Assessing the potential of harvesting rainfall runoff using pans to meet Marigat's water demand: an application of WEAP21 model

Kenya is among the ‘water-scarce’ countries of Africa (Futi et al. 2011; www.nema.ke.org) and faces extreme variations in climate particularly the Rift Valley due to its various landforms, UNEP (2010). Marigat Division in Baringo County located in an arid and semi-arid region is among such areas that are facing water scarcity. The average annual rainfall is 650 mm with weak bimodal peaks recorded from March to May and June to August. There is one rainy season from end of March to August and a prolonged drought. The rainfall is about 50% reliable and is strongly influenced by the local topography. The mean annual temperature of the area is about 30 °C and occasionally rises to over 35 °C. The period between January and March is the hottest (BDVS 2005–2015). There is plenty of water during rainy season most of which go to waste and frequently causes havoc due to lack of appropriate rainwater harvesting technologies (BDVS 2005–2015). The wastage of water during rainy season can be salvaged with improved storage and rainwater harvesting methods and such water can be used during periods of water shortages.

As climate change leads to more extreme variations, water harvesting solutions must cope with both extreme rainfall and extreme droughts. Extreme rainfall requires good flood protection and diversion structures while extreme drought requires large storage capacity (NWP 2007). To respond to water scarcity, rainwater harvesting techniques provide a direct solution especially in rural and drought prone areas.

Integrated Water Resources Management (IWRM) is a holistic approach appropriate for planning and managing scarce water resources. IWRM tools that support the planning and management of water resource processes have become more common, but frequently generic tools that can be applied to different basin settings are difficult to use because of the complex operating rules that govern individual water resource systems. These IWRM tools must be useful, easy-to-use, and adaptive to new information and stakeholder priorities, and Water Evaluation and Planning (WEAP21) model is one of the tools (Yates et al. 2005; Yilmaz & Harmancioglu 2010).

The WEAP21, IWRM model is a surface and ground water tool based on water balance accounting principles, which can test alternative sets of supply and demand conditions. WEAP21 addresses water planning and resource allocation problems and issues (Yates et al. 2005). It integrates a range of physical hydrologic processes with the management of demands and installed infrastructure in a seamless and coherent manner (Yates et al. 2005; Akivaga et al. 2010). Both the engineered and bio-physical components of a water system are represented to facilitate multi-stakeholder water management dialogue on a broad range of topics, including sectorial demand analysis, water conservation, water rights and allocation priorities, reservoir operations, hydropower generation, pollution tracking, ecosystem requirements, and project benefit–cost analysis (Akivaga et al. 2010).

In response to water scarcity in Marigat, the study adopted WEAP21 to determine the potential of rainfall runoff harvesting using water pans to meet water demand for Marigat community and propose a water supply network for Marigat community. The study assesses the potential of surface water harvesting using pans, and results can be used in planning strategies of curbing water scarcity problem in Marigat Division, Kenya.

Marigat division is one of the 14 divisions in Baringo County. The county lies between latitudes 0° 12′ and 1° 36′N and longitudes 35° 36′ and 36° 30′E. The soils are mainly clay loams with alluvial deposits derived from tertiary/quaternary volcanic and pyroclastic rock sediments that have been weathered and eroded from the uplands (Mwangi & Swallow 2005). The soils are fertile but high evapo-transpiration rates and low variable rainfall create water scarcities that limit intensive agricultural use. The major topographical features are river valleys, plains, and the floor of Rift Valley. The area is on Loboi plain and is characterized by rolling slopes that range from 5 to 25% towards downstream of the rivers. The location of the study area is given in Figure 1 and the drainage network is given in Figure 2.

The main land use practices in the study area are pastoralism, agriculture, fishing, tourism, and settlements. The land use information was identified from satellite imagery. The drainages in the study area consist of rivers, Endao, Molo, and Perkerra, and streams (Figure 2). The spatial distribution of the locations of water demand and supply points was determined using hand-held GPS and GIS techniques.

Historic monthly rainfall data for the past 39 years from 1970 to 2008 were obtained from Kenya Agricultural Research Institute, Marigat station and used in simulation. Conceptual framework for water planning in the study area was developed and subsequently customized in WEAP21 model. The river system, boreholes, and water pans were schematized from an ArcView GIS layer. This information was obtained from different water users within the Marigat sub-catchment through questionnaires and individual interviews. This information was used to determine the extent to which the water stored in the proposed reservoirs can meet the water demand for the community.

WEAP21 system is a demand, priority, and preference driven water planning model that aims at closing the gap between water management and catchment hydrology by addressing both bio-physical factors influencing the river and socio-economic factors affecting the level of domestic, agricultural, and industrial demand and management of artificial reservoirs (Rochdane et al. 2012). WEAP21 is based on the basis that defines water supply as the amount of precipitation that falls on watershed processes, human demands and interventions, or enhanced through watershed accretions (Yates et al. 2005; Mutiga et al. 2010). The water demand was classified as domestic, livestock, and agriculture with reducing order of allocation priority, respectively (Water Act 2002).

River flows from Molo, Endao, and Perkerra rivers and other seasonal streams are a source of water in the community to be supplemented with water supply from the proposed water pans in this study. Observed river flow data for Marigat Bridge station (Figure 2) were available as gauge heights and were converted using the rating curve to get the flows in cubic meters per second. The observed flows were used to calibrate and validate the model.

The total consumptive water requirement was obtained from the number and sizes of households and livestock in Marigat based on 2009 population census, with a total of 8,828 households. A unit water requirement of 50 l per person per day (FAO 2007; www.un.org) and 50 l per day for each livestock was used for WEAP21 domestic water demand calculations.

The key principles of the Kenya Water Act (2002) are sustainability and equity (Mutiga et al. 2010). In using water resources to promote social and economic development, it is essential to protect the environment while ensuring that the water needs of present and future generations can be met. This is partly achieved by leaving enough water in a river, referred to as the ‘reserve’, to protect aquatic ecosystems in order to secure ecologically sustainable development and use of the water resource (Water Act 2002).

The water demand for agriculture was estimated by multiplying the total area under irrigation with the average water requirement for the main crops that is from Perkerra irrigation scheme (Table 1). Irrigation water demand for the basin was calculated using the reference evapo-transpiration (ETo) and effective precipitation (P) concept as outlined in FAO-56 (Mutiga et al. 2010) since there were no data available on the exact amount of water used for irrigation.

The rainfall runoff method was used to simulate river flows, it was chosen since it best suited the characteristics of the study area. The type of data required to perform rainfall runoff simulation included: land use (area, Kc, effective precipitation) and climate (precipitation and ETo). Where Kc is the crop coefficients and ETo is the reference crop evapo-transpiration.

A schematic diagram of the WEAP21 model for the Marigat division in Kenya (Figure 3) shows all the demand sites and various water sources (water pans, boreholes, and rivers).

This process was guided by the priority as described in the Water Act (2002), which ranks the reserve and domestic water requirements above other uses, reserve means quantity and quality of water required to satisfy basic human needs for all people who are or may be supplied from the water resource and to protect aquatic ecosystems in order to secure ecologically sustainable development (Water Act 2002), thus, domestic and environmental flows were assigned the highest priority over all other water uses and must strictly be met before water resources are allocated to any other uses. Domestic water use and environmental flows were given priority (1) as shown in Table 2. This satisfied water demand downstream according to the water allocation hierarchy in the Water Act (2002).

WEAP model was selected for this study as it incorporates all these values into a practical tool for water resources planning (Mutiga et al. 2010). A schematic diagram of the WEAP model for the Marigat Division (Figure 3) shows all the demand sites and various water sources (streams, boreholes, and water pans).

The water year method that is in-built in the model WEAP was used in the predictions of hydrological variables based on the analysis of historical rainfall. The method uses the statistical analysis to identify the coefficients, which is used for future projection of the type of year: wet, normal, or dry. Rainbow model (Raes et al. 2006) was used in frequency analysis of historical rainfall to come up with the type of year.

A scenario can be defined as a plausible description of how the future may develop, based on a coherent and internally consistent set of assumptions about key relationships and driving forces (Arranz & McCartney 2007). Scenarios are built and then compared to assess their water requirements, costs, and environmental impacts (Akivaga et al. 2010). All scenarios inherit data from the current accounts year.

The scenarios can address a broad range of ‘what if’ questions, such as: What if population growth and economic development patterns change? What if reservoir operating rules are altered? What if a water recycling program is implemented? What if climate change alters the hydrology? ‘What if’ scenario analyses were built and done for 2009–2020 (Figure 4). The demand and supply sites were identified and using WEAP21 model, the demand and supply points were linked to come up with a water supply network in the community.

The reference scenario is the scenario in which the current situation, current account year as 2008, is extended to the ‘future’ (2008–2020). The reference scenario is helpful in understanding the current trend or the real situation in relation to water resources management in the catchment (Mutiga et al. 2010). The model mimics reality over the period of 2008–2020. The accuracy of the model is assessed by comparing observed and simulated results (Hoff et al. 2007). In calibrating the model, the observed stream flows and simulated stream flows of the reference scenario at Marigat Bridge were used to draw the line of best fit (Figure 5). The observed stream flows for the period 2008 were used due to limited data to calibrate the model. The results presented in Figure 5 indicate that the flows show the true presentation on the ground. The analysis was done where the mean error (ME) is 0.18 m³, the mean square error (MSE) is 3.86 m³, and the coefficient of efficiency (EF) was found as 74.8%. Though the magnitudes of the ME and MSE are high, the EF indicates that the model is good. The model results as shown in Figure 5 had an R squared (R²) value of 99.5%.

The calibrated model was used in simulating the unmet demand for water and the results are presented in (Figure 6). This result is of importance for planning purposes, because water shortages normally occur between November and March of most years that is during the dry periods in the study area. Further results indicate that some sub-locations in the study area such as sub-locations Kapkuikui, Perkerra, and Kaptombes and livestock in sub-locations Yatoi, Sintaan, Eldume, and Sandai as shown in Figure 6 experience water shortage mainly due to high population of humans and livestock while the water resources are limited for instance there is one water pan and a borehole supplying water to the whole sub-location.

Some sub-locations within the study area for instance Kimalel, Marigat, and Chelaba have a number of water sources to meet domestic and livestock water demands for instance streams, rivers, water pans, and boreholes and during the rainy season they harvest water; this is the reason why there is low unmet demand in such areas. WEAP21 allocates water to competing demands based on the physical system characteristics as well as user-defined criteria, so that coverage at all competing demand sites are equal (Al-Omari et al. 2014).

Marigat Division has been experiencing water shortage especially during the dry seasons. The study proposes creation of new water pans in sub-locations with unmet demand. Five new water pans were proposed in Marigat (2), Sandai (1), Kapkuikui (1), and Kimalel (1) locations. The calibrated WEAP21 model was used in simulation of creation of new water pans scenario that looks at the effect of having new water pans to meet the unmet demand. During the simulation, the demand sites and hydrology remains as in the reference scenario. The water demand increases annually because of the population increase for both humans and livestock.

Through and iterative model simulation process, results indicated that the new water pans are situated in areas where water shortage is experienced, and the pans are estimated to have a storage capacity of 50,000 m³ each. The unmet demand in the study area for domestic and livestock for the simulated period as compared to the results obtained from the reference scenario is met. The unmet domestic demand for Zarqa city also dropped to zero for the year 2050 due to the implementation of the Disis project, which was expected to start operation in 2013 (Al-Omari et al. 2014). Mutiga et al. (2010) incorporated two dams in the scenario and the result showed that building of dams would reduce the unmet water demand by about 5%.

Based on the results of new water pans scenario and existing water sources, a water supply network was proposed for the community. Demographic and water use information was used to construct a scenario that examine how total consumption of water evolved over time and if creation of new water pans will be able to meet the unmet demand. These demands scenarios were computed in WEAP21. Demand analysis is central to integrated water planning analysis with WEAP21, since all supply and resource calculations are driven by the allocation routine, which determines the final delivery to each demand node, based on the priorities specified by the user (Yates et al. 2005). A demand scenario comprised of several sectors including households, livestock, ecosystem, and agriculture.

Individual demand sites and reservoirs requirements were assigned a priority number, which is integers that range from 1 (highest priority) to 99 (lowest priority). Similar to demand priorities, supply preferences applied an integer ranking scheme to define which sources will supply a single demand site. The WEAP21 that applies IWRM principles incorporates a demand priority and supply preference approach to describing water resource operating rules, as system demands drive the allocation of water from surface and groundwater supplies to the demand centers. The water allocation problem is solved at each time step using an iterative, linear programing approach that introduces the concept of equity groups (Yates et al. 2005).

The study took into account criteria representing the views and values of different stakeholders, the process by which the model selected water pans sites is suitable for other case studies, which require multi-stakeholder engagement and community participation. Participatory approaches are complimentary, to decision support tools such as WEAP. Figure 7 shows the schematic diagram of the proposed water supply network. It was done by linking the demand and the supply in the model. For the areas facing water shortages, new water pans were created in the scenarios so as to meet the unmet demand for humans and livestock.

The water resource is scarce in Marigat, and there is need to harness surface water generated from rainfall runoff that occurs during the rainy season. In addition, the water resources are sparsely distributed although some of the sources run out of water and others are quite reliable, women and children have to walk long distance to fetch water. From the findings, it shows that harvesting of surface water runoff in water pans is a solution to water scarcity in the area. From the results, WEAP21 model, it shows domestic and livestock water demand increases due to population increase and from building of a scenario, that is, creation of new water pans in areas with unmet demand, it shows that all the unmet demands of domestic and livestock are met up to 2020. This has also been shown by linking the demand and supply to come up with the water supply network. It also shows that WEAP21 model is a practical tool for planners for water resource planning and management.

Since the community is facing water scarcity, it is highly recommended that water pans should be constructed in areas facing water scarcity and used to harvest surface runoff. It is also recommended that the community should come together and harvest surface water that occasionally causes floods and displace them during the rainy season in water pans to curb their problem of water scarcity.