1. Introduction (NPC)
1. Rift valley focus
2. Relation to overall grant
3. Overall DSS objectives and description
4. Cross-site study with Sikasso
5. Major products of this study
2. Objectives and Outcomes
1. Relation to World Food Summit Goals
2. Meeting the Future Needs of Kenya
1. Design and major thrusts of the CRV study
3. Criteria for Selection of Agricultural Technologies by KARI
1. Maize
2. Wheat
3. Beans
4. Potatoes
5. Other Crops
3. Integrated Regional and National Analysis
4. Strategic Design of the Study
3. Food Security and Natural Resources in Kenya (RK, TB, NPC)
1. Current Status of Kenya Agriculture and Food Security
2. Future Impact of Climate Change
3. Impact of National and Regional Population Projections
4. Regional Capacity of Kenya for Food and Agriculture
5. Link between natural resources and food security
4. The Rift Valley — The Bread Basket of Kenya
1. The Geography of the Central Rift Valley
2. Importance of the Central Rift Valley to Kenya Agriculture
3. Selection of the Central Rift Valley as a Regional Focus of the Study
5. Experimental Design (Methods) — Integrated Approach for Impact Assessment
1. Institutional Definition of Policy or Technology Package to be Evaluated

2. Criteria for Selection of Agricultural Technologies by KARI
1. Maize
2. Wheat
3. Beans
4. Potatoes

1. Assessment of Analytical Tools and Data Requirements
2. Spatial Characterization of the Impact Area
1. Farm Household Survey Spatial Sampling Frame
2. Simulation Zones for Biophysical Modeling and Sectoral Stratification
3. Acquisition of Primary and Secondary Data to Support the Models
1. Representative Farm Household Survey Data
2. Biophysical Data
3. Economic Data
4. Prediction of Crop and Forage yields — Modeling, Verification, Meta Models
1. Crop Simulations
2. Grazingland Forage Yields
3. Creation of Meta-Equations for Crop and Forage Yields
5. Prediction of Economic Outcomes
1. Agricultural Sector Level
2. Farm Household Level
6. Assessment of Environmental Consequences Reflecting Economic Projections
1. Impacts at the Landscape Level
2. Impacts at Farm Level
7. Geographic synthesis of Impact
8. Define Key Outcomes and Analysis of Relative to Strategic Questions Addressed

1. Results
1. Impact on Crop and Livestock Productivity
2. Impact on the agricultural sector of the Rift Valley and Kenya
1. Agricultural Sector economic impacts
2. Environmental impacts on cropland and grazingland
3. Implications of risk of hunger
1. Impact of new technologies toward achieving WFS goal
2. Yields necessary to achieve WFS goal
4. Impact of new technology on individual farms/households
1. Nature of farms in Rift Valley
2. Economic impact at the farm level by type of farm
3. Environmental impacts at the farm level by type

This study focuses on the Central Rift Valley (CRV) of Kenya and examines the impact of alternative agricultural technology options and associated policy on the contribution of this highly productive region to current and future food security in Kenya. An integrative view of both economic and environmental impacts is provided in a spatially coherent manner.

Application of a Decision Support System (DSS)". The research is conducted in collaboration between scientists in the Center for Natural Resource Information Technology (CNRIT) at Texas A&M University, national scientists, and government decision-makers in Kenya and in the U.N. Food and Agriculture Organization.

The Decision Support System project has the overall objective of developing an integrated suite of models and databases that link economic, environmental, and biophysical elements into a holistic system to evaluate policy and technology options at levels of scale from local to national. There is a strong spatial analysis component in the methodology and an active use of remotely sensed information. Some of the methods are being further developed for use in near-real-time operational applications such as marketing and early warning systems. The product of this research is a set of analytic tools that can be used by decision-makers at local, sub-national, national, and multinational levels. The primary host countries of the overall study are Kenya and Mali with participation of adjacent East and West African countries.

Part of the overall DSS project plan is to compare the results of projections of impact of policy and technology options in these two highly productive but divergent regions in East and West Africa to seek transcending principles for achieving food security through sustainable use of natural resources. The DSS project in Kenya includes a study of climate change in Kenya and adjacent countries, which are complimentary to the Central Rift Valley, study and reported elsewhere.

The major products of the overall DSS project, to which this Central Rift Valley study contributes, is a the suite of integrated models for decision-makers, a contemporary set of detailed geo-referenced databases, and the results of a series of specific studies that have been called for by senior decision-makers and conducted cooperatively with national partners. Kenyan researchers, civil servants, and senior decision-makers have been trained and have experience in the use of the DSS.

Previous, current, and related products relevant to this report are found at:

http://cnrit.tamu.edu

2. Objectives and Outcomes of the Central Rift Valley Study

The World Food Summit (WFS) has the goal of reducing hunger by 50% in the year 2015. In the meeting on the WFS + 5 years, the ultimate goal was reaffirmed with some slippage in schedule. The Government of Kenya (GOK) is a signatory of this international convention, reaffirming its intent to place high national priority on achieving improved long-term food security. International investors, including USAID, are also focused on supporting the development of efforts to enhance food security through the sustainable use of natural resources. As this study will describe, current population projections, even in the face of HIV-AIDS, the rapid urbanization of its population, and the continuing deterioration of natural resources using present farming practices point toward a decline rather than an increase in food security in Kenya by the target year of 2015. The GOK recognizes the challenges in meeting its long-term goals in this area and continues to refine its overall national strategy to reverse this trend and meet its future needs.

The proof of concept of the DSS was established in earlier studies that evaluated the impact of smallholder dairy operations in Kenya and East Africa over the last 20 years. Through their active participation in these studies, GOK decision makers and scientists in the Kenya Agricultural Research Institute (KARI) recognized the larger potential of the DSS as an analytic method for extending and making more explicit the government’s plans for future food security. The plan for the overall project, with focus of this element on the Central Rift Valley, was developed in a workshop held at KARI Headquarters in September 2000. A second more detailed workshop was held at Nakuru in August 2001 to define the scope of the CRV study. It was conducted by a team of KARI scientists, Ministry of Agriculture workers, regional Extension specialists, and CNRIT scientists. The project had two major thrusts: (1) the extension and new development of the DSS using the Rift Valley study as a platform and (2) the application of the DSS to a series of questions posed by senior GOK officials as they continue to pursue the goals of the WFS. This paper emphasizes the second thrust — the studies done to apply the DSS to specific options and outcomes that affect food security in the future.

Because the study examined options to enhance future food security, the Rift Valley was a major focus study since, as shown below, it has been called "the bread-basket of Kenya" by the FEWS and will provide the natural resource base for achieving much of the WFS goals in the future. Advanced technology packages were evaluated for four major cereal grains which are prominent in the Rift Valley and which will be pivotal in future food security.

After many years of breeding and several years of testing under the National Performance Trial programs, the Kenya Agricultural Research Institute (KARI) and its partners have developed superior varieties of food crops for multiplication by seed merchants to benefit farmers. The varieties developed are adapted to specific agro-ecological zones (Table 2-1 characterized by differences in elevation, rainfall, soils and temperatures. To date, KARI has developed and released maize (11), wheat (8), sorghum (4), beans (4), potato (2), finger millet (1), cow peas (2), pigeon peas (1) and sweet potato (5) varieties along with associated farm management practices. These varieties and their agronomic recommendations are released periodically for farmers to adopt them in an effort to enhance production in the agricultural sector addressing the problem of food security and poverty alleviation. Recent review of farmer production practices and constraints in the Central Rift Valley (Nkoge et al, 1997) indicated that some of the farmers were already planting the varieties released by KARI and recommended for this area. Beside the improved varieties, they were also planting their local varieties (landraces). Table 2-1 indicates the crop varieties characteristics either already being used by farmers or best bet varieties recommended by KARI. A specific description of each crop is provided in the following sections.

Maize is the main food crop for the country, grown on 1.5 million hectares nationwide. Production fluctuates depending on the weather changes but on average between 2.52 —2.7 million tonnes are produced. Between 50-60% of the maize area is under improved varieties while the rest is under local varieties. High yielding hybrids and composites have been developed which have the following potential yield when all factors are right varying from Katumani composite for dry areas to the latest new release, KH600-11D which has a potential to produce over 8.9 t/ha. Farmers. Depending on agro-ecological zonation in the CRV, the hybrid varies recommended are: H512, H511, H614D, H625, H626 and Kh600-11D. Farmers do use at various proportions Katumani composite and their local varieties. National maize production is on average about 2.1 t/ha, well below potential of many of the varieties available. The poor yields are associated with a combination of factors such as lack of credit for farmers to procure seeds of improved varieties, inadequate inputs (fertilizers, seeds, chemicals) and machinery and Unfavorable weather conditions.

The country produces only 0.27 million tonnes against an annual consumption of 0.54 million tonnes. The shortfall is met by imports. The ministry of agriculture and KARI releases varieties every year, and it is a common practice for farmers to use their previous seasons wheat crop seed and they make effort to buy the recommended seed varieties suited for their agro-ecological. The wheat varieties recommended for the 1999-2000 crop growing seasons were: Duma, Ngamia, Mbega and Ogema. The major constraints contributing to poor productivity include technological packages that favor large farms and poor credit facilities to small-scale farms.

This is the major source of staple protein in the country, grown in over 700,000 Ha annually with a production of about 0.45 million tons. Superior varieties such as Katumani B1, B9, Rose Coco, Red Harricot, Mwezi Moja and Canadian Wonder have been released by KARI. Production is weather dependent and is usually enough to meet the consumption requirements. This crop is grown mainly by small-scale producers as an intercrop with maize or as a second season crop.

4. Potatoes

The production of disease-free seed tubers of high-yielding varieties is recognized as an important aspect in boosting yields and controlling seed—borne diseases in farmers' fields. Farmers have a practice of either retaining or selling as seed the small-sized tubers from their harvests. This has helped in rapid spread of major old varieties such as Kerr's Pink, Nyayo, and the newly KARI released varieties Tigoni and Asante producing only what is desired by the market.

The Government places a lot of emphasis on the promotion of traditional food crops as a food security measure. Some of these crops include millet, sorghum, various root crops, an assortment of fruits and a wide range of pulses. The majority of the traditional food crops have capacity to withstand harsh ecological conditions, hence their importance in drought prone areas. These crops are grown in varying hectarages in many parts of the country. However, KARI deemed it important to address the role of low input potatoes in the Rift Valley.

Through the use of an integrated suite of models and spatially explicit data, the studies done here provide products that evaluate the technology packages at farm, sub-national, and national levels with an evolving capability to scale up the results from smaller to larger scales up to national levels. While the performance of new technology systems is evaluated at both local and national levels, estimates of changes in food security and risk of hunger were made at the national level.

The strategic design of this study includes the following elements:

The DSS, as an integrated (interactive) analytical system provides:

Economic

Biophysical

Environmental assessment of the production system options considered

3. Food Security and Natural Resources in Kenya
1. Current Status of Kenya Agriculture and Food Security

Kenya is a country that is heavily dependent on agriculture as it plays a major role in providing livelihood to over 85% of the population and contributes over 65% of foreign exchange earnings (World Bank, 1991). The agriculture sector’s contribution to GDP has progressively declined from 37% of GDP in the early 1970’s to about 25% at the end of 2000 (GOK, 2002). Given the importance of agriculture in Kenya, the government has given top priority to the agricultural sector since 1963. However, several challenges have been encountered in the efforts to further develop the sector. Over the last 20 years, Kenya has experienced frequent and lengthy droughts such as those that occurred between 1991-1993 and again from 1998 to 2000. Less than 20% of the country is suitable for rain fed agriculture, the rest being arid and semi-arid lands. Due to population pressure, more people have moved from high potential areas to settle in more fragile environments without corresponding appropriate technologies for utilizing resources in such areas.

The Kenya government has recognized the need for food security, making this a core subject in all long, medium and short-term national development plans affecting statutory policy documents produced to date. The National Food Policy Paper No. 4 of 1981 (GOK, 1981) addresses the issue of the ever-increasing demand for food. The quest for food security and the avoidance of hunger and famine is as old as civil society itself.

The definition of food security as agreed at the World Food Summit in 1996, and refined in later years thus states… "Food Security exists when all people, at all times, have physical and economic access to sufficient, safe and nutritious food to meet their dietary needs and food preferences for an active and healthy life". Food security is not just a supply issue, but also a function of income and purchasing power, and hence its strong relationship with poverty.

Agriculture’s full potential for meeting the needs of society in Kenya has been hampered by a number of factors. These include decreasing household farm size, inadequate use of appropriate technology, unreliable rainfall, poor marketing infrastructure, limited access to credit, high costs of farm inputs including agricultural machinery, poor market information and early warning systems and lack of coherent land use policy. In recent years, production of food crops has been declining largely due to unfavorable policies, weather variability, liberalization, marketing constraints, the high cost of inputs and static technologies. Land under food crops has been continuously subdivided to accommodate a growing rural population.

The most important food crops in Kenya are maize, beans, sorghum, millet, potatoes and cassava. The important cash crops are tea, coffee and sugar cane. Considering that only 13% of the country is suitable for agriculture, there is intensive competition between food and cash crops in use of land. The farmers allocate their land to different crops according to what they perceive as potential future prices.

Maize is Kenya’s staple food and on average the area under the crop annually is about 1.5 million hectares. The long rain season maize production varies between 2.3-2.7 million MT, out of which 75% is acquired through small-scale farming. Maize yields average 2 MT per hectare, but the potential exists to increase the average yield to 6 MT per hectare, thereby ensuring self-sufficiency in production and surplus for export. Wheat production has stagnated at 270,000 MT per year but occasionally drops to as low as 180,000 during drought periods. The national demand is estimated at 600,000 MT with the difference being met from imports. Rice is produced through irrigation in the Mwea, Ahero, West Kano and Bunyala irrigation schemes. The average annual production is 52,000 tonnes, which is about 34% of national consumption. There has been a declining area under sorghum and millet production for the last 30 years. This trend has been attributed to a change in eating habits, low productivity, pests and a narrow range of uses of sorghum and millet. Production of pulses has also been declining due to weather conditions, use of low quality seed and high cost of inputs.

Kenya and its partners in development have made a lot of progress towards achieving food self sufficiency and ensuring adequate supply of nutritional food in all parts of the country. The country’s main goal is to reduce poverty and hunger by at least 50% by 2015. The government is planning to achieve this by creating an enabling environment for private sector and expanding grass-root partnerships that build local food systems and replicate best practice of existing efforts. Surprisingly, many of the efforts so far employed such as introduction of new technologies in agriculture seem not to have achieved their targets. Needless to say, many communities feel more food insecure because of increasing uncertainties and breakdown of social systems.

The major factors that have contributed to increasing food insecurity are:

• Poor food policy ignoring technology and the resource base
• Lack of early warning systems
• Inadequate strategic reserves
• Inadequate research and dissemination of technological innovations
• Weak farmers’ institutions as well as poor linkages and coordination
• Inadequate credit, poor infrastructure and marketing channels
• Insecurity especially in pastoral areas
• Lack of effective control of crop and livestock diseases
• High incidences of HIV/AIDS.

1. Future Impact of Climate Change

It is generally believed that greenhouse gas induced climatic change may negatively impact agriculture. Rosenzweig and Iglesias (1994) argue that most of the developing world may be hard hit by climatic change. IAG conducted a study on climate change impact in Mali and concluded that the Malian agriculture might be negatively impacted leading to considerable food security challenges. As food security in Kenya is heavily dependent on the country’s agriculture, climate change may become a critical factor in meeting the future food demand of burgeoning Kenyan population. IAG is conducting a separate study on climate change impact on Kenya’s agriculture and its implications for the country’s food security. The findings of this study will be available for sharing with the Kenyan government and donor community towards the end of the year 2004.

2. Impact of National and Regional Population Projections

The population census of 1999 revealed that there were about 29 million people in Kenya at that time. Ten years earlier (1989), the country had one of the highest birth rates in the world, at 4%. However, with concerted government and civil society pressures for smaller family size, birth control awareness campaigns and increased incidence of HIV/AIDS, this rate has now fallen to about 2.8%. While the average population density over the whole country is low, the distribution of people is skewed. Over 75% of the total rural population lives in the Highlands where the average population density is close to 200 people per square kilometer. Close to 70% of the 29 million people live in the rural areas, while the rest live in the cities. Overall, the population growth in the cities is higher than in the rural areas as more and more people move into the cities.

The expanding urban population has two effects: it places pressure on the housing in the cities resulting in the expansion of the urban areas into surrounding countryside; and it puts pressure on the rural farmers who need more agricultural land to feed the growing urban population. The only land that offers opportunity for expansion of cropland is the nation’s scarce forestlands creating pressure for conversion to cropland. This pressure is exerted first by the rural population that lives adjacent to the forests. There is constant shifting of forest boundaries as the population increases and encroaches onto the forests. The increasing population has also put more pressure on water and wildlife resources.

Virtually all aspects of development have experienced the severe impact of HIV/AIDS at the household, community and national level. HIV/AIDS has created shortages in manpower and also overstretched social services, especially the health services and the social security system. Current estimates indicate 2.2 million Kenyans are HIV/AIDS positive; 1.5 million have died since 1984, leaving behind about 1.3 million orphans. The majority of the AIDS cases fall in the 15-49 year age group who form the pool of trained and productive manpower. Life expectancy has been reduced to about 47 years and infant and child mortality has increased by about 20% during the 1989 — 2000 period. The Pandemic poses a threat to long-term sustainable development. It is estimated that without HIV/AIDS gross domestic product would be 14.5% higher than the current levels.

HIV/AIDS increases present and future food insecurity through its impact on:

• Households’ ability to produce food, because of labor shortages and loss of knowledge about farming methods;
• Households’ ability to buy food, because of impoverishment due to the loss of productive family members and of assets;
• Communities’ ability to produce and buy food, as the epidemic reaches every home and neighbors become too overburdened to help each other with food, loans or a hand in the fields;
• Countries’ ability to import food, as HIV/AIDS reduces GDP growth per capita

HIV/AIDS effects on the agricultural sector and rural societies have been manifested in the labor forces, management inputs and financial stability. More specifically these effects can be detailed below:

• Decline in animal husbandry and livestock production
• Impaired post-production, food storage and processing
• A breakdown in support services occurs as staff fall ill
• Diversion of credit for agricultural production for care of sick relatives, funeral expenses and food. This results in a reduction in crop yields and loss of income causing families to default on loan repayments or sell assets
• Remote fields tend to be left fallow and the total output of the agricultural unit declines
• Switching from labor-intensive crops to less demanding ones has been observe

1. Regional Capacity of Kenya for Food and Agriculture

A prime objective of agricultural polices in the developing world has been to achieve food security; yet the concept has different connotations in the development community. Hunger, malnutrition, caloric deficit, insufficient access to food, low body mass index (BMI), stunting, and wasting are some of the terms used in describing the state of food insecurity. Use of these terms in the food policy literature is replete, however, without much substantive policy relevance. Maxwell (2000) cited 32 definitions of food security as given by various researchers and institutions. In this study, the indicators we will focus on are economic indicators of food scarcity or plenty and an indicator of per capita food consumption given the total availability of food and total population. The economic indicators included production, prices, and trade of major crop and livestock products. The indicator of per capita food consumption is based on the FAO (1996) methodology of computing the proportion of undernourished population, called Risk of Hunger.

The choice of these indicators was based on their relevance to policy making and the fact that they can be made part of an analytical framework to assess the impact of alternative policy options on food security.

2. Link Between Natural Resources and Food Security

Figure 3.1 The relationship of the Rift Valley study area to the general land use for Kenya

The natural resources that are available in Kenya and have a direct impact on food security, include grazinglands, livestock, forests, water, wildlife, and fish. Although Kenya is the most industrialized country in east Africa, by international standards, the country is still primarily an agrarian society. This means that people are still very dependent on the natural environment for their survival. For this reason, the line between having enough food and not having enough is very thin and subject to weather induced risk. The country depends on rain-fed agriculture to produce most of its food. If there is a failure of rain in one season, the country has to import food.

Further, most of the people in Kenya are not in the market place yet. They produce all the food they need for subsistence and do not have the wealth to buy it from the market. This means that if the environment fails to support these people, they immediately go on food relief. In addition, those people who live close to national parks or forests suffer from their food crops being damaged by wildlife. And if this happens, the farmers are not in a position to replant or to buy food from the market. They cannot simply replant, because agriculture is rain-fed. This means that crops can only be grown at certain seasons. And they cannot purchase food from the market because they do not have the money. Water sources have deteriorated in the water catchment areas of the country. This has happened when forestland was alienated and given to selected groups to convert into farmland. Degraded forests lead to flash floods as well as soil erosion in the surrounding farmlands as evident in the recent floods (citation !!)

For Kenyan people to have adequate food, it is necessary that a balance between agriculture and natural resources be maintained. This calls for policies that can adequately ensure the sustainable management of natural resources. It also calls for measures that will reduce people's direct dependence on natural resources. These measures include good food storage facilities; better animal husbandry in the drylands; and industrialization.

Development of improved agricultural farming system that keep pace with food demands will have a critical input in the social and economic welfare targets set in the long term National Development Plan of Kenya (GOK, 2002). USAID/REDSO has recognized the need for improved food security as per their strategic plan SO5: Enhanced African Capacity to achieve regional food security. The case for REDSO’s strategic decision to focus on food security is made in the U.S. Action plan on food security and the greater horn of Africa initiative strategic plan (USAID, 1997). The main thrust will be to improve food production technologies and help disseminated "best practices" through out the region that address key food security problems. The USAID mission in Kenya has also recognized food security needs by their strategic objective 7 which supports their mission goal of a well governed and more prosperous Kenya by promoting economic growth and focusing on households in rural areas, where the great majority of Kenyan’s poor reside.

1. The Geography of the Central Rift Valley

The Central Rift Valley (CRV) has been described by FEWS NET as the "Bread Basket" of Kenya. This ecologically diverse region is characterized by steep environmental gradients, diverse agriculture and widely varying population density. The primary agricultural production is located in the central and western region of CRV, representing the higher elevation areas of the valley. The Rift Valley created a geologic landscape that has many internally drained lakes on the valley floor running north and south through the region. The northern reaches of CRV are drier with steep terrain representing sparse to dense savannas and woodlands. To the northeast are productive rangelands with mixed private ranchers and group pastoral ranches interlaced with agro-pastoral communities. To the southeast in the Naivasha corridor, much of the land is un-improved pasture, given the somewhat unpredictable amounts of rainfall to support cropping. However, as you approach Nakuru from the north, you encounter a rich region of wheat and other cereal grain production. Most of the intensive agriculture of the CRV is associated with the central and western regions of CRV. The western reaches of CRV possess a major part of the forested regions of Kenya, associated with ecological problems due to competition of land use between forestry and cropping systems. Tourism is important in the central valley region given the influence of the lake system. Several game parks are located in the northeastern and northwestern region of CRV. Agricultural production can vary abruptly from bushed savanna where goats are the primary agricultural products to land growing banana and coffee in a matter of 15 to 20 km. Urban centers are typically located in the mid-elevation areas and are well distributed throughout the region. When compared to the peri-urban region of Nairobi, the Central Province or the more western provinces, population density is lower and agricultural production is high.



Figure 4-1. A geographical view of the Central Rift Valley reflecting terrain, agriculture, ecology and social infrastructure of the study region.

2. Importance of the Central Rift Valley to Kenya Agriculture

The Kenyan economy is heavily dependent on agriculture, the dominant sector in the economy. This sector produces food to feed the population and cash crops to earn money for the country. The most important food crops are maize, beans, sorghum, millet potatoes and wheat. The important cash crops are tea, coffee and sugar cane. Putting into account the fact that only 13% of the country is suitable for rain-fed agriculture, there is a lot of competition for land, between food crops, cash crops.

The eight provinces in Kenya range in geographical size from the small Nairobi province to the much larger Eastern and Rift Valley provinces. The provinces also range in population from the densely populated Nairobi Province, to the sparsely populated areas in the North East. The eight provinces are: Central, Coast, Eastern, Nairobi, North Eastern, Nyanza, Rift Valley and Western Province. Each Province in turn is subdivided into districts.

Most of the land of high or medium potential for farming lies in the Western Highlands, central Rift Valley, around Lake Victoria and Mount Kenya, and along the coast. Due to its relatively lower population density, as compared to similar climate areas in the country, the Central Rift Valley is referred to as the ﾒgrain-basketﾓ of Kenya. The Northern (Turkana, Samburu) and southern (Kajiado and Narok) districts do not have suitable climate for agriculture. Table 4-1 shows that the Rift Valley province produces 49.9%, 46.08%, 96.64% and 50.48% of maize, beans, wheat and tea respectively. It is the leading producer of maize, beans, wheat, tea and pyrethrum. Surprisingly, Eastern province produces 12.18% and 23.80% of maize and beans, respectively, the percentage land allocated to these crops is more than what is allocated in the Rift Valley (Table 4-2) indicating a more favorable production per unit area in the latter.

Table 4-1 Percent Crop production by province of Kenya.

PROVINCE	Maize	Beans	Potatoes	Wheat	Sorghum	Millet	Tea	Coffee
CENTRAL	4.86	10.76	39.03	2.40	0.36		38.20	63.37
COAST	0.61	0.18	0.01		0.07
EASTERN	12.18	23.80	37.08	0.73	39.85	53.67	2.62	20.04
NAIROBI	0.14	0.22	0.01
NEASTERN					0.04
NYANZA	17.20	11.70			48.73	30.43	8.18	1.80
RIFTVALLEY	49.90	46.08	23.65	96.64	4.86	9.95	50.48	13.97
WESTERN	15.10	7.26	0.21	0.23	6.09	5.95	0.52	0.82

Table 4-2 Percentage land allocated to crops by province of Kenya.

PROVINCE	Maize	Beans	Sorghum	Potatoes	Wheat	Millet	Tea	Coffee
CENTRAL	11.16	15.77	0.35	40.59	4.38		35.73	54.51
COAST	0.89	0.24	0.07
EASTERN	36.93	36.99	54.61	35.65	4.61	66.58	2.45	28.95
NAIROBI	0.10	0.16		0.15
NEASTERN			0.07
NYANZA	10.52	11.65	33.72			15.61	13.63	6.15
RIFTVALLEY	32.47	27.44	4.75	23.42	90.81	11.72	47.96	7.75
WESTERN	7.93	7.76	6.43	0.17	0.19	6.09	0.23	2.63

3. Selection of the Central Rift Valley as a Regional Focus of the Study

Given the need for Kenya to meet critical food security goals of Kenya by 2015, including both rural and the growing proportion of urban dwellers, and the central role that the CRV plays in meeting those needs today, the choice of this ecologically rich region as the focus of this study by the collaborating institutions in Kenya is well founded. Technologies for improvement in production of maize, beans and wheat in CRV will have a proportionally larger effect on agriculture’s ability to feed the people of Kenya. The improvement in potato production in the CRV with pest resistant varieties insures that this traditional crop will sustain caloric intake over a broader range of environments Kenya. Based on KARI’s released technologies and clear understanding of the country’s resources, the team of KARI scientist, senior policy and decision makers from Ministry of Agriculture and Livestock Development, Ministry of Finance and Planning, Office of the President and Ministry of Water, Environment and Natural Resources, together with Kenya’s development partners such as FAO and International Research Institutions recommended the CRV as the region to do this study. They also selected the districts were focus work should be done at the workshop held in Nairobi, September 2000.

Over the past five years, the approach used to assess impact of new policies and technologies has been refined by CNRIT. However, the fundamental methodology has not changed. The process must be followed in a broad manner using the steps outlined in the chart below (Figure 5-1).

This step involves first identifying relevant institutions that have a vested interest in the input and outputs of the analysis early on in the process. Using a participatory approach, key analysts and decision makers are asked to identify the key technologies to be evaluated or develop a concise definition of the policy that has been pursue or implemented. In this study, the Kenya Government’s Office of the President and Ministry of Finance and Planning (MFP) were most interested in the potential role that new agricultural crop technologies could play in meeting target food security needs of Kenya with special interest in the Central Rift Valley. A team of scientists from the Kenya Agricultural Research Institute (KARI) working with the Ministry of Agriculture and Rural Development, in collaboration with the Office of the President and MFP, organized focus groups to help identify the array of new crops and associated technologies that could best meet the food security needs of the country. The role of new crop varieties, coupled with improved farming practices, were targeted for maize, wheat, beans and potatoes by this group of planners. Although crop yields had to be reflected for the entire country in the sector analysis, the area of interest was decided to be the Central Rift Valley. The planning group also indicated that they were interested in both sectoral and farm level responses to the technology and the resulting environmental consequences of adoption of the technologies within the Central Rift Valley.

After many years of breeding and several years of testing under the National Performance Trial programs, the Kenya Agricultural Research Institute (KARI) and its partners had developed superior varieties of food crops for multiplication by seed merchants to benefit farmers. The varieties developed were adapted to specific agro-ecological zones (Table 5-1) characterized by differences in elevation, rainfall, soils and temperatures.

Recent review of farmer production practices and constraints in the Central Rift Valley (Nkoge et al, 1997) indicated that some of the farmers were already starting to plant the varieties released by KARI for this area. Table 5-1 represents the array of varieties including landraces already being used by farmers and the chosen improved varieties recommended by KARI. Specific descriptions of each crop are provided in the following sections.

Maize is the main food crop for the country, grown on 1.5 million hectares nationwide. Production fluctuates depending on the weather changes, but on average, between 2.52 —2.7 million tons are produced. Between 50-60% of the maize area is under improved varieties while the rest is under local varieties. Hybrids and composites have been developed which have high potential yields when all factors are optimal. These types vary from Katumani composite for dry areas to the latest new release, KH600-11D which has a potential to produce over 8.9 t/ha. Depending on agro-ecological zonation in the CRV, the hybrid varieties recommended were: H512, H511, H614D, H625, H626 and Kh600-11D. Farmers currently use various proportions of the Katumani composite as well as their local varieties. National maize production is on average about 2.1 t/ha, well below the potential of many of the varieties currently available. The poor on-farm yields are associated with a combination of factors such as lack of credit for farmers to procure seeds of improved varieties, inadequate inputs (fertilizers, seeds, chemicals) and machinery and unfavorable weather conditions.

The country of Kenya produces only 0.27 million tons of wheat annually against an annual consumption of 0.54 million tons. The shortfall in wheat production is met by imports. The Ministry of Agriculture and KARI release new wheat varieties every year, and it is a common practice for farmers to use their previous seasons wheat crop seed. Also, efforts are made by farmers to buy the recommended seed varieties suited for their agro-ecological zone. The wheat varieties recommended for this study were: Duma, Ngamia, Mbega and Ogema. The major constraints that contribute to poor on-farm productivity include technological packages that favor large farms over small farms, and poor access to credit facilities for small-scale farms.

Beans are the major source of staple protein in Kenya with over 700,000 ha under cultivation annually. Production has average approximately 0.45 million tons per year. Superior varieties have been developed such as Katumani B1, B9, Rose Coco, Red Harricot, Mwezi Moja and Canadian Wonder and have been released by KARI. Each of these varieties was evaluated in the technology packages for this study. On-farm production is highly dependent on climatic conditions and is usually enough to meet the consumption requirements for the country. The crop is grown mainly by small-scale producers as an intercrop with maize or as a second season crop.

The production of disease-free seed potatoes of high-yielding varieties is recognized as an important aspect in boosting yields and controlling seed—borne diseases in farmers' fields. Farmers have a practice of either retaining or selling as seed the small-sized potatoes from their harvests. This has helped in the rapid spread of major old varieties such as Kerr's Pink, Nyayo. This study focused on the newly released varieties Tigoni and Asante by KARI.

The nature of the technologies or policies and desired depth of analysis dictates the complexity of the methodology and the tools required to meet the needs of the collaborating institution. Given that the study involved complex arrays of varieties, each adapted to different types of environments and the need for large-scale landscape production and environmental response, a complex representation of agro-ecological zones was required. The analytical tools required were the use of geographic information systems (ArcView) to help manage the spatial data, the Agricultural Sector Model (ASM) to conduct the sectoral economic analysis and provide prices to the Farm Level Analysis Model (FLAM), which in turn allowed FLAM to determine impacts across representative farms in different production zones. To support the use of the economic models and provide environmental data, the SWAN crop model and PHYGROW rangeland model were used. SWAN was used primarily to predict crop yields under the various technology scenarios and to assess environmental impacts (i.e., loss/gain in soil N, humus, erosion, runoff, and nutrient loss to deep percolation) of those technology scenarios. The PHYGROW model was used to predict forage yields for livestock production under the technology scenarios and to assess environmental impacts (i.e, runoff) to grazing lands resulting from the technologies evaluated.

Secondary economic data was heavily relied on at the district and AEZ level. The weather, soils, and management input data had to be acquired to support the biophysical models. Section 5.4.2 provides a more comprehensive overview of data used in the economic and biophysical models.

Once the problem area was defined the area was stratified to support the array of analyses required to meet the needs of the planners. The methodology used in this study required farm surveys be conducted to support the farm-level analysis and biophysical models to be run on the range of environments found across the Rift Valley to support the sectoral level analysis. This process required that the Central Rift Valley be bounded for analysis. Two separate spatial analyses conducted, one for the farm surveys and the other to support the biophysical modeling efforts.

A spatially coherent sampling frame was developed to allow for a representative sampling of households across the CRV region to gather data needed for biophysical and economic models, and to allow scaling to the appropriate levels required by the decision support models. This sampling frame emphasized biophysical traits of the region such as soils, climate, and human population density rather than arbitrary political boundaries. This insured that the range of spatial differences that might influence farm yields and natural resource management across the CRV region was captured for both the economic and environmental impact assessment. The objectives for developing the spatial sampling frame were as follows:

1. Define the unique zones of similar climate.
2. Assess which climate zones were represented by the highest human population density,
3. Use climate zones weighted by population to determine the location of households for the rapid appraisal survey needed to locate representative farms (modal farms).

The first step in developing the sample frame was the identification of zones of unique climatology by using climate surface layers (5 x 5 km) derived from a procedure that interpolates historical weather station point data to a gridded surface (ANUSPLIN procedure, Hutchinson 1991). The interpolated climate surfaces were then fed into a growing season model that identifies the five consecutive months that maximize water availability in the environment (i.e. precipitation exceeds potential evapostranspiration). This growing season model has been shown to be quite effective in the identification of the growing season in Africa where water is the main limiting factor (Corbett 1995). A cluster analysis was run (Wards Minimum Variance method— SAS software) with the input variables being the climatic characteristics of each month of the five consecutive growing season months (maximum and minimum temperature, precipitation, and potential evapotranspiration). The results of this cluster analysis identified areas designated as "climate clusters" which are areas of highly similar climatology during the growing season (Figure 5-2).

To assist in determining where to locate households for the rapid appraisal farm survey, the climate cluster layer was then merged with the 1990 population density grid for the region (Diechmann, 1997) (Figure 5-2). This process allowed the delineation of areas that could be sampled that were representative of the population in the region, thus allowing polygons having higher population to be sampled more frequently. In order to assign numbers of samples to the climate zones, the total population within each zone was sorted from largest population to smallest. This approach allowed the delineation of 65 sampling zones that were representative of the environment and population of the CRV region (Figure 5-2). These sampling zones served as the areas where farms would be surveyed for the rapid appraisal.

Based on the climatic cluster and rural population density combination, a minimum of 120 household surveys were needed to achieve the desired land area coverage. The proportion of the 120 households in each sampling zone polygon was calculated based on the climatic clusters and the population. Detailed maps overlaid with roads and town were produced, and with help of agriculture officers in the CRV districts, unique landmarks such as churches, schools, cattle dips or a unique physical structure such as a rock outcrop were added in each polygon. The landmarks were numbered, and then using a random number generator, pairs of these numbers were generated. Straight lines were then drawn joining these landmarks forming the sampling transects. Using GPS to locate these transects, the first rural farm or pastoral household encountered was surveyed. If more than one household was needed in a transect, we selected the next 5^th household on the alternate side on the transect until the required number of households was sampled. The survey team was made up of scientists and senior government officers from Ministry of Agriculture, Office of the President, Ministry of Finance and Planning, Ministry of Natural Resources, Kenya Agricultural Research Institute, International Livestock Research Institute and district agriculture and livestock extension officers from various districts. The single visit survey (Appendix XXX) was carried out between January 29, 2001 and February 28, 2001.

After the survey was conducted, data were summarized for analysis to determine representative farms. The following variables were derived from the data collected and served as the primary variables used to distinguish farm types for the region:

The K-means cluster analysis method (SPSS 2001) was conducted using the variables described above. The variables were standardized to overcome problems with varying scales. Five clusters were initially designated for the 99 rapid appraisal observations (4 farms were removed because of problems with land area summations and missing data).

The clusters identified were as follows (See Table 5-2 for descriptive statistics):