 13 281 21 22 -30 22 Sorghum 121 144 14 230 8 19 146 11 41 -5 32 Soybeans 214 306 10 7 2 0 347 13 14 6 0 Groundnuts 501 529 20 -67 17 19 487 20 52 16 35 Other Grains 88 88 39 57 24 44 83 43 123 17 84 Potatoes 226 188 38 58 49 50 158 56 118 90 101 Sweet Potatoes 549 567 2 46 38 38 624 -7 94 50 64 Cassava & other 69 71 15 42 23 27 68 16 97 41 56 roots and tubers 48 Climate change affects the agriculture sector directly and indirectly through impacts on crop productivity and production, and resulting shocks on the economic system, and alteration of prices, which in turn affect food demand, calorie availability and, ultimately, human well- being. Figure 16 presents the aggregated changes in maize yield under alternative climate change scenarios taking into account a) technological change through time, b) autonomous adaptation as a result of higher food prices and thus dampened demand and pressure on increasing supplies; and c) some balance between supply and demand as a result f changes in trade-flows. In Kenya, compared to 2000 rainfed maize yields of 1.6 metric tons per ha, yields by 2050 without climate change are projected at 2.4 metric tons per ha, at an annual yield growth rate of 0.86 percent, compared to historic overall maize yield growth (rainfed and irrigated) of 0.73 percent per year from 1962 to 2006 (three-year centered moving average). This exogenous technological change is justified by a long history of crop improvements over time as a result of agricultural research (new varieities) and enhanced crop inputs; globally crop yields are expected to improve by 1% per year. Under climate change, yields change dramatically for Kenya, whereas average changes in sub-Saharan Africa are much smaller, as adverse impacts in parts of sub-Saharan Africa are compensated by beneficial impacts elsewhere in the region. Compared to a situation with historic climate, Kenyan maize yields drop by 51-55 percent under the NCAR 369, CSIRO 369, and CSIRO 532 A2 scenarios, compared to 2050 yields with historic climate. On the other hand, yields increase by 25 percent under the Hadley 369 A2a scenario. These results somehow differ from the yield changes from other studies (i.e. Cline 2007). The main reason is that the yield changes presented in Figure 16 come from a detailed integration of spatially explicit bio-physical modelling with projected technological change, supply/demand aspects and trade. Therefore they represent more than just biophysical impacts. 49 0 500 1000 1500 2000 2500 3000 3500 2000 2 0 5 0 n o C C 2050 N C A R 3 6 9 2050 C S I R O 5 3 2 2050 C SI R O 3 6 9 2050 H a d l e y 3 6 9 SSA K e n y a Figure 17. Maize yield, historic climate and alternative climate change scenarios (kg/ha) Research on the effects of climate change on world agricultural markets is still relatively limited. Crop and animal production are affected both by changes in temperature and precipitation. Climate change alters comparative advantage, setting up the possibility of changes in trade flows as producers respond to changing opportunities. More generally, agricultural trade flows depend on the interaction between inherent comparative advantage in agriculture, which is determined by climate and the resource endowments, and a wide-ranging set of local, regional, national and international trade policies. As with any change in comparative advantage, unfettered international trade allows comparative advantage to be exploited to the fullest. Figure 17 presents changes in net cereal trade and net maize trade for Kenya. As expected, net imports increase under historic climate for both maize and all cereals. Maize imports are expected to almost quadruple, from 663,000 metric tons to 2,404,000 metric tons; and total cereal imports are projected to increase from 1.5 to 3.2 million metric tons. Under climate change, maize and total cereal imports would be much higher for two out of the three scenarios examined, by between 21 and 44 percent, thus increasing the future dependency and vulnerability of local food systems under climate change. Under the Hadley scenario, on the other hand, maize imports would be 63 percent below the scenario without climate change. Trade flows are changing as a result of changing comparative advantage of locations of food production and demand; as mentioned earlier trade flows have been increasing gradually and continually over time with small drop offs as a result of recessions/depressions, but the trend is clearly upward. Under climate change scenarios, the majority of additional supply is produced in the group of the developed countries, chiefly North America and Europe, but also Latin America?with variations across 50 scenarios. This is not necessarily a win-win situation. Trade will be able to buffer deficits, but at a social cost as prices rise and the poor are excluded from the benefits of consuming the available food due to their low incomes. Higher food prices due to increased trade, dampen demand for food as affordability of nearly all agricultural commodities?including basic staples and livestock products- declines under climate change. As a result, per capita calorie availability in Kenya declines under all climate change scenarios, even the Hadley scenario that postulates significant yield growth for Kenya. In 2000, average per capita calorie availability for Kenya was estimated at 2,186 calories per day, just slightly above the minimum 2,000 calories per capita per day that are considered necessary to lead a healthy and productive life. By 2050, little improvement is expected in calorie availability for the country, with availability estimated to increase to 2,295 calories per capita per day without climate change. Under climate change, on the other hand, calorie availability would decline by -2 to -19 percent (Figure 18). Under both the CSIRO 532 A2 and the NCAR 369 A2 scenarios, calorie availability would fall below the 2,000 kilocalorie threshold, and only the CSIRO 369 A2 scenario is above the calorie availability level achieved in the year 2000. - 5000 - 4500 - 4000 - 3500 - 3000 - 2500 - 2000 - 1500 - 1000 - 500 0 2000 2 0 5 0 N O C C 2 0 5 0 C SI R O 3 6 9 2 0 5 0 C SI R O 5 3 2 2 0 5 0 N C A R 3 6 9 2 0 5 0 Ha d l e y3 6 9 M a i z e A l l Ce r e a l s Figure 18. Change in net cereal and maize trade, Kenya, alternative climate change scenarios (thousand metric tons). Negative numbers indicate net imports. Source: IFPRI Impact Simulations (2009). 51 0 500 1000 1500 2000 2500 2000 2 0 5 0 N O CC 2050 C S I R O 3 6 9 2050 C S I R O 5 3 2 2050 N C A R 3 6 9 2050 H a d l e y 3 6 9 Figure 19. Per capita calorie availability per day, alternative climate change scenarios, Kenya. Source: IFPRI Impact Simulations (2009). Food and nutrition security are closely tied to agricultural productivity. Increased food production increases local food availability. Higher production from one?s own farm or herds increases access to food and enhances household food security. The nutritional quality of the food produced is also an important consideration in reducing malnutrition, particularly for households who acquire most of their food from their own fields and herds. Particularly in sub-Saharan Africa, the most potent force for reducing malnutrition is raising food availability through increased agricultural productivity, as well as trade. Key non-food determinants of child malnutrition include the quality of maternal and child care, female secondary education, and health and sanitation (Smith and Haddad, 2000). Depressed food demand translates into direct increases in malnutrition levels, with often irreversible consequences for young children. Projections show that climate change increases the share of malnourished children in both 2025 and 2050, compared to a non-climate change scenario (Figure 19). Without climate change, the share of malnourished children is projected to decline from 19 percent in 2000 to 15 percent by 2025 and 11 percent by 2050. Thus, Kenya?s child malnutrition levels are significantly below the average in Sub-Saharan Africa in 2000 (28 percent) and projected in 2025 (29 percent). Under climate change, child 52 malnutrition levels increase under all alternative climate change scenarios, with levels raising highest under the NCAR 369 A2 scenario, and lowest under the CSIRO 532 A2 scenario. 0 2 4 6 8 10 12 14 16 18 20 2000 2 0 2 5 n o C C 2 0 2 5 C S I R O 369 2 0 2 5 N C A R 369 2 0 2 5 H a d l e y 369 Figure 20. Kenya: share of malnourished children, historic climate and alternative climate change scenarios, 2025 (percentage). Source: IFPRI (2009). 0 2 4 6 8 10 12 14 16 18 2 0 5 0 n o C C 2050 C S I R O 3 6 9 2050 N C A R 3 6 9 2050 C S I R O 5 3 2 2050 H a d l e y 3 6 9 Figure 21. Kenya: share of malnourished children, historic climate and alternative climate change scenarios, 2050 (percentage). Source: IFPRI Impact Simulations (2009). 53 5 Variability, vulnerability and livelihoods As has been noted already, currently there is little that can be said concerning the details of the increases in climate variability that, it is envisaged, will affect East Africa (indeed all places) during this century. This section contains two brief examples of some of the impacts that increased climate variability may bring about in livestock systems: one looks at possible impacts of increasing climate variability on herd structure, and the other at possible shifts in livehoods that may be induced by changes in climate and climate variability. Impact of increased climate variability on livestock assets of pastoralists (This subsection was adapted from Thornton and Herrero, 2010, with additional analyses). In general, pastoralists live in regions where the impacts of climate change are likely to be large (Thornton et al., 2006), including the Sahelian rangelands, southern Africa, and parts of East Africa. These are some of the most vulnerable livestock keepers on the planet. Livestock provide many benefits to pastoral families in the form of milk, meat, hides, manure, and socio-cultural capital. At the same time they represent a considerable asset that can be traded or sold in hard times or for purposes such as paying school fees or providing a dowry (Nkedianye et al., 2009). The impact of drought on herd performance and asset values has been widely documented. In large areas of Africa, highly variable climate with frequent droughts can decimate herds and displace pastoralists. Emergency services and humanitarian relief efforts are often needed to support pastoralist families during considerable parts of the year in these regions. We ran a herd dynamics model (Lesnoff, 2007) to investigate the potential impacts of increased climate variability, represented here as increased drought frequencies, on herd dynamics and livestock numbers. We used baseline information on mortality, reproduction and herd structures from pastoralist herds in Kajiado, Kenya (Boone et al., 2005). The model was run over 20 years assuming a herd baseline size of 200 animals, of which 60 where adult females. We ran two scenarios: a baseline scenario simulating realistic climate variability of one drought every five years and an alternative scenario of increased frequency of droughts ? one year in three. Such increases in climate variability may be anticipated as a result of global warming. In years of drought, animal mortality rates increase and reproductive performance 54 of adult females declines, potentially resulting in lower numbers of offspring and a declining herd size. Results indicate that a drought once every five years (i.e., representative of current conditions) keeps herd sizes stable (Figure 20), and this has in fact been observed in Kajiado for a long time (Rutten, 1992). At the same time, the district has seen substantial increases in human population, meaning that the proportion of the population that can thrive in a pastoral setting has plummeted, because animal numbers per adult equivalent are simply not sufficiently high to support pastoralism. This might reflect that the ecosystem simply cannot support more animals (except at the possible expense of wildlife, with other income-related effects). When we increased the probability of drought to once every three years, herd sizes decreased as a result of increased mortality and poorer reproductive performance (see Figure 21). This decrease in animal numbers would affect food security and would compromise the sole dependence of pastoralists on livestock and their products, as well as the additional benefits they confer. This simple analysis shows that under increased climate variability, the need for diversification of income, a strategy often (and increasingly) observed in pastoral areas, becomes ever-more important. Climate change and increasingly climate variability will have substantial impacts on environmental security as well, as the conflicts (usually over livestock assets) often observed in these regions are likely to escalate in the future (Bocchi et al., 2006). Figure 22. Evolution of total herd size and the number of adult females under two scenarios of climate variability: (1) a drought every five years, and (2) a drought every three years. 55 We upscaled the results to the ASAL regions in Kenya and estimated that 1.8 million animals would be lost to 2030 due to increased drought frequencies (Table 8). In terms of economic losses, the biggest losses are in terms of the livestock assets, as in these regions milk and meat production are low. This is important as livestock accumulation represents an important risk management strategy for pastoral societies as animals can be sold in times of hardship. They also play an essential cultural role (prestige, dowry) or for paying school fees, food purchases, etc. It is essential to increase the resilience and adaptation of agro-pastoralists to protect their livelihoods if these kinds of extreme events increase in frequency (Herrero et al. 2010b). This can be done in many ways. Some examples are: by a) implementing schemes to protect their assets such as index-based insurance schemes, or the development of easy to implement early warning systems, b) creating incentives to incorporate pastoralists into the market economy to generate cash income. This would imply investing in market and value chain development to enhance the ability to obtain inputs and sell livestock products. c) develop safety nets so that disenfranchised people can access food, health services and others in times of hardship. This would involve the development of food storage systems, improving water accessibility and developing institutional networks to support pastoralists (government, civil societies, NGOs, others). Table 7. Economic impacts of increased drought frequencies in pastoral and agropastoral systems in arid and semi-arid regions of Kenya Indicator value Cattle numbers 2000 (million TLU)1 Cattle numbers 2030 drought 1 in 5 years (million TLU) 1 Cattle numbers 2030 drought 1 in 3 years (million TLU) 2 Animals lost due to increased drought freq. (million TLU) 2 Cumulative milk production lost (million kg) 3 Cumulative meat production lost (million kg) 4 Value of lost animals (million $) Value of lost milk production (million $)3 Value of lost meat production (million $)4 Total economic losses (million $) 5.6 5.9 4.1 1.8 837 1.4 458 167 5 630 1 Data from Herrero et al. (2008) 2Estimated with the model of Lesnoff (2007) 3Assumptions: price of 1 animal $250, milk production 150 kg/yr, 20% females in milk, milk price 20 Ksh 4Assumptions: 10% offtake, 50% dressing percentage, meat price 250 Ksh / kg 56 Croppers to livestock keepers: Possible livelihood transitions due to climate change (This subsection is based on Jones and Thornton, 2009). Various studies estimate that warming and drying may reduce crop yields by 10 to 20% overall to the middle of the century, and increasing frequencies of heat stress, drought and flooding events will result in yet further impacts on crop and livestock productivity. The local effects of climate change may be severe in places, to the point where the existing livelihood strategies of rural people may be seriously compromised. These places are likely to include parts of East Africa that are already marginal for crop production; as these become increasingly marginal, through a combination of increasing temperatures, changing rainfall amounts and patterns, and increasing climate variability, then livestock may provide an alternative to cropping. Some of these areas in sub-Saharan Africa have been identified where such transitions might occur. For the currently cropped areas of the continent, a recent study estimated probabilities of failed seasons for current climate conditions, and compared these with estimates obtained for future climate conditions in 2050, using downscaled climate model output for two contrasting greenhouse-gas emission scenarios. Results are shown in Figure 22, in terms of the parts of the continent in the mixed crop-livestock rainfed arid? semiarid systems in which the number of Reliable Crop Growing Days (RCGD) falls below 90 between 2000 and 2050, as projected using the HadCM3 model and the A1FI high- ermissions scenario (Jones and Thornton, 2009). RCGDs are an indicator of growing season length and reliability, and are a probabilistic measure related to LGP (see above). Cropping in areas with an RCGD less than 90 becomes highly marginal, and so this value can be used as a cut-off point below which cropping is likely to be too risky for the household. Areas in red in Figure 22 are "transition zones", where cropping may be possible now but will probalby not be possible in 2050. For Kenya, these areas are relatively small. They are located around coastal areas and also in transition zones between the highlands and the lowlands. Areas like Machakos, where significant decreases in LGP may force farmers to rely more on livestock, substitute crops and/or diversify into other activities. 57 Figure 23. Transition zones in the mixed rainfed arid?semiarid system, in which the Reliable Crop Growing Days (RCGD) falls below 90 between 2000 and 2050, as projected using the HadCM32 model and the A1FI scenario (Jones and Thornton, 2009). Even under even a moderate greenhouse gas emission scenario for the coming decades, there are likely to be substantial shifts in the patterns of African cropping and livestock keeping to the middle of the century. Potential livelihood transition zones can be identified, and these zones differ in their accessibility, which may have considerable impacts on the type of adaptation options that may be viable. For those that are relatively close to large human settlements, for example, there may be options for both integration of livestock systems into the market economy and for off-farm employment opportunities. For transition zones that are more remote, on the other hand, both market and off-farm employment opportunities may be much more limited. There are currently significant populations of people in these more remote transition zones, and they are widely spread throughout West, East and southern Africa. Substantial changes may be required to people's livelihoods and agricultural systems if food security is to be improved and incomes raised. The results also highlights the fact that poverty rates in the marginal cropping lands of Africa are already high, and generally increase as accessibility decreases (Jones and Thornton, 2009). There will be an increasing need in these areas for highly-targeted schemes that promote livestock ownership and facilitate risk management where this is appropriate, as well as efforts to broaden income-generating opportunities in parts of the continent where this is feasible. 58 Conclusions Notwithstanding the uncertainty in analysing the impacts of climate change and variability on the agricultural sector in Kenya, below are a few points that summarise the main conclusions from our report. In East Africa there are very few places where rainfall means are likely to decrease. The increase in rainfall in East Africa, extending into the Horn of Africa, is robust across the ensemble of GCMs, with 18 of 21 models projecting an increase in the core of this region, east of the Great Lakes. The increases in rainfall and temperature will only translate in increased agricultural productivity in specific locations. Increases in rainfall may not lead to increases in agricultural productivity in lowland regions since increases in temperature will also increase evapotranspiration and offset any potential increase in productivity. On the other hand, increases in temperature may remove crop growth constraints in the highlands, thus potentially leading to higher yields. However, to really capitalise on the potential yield increases it will be necessary to invest in inputs and services. Even with modest increases in maize and bean production in the highlands, Kenya will experience country-wide losses in the production of key staples, due to increased evapotranspiration in large cropland areas. There is large uncertainty about the magnitudes of the country-wide staple production losses., but they may be between minus 10 and 55 % depending on the scenario, crop model and GCM run. Trade in key staples could offset lower crop production caused by climate change. Trade in cereal is likely to increase to satisfy internal consumption. Under climate change, maize and total cereal imports would be much higher for two out of the three scenarios examined, by between 21 and 44 percent. Under the Hadley scenario, on the other hand, maize imports would be 63 percent below the scenario without climate change. 59 However, the whole picture is more complex. Prices of key staples are likely to increase and this will dampen demand for food, as affordability of nearly all agricultural commodities? including basic staples and livestock products declines under climate change. As a result, per capita calorie availability in Kenya is likely to decline under all climate change scenarios Lower food accessibility due to increased commodity prices is likely to translate in increases in malnutrition, especially of young children. Climate change is likely to increase the number of malnourished children in both 2025 and 2050. Without climate change, child malnutrition levels are projected to decline from 19 percent in 2000 to 15 percent by 2025 and 11 percent by 2050. Under climate change, child malnutrition levels increase under all alternative climate change scenarios. These effects will probably be exacerbated in areas of high vulnerability, like the ASALs. Increased drought frequencies to more than a drought every five years could cause significant, irreversible decreases in livestock numbers in arid and semi-arid areas. Results indicate that a drought once every five years (i.e., representative of current conditions) keeps herd sizes stable in ASALs, and this has in fact been observed in places like Kajiado for a long time. Increased probability of drought to once every three years, could decrease herd sizes as a result of increased mortality and poorer reproductive performance of the animals. Pastoralists whose food security and entire livelihood depends solely on livestock would be severely affected by decreased animal numbers.. This highlights how under increased climate variability, diversification of income sources is a key adaptation strategy. There are some signs of livelihood diversification in pastoral areas, but it will need to be encouraged further. Climate change and increasingly climate variability will have substantial impacts on environmental security as well, as the conflicts (usually over livestock assets) often observed in these regions are likely to escalate in the future. Kenya will have significant areas in the ASALs where cropping might no longer be possible as a result of climate change and where the role of livestock as a livelihood option is likely to increase. Even under even a moderate greenhouse gas emission scenario for the coming decades, there are likely to be substantial shifts in the patterns of African cropping and livestock keeping to the middle of the century. Potential livelihood transition zones can be identified, and these zones differ in their accessibility, which may have considerable impacts on the type of adaptation options that may be viable. For transition zones that are remote, both market and off-farm employment opportunities may be limited. Substantial changes 60 may be required to people's livelihoods and agricultural systems if food security is to be improved and incomes raised. There will be an increasing need in these areas for highly- targeted schemes that promote livestock ownership and facilitate risk management where this is appropriate, as well as efforts to broaden income-generating opportunities in parts of the continent where this is feasible. Strengthening the adaptive capacity of vulnerable populations and of the agriculture sector as a whole requires a comprehensive assessment of the impacts of climate change and variability, the risks these changes pose to agricultural production, the constraints to adaptation households and communities face, and the potential policy options that can facilitate adaptation. Responses to climate change need to encompass several levels, including crop and farm-level adaptations, collective action at the community level, and agricultural and supporting policies and investments at national, regional and global levels. Adaptation will require the involvement of multiple stakeholders, including policymakers, extension agents, NGOs, researchers, communities, and farmers. Potential strategies will include infrastructural investment, water management reform, land-use policy, and food trade. The fact that the study of climate change is an uncertain discipline is no excuse for inaction. Using the best information available, the Kenyan agricultural sector, donors and other stakeholders need to be responsive and act in a timely, targeted fashion to ensure that millions of smallholders can adapt to climate change and maintain or improve their livelihoods and the ecosystems they rely upon. 61 References Anyah, R.O., Semazzi, F.H.M. 2007. Variability of East African rainfall based on multiyear RegCM3 simulations. International Journal of Climatology, 27, 357?371. Bocchi S, Disperati S P, Rossi S, 2006. Environmental security: A geographic information system analysis approach - the case of Kenya. Environmental Management 37, 186-199. Boko, M., I. Niang, A. Nyong, C. Vogel, A. Githeko, M. Medany, B. Osman-Elasha, R. Tabo and P. Yanda, 2007: Africa. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson, Eds., Cambridge University Press, Cambridge UK, 433-467. Boone R B, BurnSilver S B, Thornton P K, Worden J S, Galvin KA, 2005. Quantifying declines in livestock due to land subdivision in Kajiado District, Kenya. Rangeland Ecology and Management 58, 523-532. Challinor, A.J., Wheeler, T.R. 2007. Crop yield reduction in the tropics under climate change: Processes and uncertainties. Agricultural and Forest Meteorology, 148(3), 343-356. Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli, W.-T. Kwon, R. Laprise, V. Maga?a Rueda, L. Mearns, C.G. Men?ndez, J. R?is?nen, A. Rinke, A. Sarr and P. Whetton, 2007: Regional Climate Projections. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Comprehensive Assessment. 2007. Water for Food, Water for Life: A Comprehensive Assessment of Water Management in Agriculture. Earthscan, London and IWMI, Colombo. Conway, D., Hanson, C.E., Doherty, R., Persechino, A. 2007. GCM simulations of the Indian Ocean dipole influence on East African rainfall: Present and future. Geophysical Research Letters, 34(3), L03705.1-L03705.6. Cooper, P.J.M., Dimes, J., Rao, K.P.C., Shapiro, B., Shiferaw, B., Twomlow, S. 2008. Coping better with current climatic variability in the rain-fed farming systems of sub-Saharan 62 Africa: An essential first step in adapting to future climate change? Agriculture, Ecosystems and Environment, 126, 24?35. Doherty, R.M., Sitch, S., Smith, B., Lewis, S.L., Thornton, P.K., 2009. Implications of future climate and atmospheric CO2 content for regional biogeochemistry, biogeography and ecosystem services across East Africa. Global Change Biology, doi:10.1111/j.1365- 2486.2009.01997.x. Fischer, G., Shah, M., and Velthuizen, H., 2002. Climate change and agricultural vulnerability. International Institute for Applied Systems Analysis (IIASA). http://www.eldis.org/static/DOC10242.htm FAO, 1978-81. Report on the Agro-ecological Zones Project. Vol.1, Methodology and results for Africa; Vol.2, Results for Southwest Asia; Vol.3, Methodology and results for South and Central America; Vol.4, Results for Southeast Asia. [FAO] World Soil Resources Report 48/1,-4. FAO, 1976. A Framework for Land Evaluation. FAO Soils Bulletin, 32. FAO, 1978-81. Report on the Agro-ecological Zones Project. Vol.1, Methodology and results for Africa; Vol.2, Results for Southwest Asia; Vol.3, Methodology and results for South and Central America; Vol.4, Results for Southeast Asia. [FAO] World Soil Resources Report 48/1,-4. Fischer, G.W. and J. Antoine, 1994. Agro-ecological land resources assessment for agricultural development planning: A case study of Kenya: Making land use choices for district planning. Food and Agriculture Organization of the United Nations and International Institute for Applied Systems Analysis, Rome. Funk, C., Dettinger, M.D., Michaelsen, J.C., Verdin, J.P., Brown, M.E., Barlow, M., Hoell, A., 2008. Warming of the Indian Ocean threatens eastern and southern African food security but could be mitigated by agricultural development. PNAS 105, 11081?11086. Freeman, A., Notenbaert, A., Herrero, M., Thornton, P.K. and Wood, S. 2008. Strategies and priorities for integrated agriculture for development in the SSA CP. Presentation on the selection of Pilot Learning Sites for the Sub Saharan Africa Challenge Program (SSA CP), Accra, Ghana, February 2008. Herrero, M., Thornton, P.K., Kruska, R., Reid, R.S. 2008. Systems dynamics and the spatial distribution of methane emissions from African domestic ruminants to 2030. Agriculture, Ecosystems and Environment 126, 122?137. Herrero, M., Thornton PK., Notenbaert AM., Wood S., Msangi S., Freeman HA., Bossio D., Dixon J., Peters M., van de Steeg J., Lynam J., Parthasarathy Rao P., Macmillan S., 63 Gerard B., McDermott J, Ser? C., Rosegrant M. 2010a. Smart investments in sustainable food production: revisiting mixed crop-livestock systems. Science 327, 822- 825. Herrero, M., Thornton, PK., Nouala, S., Notenbaert, A., Ericksen, P., de Leeuw, J. 2010b. Coping with drought and climate change in the pastoral sector in Sub-Saharan Africa: policy considerations. Policy paper as input into the 8th Conference of Ministers of Livestock and Animal Health in Africa, May 2009, Kampala, Uganda, 6 p. Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A., 2005. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25, 1965?1978. Hulme, M., R.M. Doherty, T. Ngara, M.G. New, and D. Lister. 2001. "African climate change: 1900-2100," Climate Research, Vol. 17, No. 2, pp. 145-168. Indeje, M., Semazzi, F.H.M., Ogallo, L.J., 2000. ENSO signals in East African rainfall and their prediction potentials. Int. J. Climatol. 20, 19?46. IPCC. 2001. Climate Change 2001. Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. IPCC (Intergovernmental Panel on Climate Change) (2007). The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK. Jones, P.G., 1987. Current availability and deficiencies data relevant to agro-ecological studies in the geographical area covered in IARCS. In: Bunting, A.H. (Ed). Agricultural environments: Characterisation, classification and mapping. CAB International, Oxfordshire. pp 69-82. Jones P G, Thornton P K, 2000. MarkSim: Software to generate daily weather data for Latin America and Africa. Agronomy Journal 93, 445-453. Jones, P.G. and Thornton, P.K., 2003. The potential impacts of climate change in tropical agriculture: the case of maize in Africa and Latin America in 2055. Global Environmental Change 13, 51-59. Jones P G and Thornton P K, 2009. Croppers to livestock keepers: Livelihood transitions to 2050 in Africa due to climate change. Environmental Science and Policy 12, 427-437. 64 Karanja, F., Ogallo, L.J., Mutua, F. M., Oludhe, C., Kisia S. (2002), ?Kenya Country Case Study: Impacts and Response to the 1997-98 El Ni?o Event?, Available at http://www.ccb.ucar.edu/un/kenya.html Kenya Red Cross. 2009. Drought Appeal 2009 - Alleviating Human Suffering. Kenya Red Cross Society (KRCS), District Steering Groups (DSGs), Kenya Food Security Steering Group (KFSSG). http://www.kenyaredcross.org KNMI. 2006. Climate change in Africa. Changes in extreme weather under global warming, Royal Netherlands Institute of Meteorology, http://www.knmi.nl/africa_scenarios/. KSS, 1982a. Exploratory Soil Map and Agroclimatic Zone Map of Kenya. Eds. Sombroek, W.G., Braun, H.M.H., & van der Pouw, B.J.A. Nairobi: Kenya Soil Survey. Kurukulasuriya, P., Mendelsohn, R., Hassan, R., Benhin, J., Deressa, T., Diop, M, Eid, H.M, Yerfi Fosu, K., Gbetibouo, G., Jain, S., Mahamadou, A., Mano, R., Kabubo-Mariara, J., El-Marsafawy, S., Molua, E., Ouda, S., Ouedraogo, M., Sene, I., Maddison, D., Seo, N., and A. Dinar. 2006. Will African Agriculture Survive Climate Change? The World Bank Economic Review. 22p. Lesnoff M, 2007. DynMod ? A tool for demographic projections of ruminants under tropical conditions. User?s Manual. International Livestock Research Institute, Nairobi, Kenya. 29 pages. McSweeney, C. M. New and G. Lizcano. In press. UNDP Climate Change Country Profiles ? Kenya. Available on http://country-profiles.geog.ox.ac.uk/. Meze-Hausken, E. 2004. Contrasting climate variability and meteorological drought with perceived drought and climate change in northern Ethiopia. Climate Research, 27, 19? 31. Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well-being. Volume 1: Current State & Trends Assessment. Millennium Ecosystem Assessment. Island press. Mitchell, J.F.B, Johns, T.C., Senior, C.A. 1998. Transient response to increasing greenhouse gases using models with and without flux adjustment. Hadley Centre Technical Note 2. UK Met Office, Bracknell. Mitchell, T.D., Carter, T.R., Jones, P.D., Hulme, M., New, M. 2004. A comprehensive set of high-resolution grids of monthly climate for Europe and the globe: the observed record (1901-2000) and 16 scenarios (2001-2100). Tyndall Centre for Climate Change Research Working Paper 55. Tyndall Centre, Norwich. 65 Mogaka, H., Gichere, S., Davis R., and Hirji, R. 2006. Climate variability and water resources degradation in Kenya: improving water resources development and management. World Bank working paper nr 69. World Bank, Washington. Moore, N., Alagarswamy, G., Pijanowski, B., Thornton, P.K., Lofgren, B., Olson, J., Andresen, J., Yanda, P., Qi, J., Campbell, D., 2009. Food production risks associated with land use change and climate change in East Africa. IOP Conference Series: Earth and Environmental Science 6, 342003. doi:10.1088/ 1755-1307/6/4/342003. Mutai, C.C., Ward, M.N., Coleman, A.W., 1998. Towards the prediction of the East Africa short rains based on sea-surface temperature-atmosphere coupling. Int. J. Climatol. 18, 975?997. Nelson, Gerald C., Mark W. Rosegrant, Jawoo Koo, Richard Robertson, Timothy Sulser, Tingju Zhu, Claudia Ringler, Siwa Msangi, Amanda Palazzo, Miroslav Batka, Marilia Magalhaes, Rowena Valmonte-Santos, Mandy Ewing, and David Lee. 2009. Climate change: Impact on agriculture and costs of adaptation. IFPRI: Washington DC. Mogaka, H., S. Gichere, R. Davis, and R. Hirji. 2006. Climate Variability and Water Resources Degradation in Kenya: Improving Water Resources Development and Management. World Bank Working Paper No. 60. Washington, D.C.: World Bank. Nicholson, S.E. 1996. A review of climate dynamics and climate variability in Eastern Africa. In: Johnson, T.C., Odada, E.O. (Eds), The Limnology, climatology and paleoclimatology of the East African lakes. Gordon and Breach Publishers, SA Nkedianye D, Radeny M, Kristjanson P, Herrero M, 2009. Assessing returns to land and changing livelihood strategies in Kitengela. Pp 115-150 in K Homewood, P Kristjanson, P Chevenix Trench (eds), Staying Maasai? Livelihoods, Conservation and Development in East African Rangelands. Springer, Dordrecht. Ogallo, L.J., Janowiak, J.E., Halpert, M.S., 1988. Teleconnection between seasonal rainfall over East Africa and global seas surface temperature anomalies. J. Met. Soc. Jpn. 66, 807?822. Osbahr, H and Viner, D., 2006. Linking Climate Change Adaptation and Disaster Risk Management for Sustainable Poverty Reduction. Kenya Country Study. A study carried out for the Vulnerability and Adaptation Resource Group (VARG) with support from the European Commission. Otiende, B. 2009. The Economic Impacts of Climate Change in Kenya: Riparian Flood Impacts and Cost of Adaptation. Kenya National Advisory Committee for the DFID funded study on the Economic Impacts of Climate Change in Kenya. 66 Reid, R.S., Thornton, P.K., Kruska, R.L., 2004. Loss and fragmentation of habitat for pastoral people and wildlife in East Africa: concepts and issues. South Afr. J. Grass For. Sci. 21, 171?181. Reid, R.S., Gichohi, H., Said, M.Y., Nkedianye, D., Ogutu, J.O., Kshatriya, M., Kristjanson, P., Kifugo, S.C., Agatsiva, J.L., Adanje, S.A., Bagine, R. 2008. Fragmentation of a Peri- Urban Savanna, Athi-Kaputiei Plains, Kenya. Pp 195-224 in: K.A. Galvin et al (eds) Fragmentation in Semi-Arid and Arid Landscapes: Consequences for Human and Natural Systems. Springer Netherlands. Ropelewski, C.F., Halpert, M., 1987. Global and regional scale precipitation patterns associated with the El Ni?o/Southern Oscillation. Monthly Weather Rev. 115, 1606? 1626. Rutten M M E M, 1992. Selling Wealth to Buy Poverty; the Process of the Individualization of Land Ownership among the Maasai Pastoralists of Kajiado District, Kenya, 1890- 1990. Verlag Breitenbach Publishers, Saaarbrucken. Smith, L.C. and L. Haddad, 2000. Explaining Child Malnutrition in Developing Countries: A Cross-Country Analysis. Research Report 111. International Food Policy Research Institute, Washington, DC, U.S.A. Thornton PK, Jones PG, Owiyo TM, Kruska RL, Herrero M, Kristjanson P, Notenbaert A, Bekele N and Omolo A, with contributions from Orindi V, Otiende B, Ochieng A, Bhadwal S, Anantram K, Nair S, Kumar V and Kulkar U, 2006. Mapping Climate Vulnerability and Poverty in Africa. Report to the Department for International Development, ILRI, PO Box 30709, Nairobi 00100, Kenya. Pp 200. Thornton, P.K., Jones, P.G. 2008. Unpublished data layers. ILRI, Nairobi. Thornton P K, Jones P G, Alagarswamy G, Andresen J, Herrero M, 2009. Adapting to climate change: agricultural system and household impacts in East Africa. Agricultural Systems (in press), online at http://dx.doi.org/10.1016/j.agsy. 2009.09.003 Thornton P. K. and Herrero, M. 2010. The inter-linkages between rapid growth in livestock production, climate change, and the impacts on water resources, land use, and deforestation. Background paper for the 2010 World Development Report. Policy Research Working Paper 5178, The World Bank, Washington, US, 82 p. Sivakumar, M.V.K., H.P. Das, O. Brunini. 2005. Impacts of present and future climate variability and change on agriculture and forestry in the arid and semi-arid tropics. Climatic Change (2005) 70: 31?72 67 UNEP/GoK. United Nations Environment Programme (UNEP) and Government of Kenya (GoK) 2000. Devastating Drought in Kenya. Environmental Impacts and Responses. UNEP/GoK. Nairobi. Van de Steeg JA, Herrero M, Kinyangi J and Thornton PK, 2009. The influence of climate variability and climate change on the agricultural sector in East and Central Africa - Sensitizing the ASARECA strategic plan to climate change. Report 22. ASARECA (Association for Strengthening Agricultural Research in Eastern and Central Africa), Entebbe, Uganda, and ILRI (International Livestock Research Institute), Nairobi, Kenya. Washington, R., Harrison, M., Conway, D., Black, E., Challinor, A., Grimes, D., Jones, R., Morse, A., Kay, G. and Todd, M. 2006. African climate change: taking the shorter route. Bulletin of the American Meteorological Society 87:1355?1366. World Resources Institute; Department of Resource Surveys and Remote Sensing, Ministry of Environment and Natural Resources, Kenya; Central Bureau of Statistics, Ministry of Planning and National Development, Kenya; and ILRI, 2007. ?Nature?s Benefits in Kenya, An Atlas of Ecosystems and Human Well-Being?, (Washington, DC and Nairobi: World Resources Institute). Available at http:// www.wri.org/ You, L., Wood, S., 2004. Assessing the spatial distribution of crop production using a cross- entropy method. EPTD Discussion Paper 126. IFPRI, Washington D.C. 48 p. You, L. and S. Wood, 2006. An entropy approach to spatial disaggregation of agricultural production. Agricultural Systems Vol.90, Issues1-3 p.329-347. 68 Appendices 69 Appendix A Figure 1: Distribution and yield in MT/Ha of Maize, 2000 Arid Semi-Arid Sub-Humid Humid 70 Figure 2: Distribution and yield in MT/Ha of Sorghum, 2000 Arid Semi-Arid Sub-Humid Humid 71 Figure 3: Distribution and yield in MT/Ha of Millet, 2000 Arid Semi-Arid Sub-Humid Humid 72 Figure 4: Distribution and yield in MT/Ha of Wheat, 2000 Arid Semi-Arid Sub-Humid Humid 73 Figure 5: Distribution and yield in MT/Ha of Cassava, 2000 Arid Semi-Arid Sub-Humid Humid 74 Figure 6: Distribution and yield in MT/Ha of Potato, 2000 Arid Semi-Arid Sub-Humid Humid 75 Figure 7: Distribution and yield in MT/Ha of Sweet potato, 2000 Arid Semi-Arid Sub-Humid Humid 76 Figure 8: Distribution and yield in MT/Ha of Coffee, 2000 Arid Semi-Arid Sub-Humid Humid 77 Figure 9: Distribution and yield in MT/Ha of Sugarcane, 2000 Arid Semi-Arid Sub-Humid Humid 78 Figure 10. Livestock density maps (TLU/km2) 79 Appendix B Figure A1. Kenya: Spatial patterns of projected change in monthly precipitation for 10-year periods in the future under the SRES A2 scenario. All values are anomalies relative to the mean climate of 1970- 1999 . Source: McSweeney et al, in press, http://country-profiles.geog.ox.ac.uk/ 80 Appendix C. Production systems in Kenya. Livestock production systems consist mostly of pastoralists, while mixed systems represent crop-livestock systems where dairy predominates and where different crops, primarily maize and beans are planted in single stands or intercropped. Choice of crop is also determined by agro-ecology. In the mixed systems in the highlands maize and potatos predominate together with cash crops such as coffee and tea. Sugar cane, sweet potato and maize grow mostly in the humid areas while millets and sorghum are restricted to the semi-arid regions. In terms of livestock, most meat production predominates in arid and semi-arid regions and comes from a mixture of cattle, sheep and goats. Sheep and goat production is growing at faster rates than cattle production in these areas (Herrero et al 2008). Camels are also replacing cattle in these environments. Dairy predominates in the highlands. Livestock based systems Mixed irrigated systems Mixed rainfed systems 81 Appendix D - Generating plausible crop distribution and performance maps This text is based on the abstract in ?Generating Plausible Crop Distribution and Performance Maps for Sub-Saharan Africa Using a Spatially Disaggregated Data Fusion and Optimization Approach? by Liangzhi You, Stanley Wood and Ulrike Wood-Sichra (2007). IFPRI Discussion Paper 725. Agricultural production statistics reported at country or sub-national geopolitical scales are used in a wide range of economic analyses, and spatially explicit (geo-referenced) production data are increasingly needed to support improved approaches to the planning and implementation of agricultural development. However, it is extremely challenging to compile and maintain collections of sub-national crop production data, particularly for poorer regions of the world. Large gaps exist in our knowledge of the current geographic distribution and spatial patterns of crop performance, and these gaps are unlikely to be filled in the near future. Regardless, the spatial scale of many sub-national statistical reporting units remains too coarse to capture the patterns of spatial heterogeneity in crop production and performance that are likely to be important from a policy and investment planning perspective. To fill these spatial data gaps, You et al. (2007) developed and applied a meso-scale model for the spatial disaggregation of crop production. Using a cross-entropy approach, the model makes plausible pixel-scale assessment of the spatial distribution of crop production within geopolitical units (e.g. countries or sub-national provinces and districts). The pixel-scale allocations are performed through the compilation and judicious fusion of relevant spatially explicit data, including production statistics, land use data, satellite imagery, biophysical crop ?suitability? assessments, population density, and distance to urban centers, as wells as any prior knowledge about the spatial distribution of individual crops. Using the modified spatial allocation model, they generated 5-minute (approximately 10-km) resolution grid maps for 20 major crops across Sub-Saharan Africa, namely barley, dry beans, cassava, cocoa, coffee, cotton, cowpeas, groundnuts, maize, millet, oil palm, plantain, potato, rice, sorghum, soybeans, sugar cane, sweet potato, wheat, and yam. An example of estimated distribution maps for Sorghum, Maize and Millet are given in the Figure B1. The approach provides plausible results but also highlights the need for much more reliable input data for 82 the region, especially with regard to sub-national production statistics and satellite-based estimates of cropland extent and intensity. Figure B1. Estimated crop distribution maps of Sub-Saharan Africa 83 Appendix E. Crop and economic modelling methods Generating locale-specific yield responses to climate change Biophysical yield responses to soil, nutrients and climate change generated by the Decision Support System for Agrotechnology Transfer (DSSAT) crop simulation model distributed across the globe based on crop calendars, soils, and the ISPAM dataset of crop location and management techniques (You and Wood, 2006; see also www.mapspam.info). Distributed crop simulation model results are then aggregated into the 281 food producing units that form the basic elements of IMPACT. On the water side, results from the GCMs are fed into a global hydrologic simulation model to account for impacts on runoff and evapotranspiration from changes in temperature and precipitation patterns. Modeling Climate Change Impacts Climate change effects on crop productivity enter into the IMPACT model by affecting both crop area and yield. For example, crop yields are altered through the intrinsic yield growth coefficient in the yield equation as well as the water availability coefficient for irrigated crops. Intrinsic growth coefficients, or technological change, depend on crop, management system, location, yield trends, and agricultural research investments. For most crops, the average is about 1 percent per year. We generate relative climate change productivity effects by calculating location-specific yields for each of the five crops modeled with DSSAT for 2000 and 2050 climate as described above and then constructing a ratio of the two. The ratio is then used to alter the intrinsic rate of technological change. Rainfed crops react to changes in precipitation and temperature as modeled in DSSAT. For irrigated crops, the effect of temperature is derived from the DSSAT results and Water stress effects are captured in the hydrology model connected with IMPACT, reducing water availability for irrigation. The role of carbon fertilization Scenarios can be run with or without increased carbon fertilization effects. Plants produce more vegetative matter as atmospheric concentrations of CO2 increase. The effect depends on the nature of the photosynthetic process used by the plant species. C3 plants use CO2 less efficiently than C4 plants, which benefit from elevated atmospheric concentrations of CO2. Uncertainty remains regarding the translation of mostly laboratory results to actual field conditions. DSSAT has an option to include CO2 fertilization effects at different levels of CO2 atmospheric concentration. However, when compared to recent evidence in field trials, it appears that the CO2 fertilization effects currently embedded in the DSSAT models may overstate the benefits of carbon fertilization [Personal communication, Kenneth J. Boote, 84 Professor, Agronomy Department, University of Florida, March 2009].. To capture the uncertainty in actual field effects, we simulate two levels of atmospheric CO2 in 2050: 369 ppm (the level in 2000) and 532 ppm, the expected CO2 levels 2050 for one of the scenarios (A2 for a comparison of results). Thus, we compare a situation with CO2 fertilization with a situation without.