By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
 Upstream-downstream water conflicts are common phenomenon in most basins. Such conflicts take a different dimension when they occur during dry season with an already low level of river flows. This context aggravates the spatial externalities on downstream communities and water-dependent ecosystems, causing serious socioeconomic and ecological effects. Utilizing an analytical framework to capture the causes and consequences of the spatial nature of these externalities and relying on both primary and secondary data pertaining to the Upper Ewaso Ngiro North basin in Kenya, this paper (1) shows how these externalities are caused more by upstream land and water use changes than by any climatic and hydrological factors, (2) evaluates the nature and magnitude of their impacts on different downstream community groups, (3) assesses the extent and success of past policy responses and local initiatives, and (4) concludes by indicating the needed approach and interventions for finding a durable solution to the problem of upstream-downstream water conflicts and externalities.
 Like all producers, agricultural producers involved in crop, forestry, livestock, and fishery activities also manage their production systems to maximize output while minimizing production costs. Their actions may conflict or complement with other activities, leading to negative or positive externalities, respectively. Negative externalities are particularly serious, especially when there are no mechanisms to generate enough incentives to internalize their impacts on production and resource use decisions. The issue of developing such mechanisms is much more challenging when the externalities are generated by the joint effects of one group of producers and the impacts are spread over another group of producers. A classic example of such negative externalities is evident in the spatial pattern of water use within river basins. The externalities emerge in this case because the production systems and water use decisions of upstream groups alter the flow rate, quantity, quality, and temporal pattern of water, causing several negative impacts on downstream communities. These consequences are more serious, particularly during the dry seasons and drought years when the vulnerable groups become the main victims of such upstream-downstream water use conflicts [Thomas, 1999; Gichuki et al., 1998].
 The problems of upstream-downstream externalities have been with us for a long time, and numerous strategies for managing externalities have been proposed [Thomas, 1999; Bowers and Young, 2000; Siebert et al., 2000]. Despite the growing literature on the subject, there is still a knowledge gap on how land and water management in upper watersheds can adversely affect downstream communities and what could be done to minimize the impacts of negative externalities, especially in a basin with a high degree of hydrological dependence and vulnerability as well as social, cultural, economic, and ecological diversity. This paper aims to address this important issue using an illustrative case study of the Upper Ewaso Ngiro North Basin in Kenya. It examines externalities and their causes and effects and explores past responses to mitigating them. On the basis of the analysis it also identifies various policy interventions for dealing with the externalities on the basis of promoting agricultural water productivity, water-saving production systems, water storages, and nonagricultural sources of rural livelihoods.
2. Study Area and Analytical Approach
 We begin with a brief description of the study area and the analytical approach used for evaluating the problem of upstream-downstream externalities as induced by man-made alterations in dry season river flow.
2.1. Study Area
 The Upper Ewaso Ngiro North Basin provides the empirical context for this study. It is selected because its physical, socioeconomic, and ecological diversity provides an interesting context to study the upstream-downstream externalities and derive more generic policy interventions for managing these externalities. This can be understood from a brief description of the study area. The study basin is defined by areas draining at the Archers' Post River gauging station. It is a subbasin of the Jubba/Shebelli Basin, an international basin shared by Kenya, Ethiopia, and Somalia. The study basin has an area of 15,200 km2 and falls under seven administrative districts (Nyandarua, Nyeri, Laikipia, Meru, Samburu, Isiolo, and Nyambene) and three provinces (Rift Valley, Central, and Eastern) of Kenya. It is inhabited mostly by six ethnic communities: Kikuyu and Meru (agropastoralists) and Maasai, Samburu, Boran, and Somali (nomadic pastoralists). The total population of the area has increased from 0.5 to 1.3 million people during 1969–1999 [Central Bureau of Statistics, 2000]. Population density varies from 3 to 1150 persons per square kilometer, and ∼72–82% of them are living in rural areas.
 High population growth rates (4–8%) are observed in the upper parts of the basin, where agroclimatic conditions are more favorable for agricultural production. Such a high population growth during the last three decades is a result of the transfer of land ownership from white settlers to cooperative farmers/ranchers with a subsequent subdivision of large holdings into 1–10 ha parcels of farm settlements [Wiesmann, 1998]. As a result, most of the agricultural producers are resource-poor small holders or nomadic pastoralists with low levels of education and limited access to both input and output markets. Differences in elevation, climate, and soil conditions give rise to ecological zones, ranging from humid to arid zones within a distance of 150 km. The mean annual rainfall in the basin ranges from 300 to 1500 mm/yr, whereas the potential evapotranspiration varies from 1000 to 2600 mm/yr. Annual rainfall exceeds annual evapotranspiration only in 1% of the basin area. On average, a water deficit (annual evapotranspiration minus annual rainfall) of >1000 mm/yr is experienced in 51% of the basin area, and in 11.5% of the area the deficit exceeds >2000 mm/yr [Gichuki et al., 1998]. These physical, agronomic, socioeconomic, and ecological diversities present considerable challenges as they involve (1) a multijurisdictional setting with diverse development objectives and weak coordinating mechanisms, (2) a multiethnic and highly diverse society experiencing demographic, political, economic, and resource use transitions, and (3) a highly diverse ecological environment with a growing uncertainty and scarcity of water resources.
2.2. Analytical Framework
 In order to understand the magnitude and consequences of upstream-downstream externalities we take a broader view of the upstream-downstream situation. That is, we consider downstream community to include all people living downstream of the point of action, ranging from the immediate neighbor to those living in the downstream areas of the study basin. As required by the context of the study basin, we focus only on the negative externalities. Using the cause and effect relationship proposed by Bowers and Young , we try to trace the linkages between the actions of the upstream resource users and the impacts on the downstream users, including the environment. This framework is presented in Figure 1. As can be seen, we group the land and water use activities that contribute to reducing the quality and quantity of river water in the dry season into three categories: (1) those that reduce groundwater recharge and its contribution to dry season flow, (2) those that increase river water depletion, and (3) those that contribute to river water pollution. The resultant externalities are grouped into four categories: (1) reduction and pollution of dry season river flow, (2) impacts on agriculture-based livelihoods, (3) impacts on rural nonagricultural livelihoods, and (4) impacts on environment and biodiversity.
 Within this framework we assess the upstream-downstream externalities in terms of various socioeconomic impacts on the affected communities. The externalities considered here include both tangible (pecuniary in nature with market values) as well as intangible (real in nature with no market values). The valuation of tangible externalities is imputed from the increase or reduction in revenue or costs associated with the externality [Bowers and Young, 2000]. The valuation of intangible externalities was not undertaken because of lack of data.
2.3. Data and Analysis
 For analyzing the problem of upstream-downstream externalities we rely on different but related sources of information. We utilize the hydrometeorological data collected during 1960–2000 to analyze river flow and rainfall trends in the study basin. River flow data at the basin outlet, i.e., the Archers Post River gauging station, are analyzed to assess the nature and extent of declining low flows in the lower reaches of the basin. Similarly, corresponding rainfall data from selected stations located in the basin for the same period are used to assess the contribution of rainfall variations to the decline in dry season river flow. We have used the secondary data available in Natural Resources Monitoring Modeling and Management (NRM3) database as well as information available from existing literature related to the basin to document the impacts of upstream population growth, land use changes, surface water development and abstraction, and agricultural intensification on dry season flows. We also utilize the information collected through questionnaire and group discussions during the 2000 drought period to assess (1) the water problems associated with agricultural growth, (2) the externalities resulting from agricultural water use and management, and (3) the past policy responses and the required future interventions.
3. Declining Quantity and Quality of Dry Season Flow
 The root cause for all forms of upstream-downstream externalities in the context of the study basin lies in the hydrological phenomenon of declining dry season flow as well as its impact on water quality. Therefore, in order to understand well the nature and severity of the impacts of these externalities on down stream communities it is necessary to examine the main causes for and the true magnitude of the problem of declining dry season flows.
3.1. Major Causes
 The problems of the declining quantity and quality of dry season flows available for downstream areas are contributed by the hydrological alterations due to factors such as landscape changes, increasing dry season irrigation, and urban expansion in the upstream areas of the basin. How these factors reduce dry season flow is described in sections 3.1.1–3.1.3.
3.1.1. Land Use Changes and Reduction in Groundwater Recharge
 Land use changes alter the hydrology as they affect not only how rainfall is partitioned between runoff and infiltration but also how soil water is partitioned into plant water use and deep percolation. These changes therefore directly influence groundwater recharge and dry season river flows. Runoff plot experimental data in forested catchments show that change from natural forest to plantation forest increased runoff by as much as 17% during the 4–6 year period between tree harvesting and the establishment of next plantation. This increasing runoff correspondingly reduces the level of groundwater recharge. Agroforestry trees yield mulch material, which if effectively applied on cropland, can enhance infiltration and soil water stocks [Liniger et al., 1998]. However, it also true that the agroforest tree species have a higher annual transpiration than annual crops grown in semiarid areas. Therefore agroforestry trees deplete more water from the soil water stocks than crops, causing a reduction in the proportion of rainwater that would have percolated to recharge groundwater. It is estimated that on average, agroforestry trees deplete an additional 180 mm over what a crop would deplete per year. This amounts to an additional depletion of 117,000 m3 of water annually over the entire basin, part of which could have supplemented the dry season flows.
 The way the rain-fed areas are managed also has major implications for the level of dry season flows. The area under rain-fed crop in the basin has increased from 500 to 13,500 ha during the last 40 years. In well managed fields, runoff is reduced considerably, resulting in increased infiltration and higher soil water stock. Analysis of cropland runoff plot experiment data shows that on average, an additional 53 mm percolates below the root zone in a well-managed rain-fed cropping system. In contrast, in poorly managed fields the runoff rates are 11–36% higher, contributing substantially to flood flow, but very little to groundwater recharge. Infiltration studies in semiarid areas of the basin show that the mean final infiltration rate for bush land sites with good ground cover and high soil organic matter content (250 mm/hr) was ∼1.5 times that of grassland sites (160 mm/hr) but 4 times that for bare field (65 mm/hr) [Liniger et al., 1998] (see also NRM3 database).
 Degradation of pasturelands in the basin area also has negative effects on rainfall infiltration, groundwater recharge, and dry season flows. Livestock population in the study area has increased from 280,000 to 566,500 tropical livestock units (1 tropical livestock unit is livestock equivalent to 1 cattle with a body weight of 250 kg) during 1969–1999 [Muchoki, 1998] (see also NRM3 database). To support this increase in livestock population, bush lands have been cleared to promote grass production. In communal grazing lands, as much as 30% of the grazing lands have lost their grass cover because of overgrazing. This has resulted in reduced rainfall infiltration and groundwater recharge. Since most of grazing lands have very deep groundwater levels (>100 m), high potential evaporation (>800 mm/yr), and low rainfall (<500 mm/yr), their potential contributions to groundwater recharge and dry season river flow are insignificant. There are also man-made physical and hydrological alterations within the river itself. For instance, sand mining for construction activities reduces the quantity of subsurface water available in the river bed by damaging subsurface storage capacity and enhancing evaporation from river channels. It is estimated that sand-mining activities have lowered the ephemeral river bed by 0.5–2 m (see NRM3 database).
3.1.2. Increasing Dry Season River Water Withdrawal
 Land use changes and urban expansion in the upstream catchments of the basin have the most direct as well as the highest impact on the level of dry season flow available for the downstream users including the wildlife-based ecosystems. Water diversions for irrigation top the list followed by that for domestic consumption in urban areas. The upstream irrigated area has increased from 227 to 4088 ha during 1969–1999, whereas the irrigated area downstream of the basin outlet remained ∼217 ha (see NRM3 database). Nonirrigation river water use in 1999 was estimated at 5866 m3/d, of which livestock accounts for 51.6%. Thus the total water diversion for both irrigation and nonirrigation uses during the dry season of 2000 was reckoned to be 65,836 m3/d, with ∼91% being used for irrigation. This is an extremely high level of water withdrawal considering the fact that the dry season naturalized river flow at Archers' Post is lower than this amount 78% of the time.
 In terms of prevailing regulations, irrigated agriculture is not permitted to abstract river water during the low-flow period that generally occurs during the dry season. However, illegal abstractions do occur extensively, especially in the upper catchments of the basin areas. A snapshot survey of irrigation water withdrawal during the 2000 dry season showed that ∼1000 hectares were supplementary irrigated using unauthorized withdrawal of dry season river water. Most of these illegal abstractions from the river are made by irrigators who have not constructed the 90 days water storage works needed to sustain irrigation during dry season. These on- or off-stream water storage works are mandated by government regulations, and these storages are supposed to be filled by abstractions during the flood flow period. Irrigation water demand is further aggravated by the high conveyance (35–55%) and application (30–67%) losses [Gichuki et al., 1998].
 There are a variety of reasons for the expansion of illegal water abstractions. They are (1) the perception that water is a free and abundant resource that is wasted or unproductively utilized by downstream communities, (2) inadequate monitoring of river water withdrawals and the differential interpretations of the terms of water permit particularly those related to what is flood, normal, and low flow conditions in a given river reach, (3) failure to comply with the mandatory 90-day irrigation water storage, and (4) inadequate law enforcement and low fines for unauthorized water withdrawal.
3.1.3. Water Pollution
 Rapid urban expansion and increasing chemical application in irrigated areas also has a severe impact on the quality of dry season river flow. Besides the return flows from urban centers and irrigated areas the washing of trucks involved in sand-mining activities also causes oil spills. As a result of all these urban, chemical, and petroleum-based pollutants the quality of dry season flow is low, especially in the downstream areas of the basin.
3.2. Magnitude of Decline in Dry Season Flow
 The dry season flow in the basin is generally low because of the low rainfall and high evaporation. The problem gets further aggravated by the human-induced factors that are described in section 3.1. In this section the magnitude of the problem is indicated on the basis of available information on precipitation and hydrology. The mean, minimum, maximum, and standard deviation values for the monthly river flow observed at Archer's Post are 21.20, 0.10, 650, and 40.20 m3/s, respectively. Analysis of monthly river flow trends, although indicating the general hydrologic response to climate variability, has not indicated any decline in rainfall during the high rainfall months of April, August, and November. In contrast, the low-flow periods of January–March and June–September have experienced a 56 and 35% reduction, respectively. Notably, the lowest flows for 7, 90, and 120 day durations were observed during the 1980–1986 period when there was the longest consecutive period of below normal rainfall in the basin. The temporal pattern of low flows corresponds closely with the droughts experienced during 1974–1976 as well as during the years 1981, 1983, 1984, 1994, and 2000.
 Information from an analysis of flow data as well as from the interviews conducted with downstream water users have established that the Ewaso Ngiro North river dried ∼60 km upstream of Archer's Post River gauging station during the years 1984, 1986, 1991, 1994, 1997, and 2000 (see NRM3 database). Springs located in the game reserves near the study basin outlet yield 0.2–0.6 m3/s that sustain the flow to an average of 123 km downstream of Archer's Post River gauging station, even when the river dries up upstream of the gauging station. The number of days with flows at Archer's Post <1 m3/s has also increased over the years (see Figure 2). Analysis of year-to-year and long-term variability of annual rainfall (1925–2000) across the basin showed that there is no clear trend of decreasing or increasing rainfall, but there are clear fluctuations between periods of above and below average rainfall. This means that the decline in dry season flows, although contributed to by low rainfall in the drought years, can be attributed mainly to land uses changes and increasing water extraction prompted by agricultural intensification, livestock expansion, and urban growth. Thus it is the man-made alterations in upstream landscape and hydrology rather than the natural factor of rainfall that contribute more to the problem of declining quantity and quality of dry season flow in the study basin.
4. Impacts of Upstream-Downstream Externalities
 Although the externalities associated with the upstream activities initially manifest in the form of the hydrological phenomenon of the declining dry season river flows, their socioeconomic and environmental impacts are very severe on downstream communities and ecosystem that rely on dry season river flows for their livelihoods and sustenance. The affected groups include not only the downstream pastoralists and irrigators but also the hoteliers and tour operators relying on downstream wildlife sanctuaries. Table 1 indicates how these different groups suffer from the problems of upstream-downstream externalities. Here we also attempt to provide some estimate for the economic and environmental losses caused by these externalities.
Table 1. Impacts of Water-Related Externalities in the Ewaso Ngiro North Basin, Kenya
Nature of Externality
Downstream nomadic pastoralists
economic loss associated with loss of livestock resulting from water shortage and the long walk to watering point social impacts associated with the herders (men) moving far away from their families in search of livestock water and/or grazing resources conflicts arising from nomadic pastoralist livestock moving into private land additional cost/labor needed to fetch drinking water for families dependent on river water
Hoteliers and tour operators
loss of revenue associated with wildlife moving out of traditional viewing areas additional costs incurred in developing water resources that would retain the wildlife in the traditional viewing areas
loss of revenue associated with reduced irrigated area and crop yield additional cost incurred in developing additional water resources to improve water supply in dry season
Table 2 summarizes the economic losses associated with the negative externalities. The reduction in irrigation areas in the lower reaches of the basin varies from 60 to 100%. Since the magnitude of reduction of irrigation has varied from 131 to 217 ha, there is an economic loss equivalent to the monetary value of the loss of output from this area. Given the cropping pattern, low level of input use, and unfavorable market conditions prevalent in the downstream areas of the basin, the magnitude of the economic loss of irrigation reduction can estimated to be 260–434 tons of maize per year or US$16,000–27,000 per year; losses of livestock production benefits attributed to decline dry season river flows are difficult to quantify partly because during this period, livestock are weak and generally suffering from feed deficiencies. However, the pastoralists reported that on average, a pastoralist may lose 2–5 livestock because of water constraints (see NRM3 database). On the basis of this consideration it is possible to make a conservative estimate of a loss of 60–100 cattle for the pastoralists who rely on the riparian zone of the downstream. The value of this loss per major drought event can be in the range of US$7700–12,820. On the other hand, people relying on fishing activities in the downstream regions have reported to have lost fish production with an estimated value of US$5300 because of the drying up of the downstream reaches (see NRM3 database).
Table 2. Economic Losses Caused by Water-Related Externalities in the Ewaso Ngiro North Basin, Kenyaa
Quantity of Loss
Value of Loss in U.S. Dollars
Source: Natural Resources Monitoring Modeling and Management database.
Irrigated agriculture area reduced by 131 to 217 hectares
260–434 tons of maize
Fetching domestic water
800 man days of labor
 The tourist industry reports both positive and negative externalities associated with declining dry season flows. The nature and magnitude of these externalities depend on the low flow levels. A slight decline is considered to have a positive externality as the wildlife move to man-made watering points strategically located to maximize wildlife viewing. This increases the chances of attracting more tourists to the game reserves and therefore to the hotels. However, when dry season river flow declines below a certain threshold level, wildlife population are forced to move upstream, and the number of tourists visiting game reserves in the lower parts of the basin obviously declines. The game park attendants and hotel managers reported that on average, the number of tourists has dropped by 5–10 per day over a period of 25–60 days during the 2000 dry season flow period.
 Besides its negative effects on downstream tourism industries the movement of wildlife upstream also leads to similar effects on upstream farms and communities as such movement often leads to wildlife-human conflicts, resulting in the destruction of fences, watering points, crops, and in some extreme cases, loss of human and animal lives. Decline in dry season flow also has serious effects on downstream ecosystem and biodiversity. These effects are due to (1) the encroachment of wetlands for dry season grazing and crop production and (2) overgrazing in game reserves attributed to the movement of livestock into game reserves during the dry season. The downstream communities that depend on dry season river flow as their primary source of drinking and livestock water reported that they had to incur extra costs in fetching water by having to dig shallow wells in the river bed and for lifting the water an additional 1–3 m.
5. Policies for Managing Externalities
 There were a number of policy responses to the problems emerging from upstream-downstream externalities and conflicts. Before dealing with them, it is also necessary to understand how some of the past policies and emerging economic and political conditions have also contributed to these externalities and conflicts. The policy factors, though strictly exogenous to the basin, have irreversible effects on the land, water, and forest resources of the study area. Some of the major policies are noted below. First, the settlement policies of the 1970s have encouraged migration into the basin, causing a 5–8% population growth. Second, the expansion of commercial forestry, as induced by an increasing demand for timber products, has led to a widespread conversion of natural forest areas into plantation forest lands. Finally, the agricultural and irrigation policies have led to an increase in irrigated area from 227 to 4088 ha during 1966–1999. These policies were driven both by the need to ensure household food security in the basin areas as well as by the desire to take advantage of the high returns in horticultural activities with attractive export markets. Besides these policies, there were also other factors that have intensified the upstream-downstream resource use conflicts over the years. These include (1) cattle rustling and related insecurities, which constrained the use of grazing resources in the lower areas of the study basin and intensified the pressures on grazing and water resources in upper and middle part of the basin, (2) the limited opportunities for nonfarm sources of livelihood that increased the dependence on agricultural activities, (3) high levels of poverty that constrain investments in enhancing natural resources productivity, and (d) political transition to multiparty democracy that has widened the political differences between the upstream and downstream communities.
 The government, community, and individual responses to the externality problems have been largely determined both by the emerging socioeconomic, resource, and institutional realities as well as by the evolving perceptions of basin resource users. While these responses are many and varied, we highlight here the interventions aimed at protecting water catchments and reducing river water depletion. The Government of Kenya (GOK) identified water catchments areas that are to be kept under natural vegetation with limited human interference so that water yield can be sustained [Ministry of Water Development (MOWD), 1992]. With increasing population pressure some of these water catchment areas were converted into plantation forest or cropland. Over the years, the pressures from downstream communities and environmental lobbyists have led to a change in government policies, including the empowerment of state environment protection agencies. Such a change in government policies has been quiet effective as it has reduced the encroachments in water catchments to acceptable levels. The Kenyan government, with support from donors and cooperation from land owners, has also created a soil conservation movement in the 1980s that has eventually slowed down the rate of soil erosion in the basin areas [Tiffen et al., 1994]. Fortunately, with increasing value of farm output, farmers have also started investing in soil and water conservation. As a result, soil erosion in croplands is now a major concern only in localized areas, though it is still a serious problem in communal grazing areas.
 During the 1970s and 1980s, the government focused on promoting irrigation expansion and rain-fed cultivation in the arid and semiarid areas as the main strategy to enhance household food security and counter urban migration. The initial attempts for improving agricultural productivity in rain-fed and irrigated areas were centered on farm-level strategies for resource use and cost minimization and output maximization. As a result, the anticipated positive externalities from improvement in irrigation efficiency, i.e., the releasing of water saved to downstream communities, have not occurred because the upstream irrigators retain the saved water to use for expanding their irrigated area. During 1980s, the water-related negative externalities were not considered as an outcome of upstream agricultural intensification but were perceived as a natural drought phenomenon. However, studies carried out by Laikipia Research Program highlighted the nature and extent of informal irrigation expansion through illegal water withdrawals during the low flow period [Gichuki et al., 1998]. When the problem of illegal irrigation was recognized, the government adopted a policy of banning irrigation during drought periods. Such a ban was in place during the 1991, 1994, 1997, and 2000 drought periods.
 In the mid-1990s, the government initiated water sector reform [GOK, 1999]. This reform and the process of organizational decentralization and the transition to a multiparty democracy have enhanced the role of water users in governance arrangements at various levels. For instance, water users associations (WUAs) have became proactive in addressing problems associated with declining dry season flows. Before 1997, WUAs were mostly project-based with a main focus on water management within water projects. During the 1997 drought, as the downstream reaches of the river dried up, downstream communities marched upstream to physically force irrigators to stop or reduce water abstraction. This phenomenon led to the creation of river reach WUAs with a larger mandate and responsibilities which include (1) negotiating with the government on how much water to flow at strategic river gauging stations, (2) ensuring that minimum dry season flow targets are met, and (3) facilitating negotiated allocation of bulk water across projects in the river reach. Besides the involvement of water users other equally important institutional arrangements have also emerged to strengthen the participatory framework needed for managing the upstream-downstream resource use conflicts and externalities. For instance, the hydrometeorological and water use monitoring program operated and maintained by a partnership consisting of the MOWD, Laikipia Research Project, and NRM3 Project (jointly implemented by the University of Nairobi (Kenya) and the University of Berne (Switzerland)) (1) provides the data needed to assess the status and trends of dry season flows, (2) evaluates communities needs, attitudes and perception on water resource issues and shows how they influence their decisions, and (3) enhances wider public awareness on the need of and methods for dealing with water-related problems [Mathuva, 1997].
6. Conclusions and Recommendations
 While the policy responses and other user-based initiatives are significant, they have not been effective enough to mitigate the problem of declining dry season river flows and the attendant water-related conflicts and externalities [Namunane, 2000]. Most downstream communities are dissatisfied with the efforts made so far in controlling upstream water withdrawals. There is a general perception that the government should also address the problem through development and safety net policies for (1) developing additional water resources, (2) enhancing alternative rural livelihood options through industrial and infrastructural development programs, (3) protecting farmers from the ill effects of liberalization and privatization, and (4) enhancing food security. As current interventions are inadequate in terms of coverage and effectiveness, there is an urgent need for developing more comprehensive approach for solving the externality problems in particular and poverty issues in general. The required technological and management interventions are summarized in Table 3.
Table 3. Required Technological and Management Interventions in the Ewaso Ngiro North Basin, Kenya
Required Technological and Management Interventions
Reducuction in degradation of land and water resources
soil and water conservation practices, appropriate land use and management, minimizing salinization and water logging, slowing conversion of forest and swamp land into cropland, and channeling polluted water to sinks
Developing additional water resources
water storage technologies, groundwater exploration, water lifting technologies, zoning suitable communal dam sites, and enhancing groundwater recharge
Achieving fair and equitable allocation of water resources
water abstraction structures improved, flood flow storage, community participation in catchment-level water allocation, and allocating water to marginalized communities
Reduction of water loss
conveyance and application technologies, providing better irrigation scheduling, reducing evaporation of water stored in the soil and dams, reducing water lost into unrecoverable deep percolation or surface runoff, reusing return flows, adopting deficit, and supplemental and precision irrigation
Enhancing marketable agricultural output
improving cropping pattern, growing higher-yielding varieties, switching from high- to less-water-consuming crops, and alleviating soil, water, nutrient, pest, and diseases constraints to crop production
Enhancing financial returns
reallocating water from low- to higher-value uses and users, reducing production costs, enhancing product value through postharvest technologies, and alleviating transportation and marketing constraints
 From the perspective of the magnitude and sustainability of the impacts of development interventions in downstream and other poor areas the time and method of their implementation are as critical as their focus and coverage. For instance, the general policy interventions aiming at improving irrigation efficiencies were implemented with a hope to save water in upstream areas and move it to downstream communities. However, in view of their limited nature, water use efficiency and agricultural water productivity have increased but not water equity as the saved water was used to expand irrigation in the upstream area itself. Considering how uncertain the rainfall pattern in the basin area (heavy floods during 1997/1998 was followed a severe drought during 2000) is, catchment management programs are as important as the programs for enhancing water storage capacity. Although upstream dam sites are more cost-effective because of hydrologic and topographic factors, benefit-wise, downstream dams are more important in view the socioeconomic and ecological benefits.
 The past experience and future requirements of the basin suggest that there should be a two-pronged approach: one that enhances the incentives of upstream resource users to internalize the externalities they generate and the other that enhances the capacity of downstream communities to cope with these externalities. Continuing dependence on the fragile natural resource base under conditions of expanding population with limited alternative livelihood avenues will tend to deplete resources and aggravate the impact of externalities. While the development of industrial, infrastructural, and other nonfarm activities are clearly required, what is more important is the framework for their development because these activities cannot be developed within an economic and institutional vacuum. Such framework should be rooted in investment partnerships (state, private, and international donors) and organizational partnerships (upstream and downstream communities, research organizations, government departments, and local agricultural producer organizations). Such a framework can facilitate investment, empower local communities and user groups, and enable the development of institutional arrangements for effective program planning, implementation and monitoring in the study basin. While there are intentions and initiatives for developing such a framework, they have to go a long way in providing a durable solution to the upstream-downstream resource use conflicts and their socioeconomic and ecological impacts.
 This paper is drawn from a larger study on the Upper Ewaso Ngiro North river basin. The funding support of the Rockefeller Foundation and Swiss Development Cooperation, the research, and field support are acknowledged as are the valuable comments and suggestions received from the reviewers of this journal article.