Optimization of rainwater harvesting system design for smallholder irrigation farmers in Kenya: a review

The adverse effects of climate change on agriculture have been felt across the globe. Smallholder farmers in sub-Sahara Africa are particularly more vulnerable to the effects of climate change leading to loss of income and livelihood thus affecting global food security. Rainwater harvesting (RWH) is emerging as a viable option to mitigate the negative effects of climate change by supporting rain-fed agriculture through supplemental irrigation. However, smallholder farmers are still grappling with a myriad of challenges hindering them from reaping the bene ﬁ ts of their investment in RWH systems. This review explores some of the factors behind the poor performance of RWH systems in Kenya and also seeks to suggest techniques that can be applied to optimize the design parameters for improved performance and the adoption of RWH systems. According to the review, RWH has the potential to mitigate the adverse effects of climate change among smallholder farmers. It allows for crop production beyond the growing season through supplemental irrigation. However, their impacts have been minimal due to the consistent poor performance of RWH systems. This is attributed to inef ﬁ ciencies in design and construction brought about by lack of required technical skills among RWH system designers and implementers. Proper design and implementation are therefore paramount for better performance and adoption of RWH systems in the region. This will ensure that RWH systems are reliable, technically and economically feasible as well as possess a desirable water-saving ef ﬁ ciency.


INTRODUCTION
RWH in this context can be described as 'a method of inducing, collecting, storing & conserving local surface runoff for agriculture production' (Ibraimo ). RWH has the potential of increasing rain-fed agriculture production as well as reducing effects of soil erosion. It also allows crop production beyond the growing season through supplemental irrigation (Li et al. ). RWH can also be an effective source of water for domestic water supply in areas with a limited water supply such as arid and semi-arid areas. It can be easily collected and used without significant treatment (Nolan & Lartigue ). A study by Qin et al. () proved that rainwater buffer tank significantly reduced the runoff peak flow hence had the capacity to protect against the adverse effects of flood such as damage to properties and loss of life.
Appropriate RWH techniques coupled with an efficient Decision Support System have become indispensable tools for conducting sustainable agriculture in the face of climate change (Fenu & Malloci ). The significance of RWH in this context is to make use of the available rainwater that is experienced in Kenya to increase crop production and enhance resilience to climate change. Research has shown that RWH can be effective in increasing crop production and improving adaptability to climate shock. For instance, a study by Kahinda et al. () showed that supplemental irrigation of maize in Zimbabwe using rainwater increased yields and reduced the risk of crop failure. The collection and use of rainwater is therefore seen as a suitable remedy for adverse effects of drought among smallscale farmers.
Despite these known benefits, the adoption of RWH systems by small-scale farmers in Kenya for irrigation is relatively slow (Annastacia ). A study by Matiti () showed that many RWH systems in Kenya are not performing as expected. The study indicated that a number of RWH systems are not reliable and are technically inefficient. Furthermore, a report by Chamwada () indicated that a number of farm ponds and reservoirs constructed under the government household irrigation and water storage program do not fill to their capacity even after prolonged rainy seasons.
Wachira () also noted that many smallholder farm ponds suffer high water losses through seepage and evaporation such that the ponds dry up before the end of the growing season. A study by Matiti () found out that a good number of reservoirs have reduced capacity as a result of heavy siltation while some are completely damaged by water that overtops their banks leading to collapse. Consequently, the impact of the systems has been minimal with farmers taking relatively a long time to return their investment, while other investments on the same are completely lost as a result of the system failure (Kiggundu et al. ).
This review aimed at exploring the factors behind the poor performance of the many RWH systems in Kenya while attempting to render solutions for better design and implementation of the system in the region. It provides guidelines for the effective design and construction of an optimal RWH system for smallholder irrigation farmers.
This will help designers and implementers of RWH systems to understand various approaches that can be applied to optimize the design parameters of the systems in order to contribute to better performance and improved adoption by smallholder farmers in the region hence enable them to develop resilience to climate shock.
The review is structured as follows: First, a brief description of RWH techniques used by smallholder irrigation farmers in the region is provided. Secondly, factors affecting the performance of RWH systems used by smallholder farmers have been explored. Finally, the review looked into the design aspects for an optimal RWH system for the region; consideration was made on ways of optimizing the design parameters to improve the RWH systems reliability, technical feasibility as well as the water-saving efficiency in the area. At the end, conclusions and recommendations are provided.

RWH SYSTEMS TECHNIQUES AND INTERVENTIONS FOR SMALLHOLDER IRRIGATION IN KENYA
There are basically two methods of harvesting rainwater commonly used by smallholder farmers in Kenya, namely in situ and ex situ RWH. In situ RWH techniques involve capturing and using the rainwater at the farm through employing methods that increase the amount of water stored in the soil. Smallholder farmers have also used hand-dug wells to store rainwater for supplemental irrigation. The wells have been constructed mainly in areas, where the water table is relatively high and at the river beds of seasonal rivers (Matiti ). They have been effective in storing rainwater during wet seasons for supplemental irrigation in dry periods. Water from the well can be manually lifted or pumped to a raised reservoir for subsequent application to irrigation fields (Aroka ).

Factors affecting RWH system performance in Kenya
Ex situ RWH techniques have been in existence in Kenya for many decades. Farmers and households have used the techniques to supplement rain-fed agriculture as well as augment the available water resources (Black et al. ). In the last three decades, there has been increasing interest toward the promotion of ex situ RWH systems in Kenya. The national governments, NGOs and other development agencies have initiated a number of projects to harness and store rainwater for irrigation in Kenya. However, many of the RWH systems constructed are not performing as expected (Matiti ). As a result, their adoption by smallholder farmers in Kenya is relatively slow. This can be attributed largely to technical and socioeconomic factors.  According to a study by Gachene & Kimaru (), a greater portion of arid and semi-arid lands of Kenya, where RWH is practiced, consist of black cotton soils (vertisols) which have a cracking tendency especially when dry hence cannot hold water during the wet season. A report by Chamwada () further indicated that the majority of farm ponds implemented under the government water harvesting program have failed to supply the expected water demand owing to high seepage and evaporation losses. Table 2 shows seepage losses for various soils.
In addition, RWH systems for irrigation have also been limited in a capacity as a result of heavy siltation. A number of farm ponds and small earth dams constructed across the country have been abandoned due to reduced capacity caused by the deposition of sediments (Thome

). A study by Karara () on the Kabiruini dam in
Nyeri, Kenya noted that heavy siltation significantly hampered the dam's ability to supply the demand water for irrigation. It indicated that the dam depleted within two years of its desiltation due to sedimentation from the over-

Design and operations aspects of an optimal RWH system for irrigation in Kenya
The main objective of RWH system for irrigation is to ensure the availability of an adequate amount of water for crop pro-

Optimization of RWH system design for reliability
Reliability of RWH system is the probability that the system will supply the required demand of water within a given time (Baek & Coles ). It is calculated from the water balance method by taking into account the catchment size, the storage volume, water demand and the evaporation losses as in the following equation (Ndomba & Wambura ): where V t is the volume of water in the pond at present, V tÀ1 is the volume of water in the pond that remained from the previous rainy season, Q t is the rainwater collected at present, D t is the total consumption per month and V s is the volume of the pond. The reliability index can be obtained from the water balance model using the following equation (Baek & Coles ): where R is the reliability index of the catchment -RWH system, D t is the total demand for water for the crop season and De t is the total deficiency for water.
However, in practice, the reliability index is determined using probability distributions or performance functions using the period when the RWH system does not meet the crop water demand. The following equation (Baek & Coles ) is applied: The reliability index has been used as a good indicator for evaluating the performance of RWH systems. A value closer to 1 indicates high system reliability (Matiti ).
The reliability of RWHS systems is ensured by the optimization of the design parameters including design rainfall, catchment area to storage volume ratio as well as the water-use techniques (Njuguna ). Design rainfall 'the total amount of annual rainfall at which or above which the catchment will provide sufficient water to meet crop needs' is a critical factor in determining the reliability of RWH system (FAO a, b). Appropriate choice of design rainfall is necessary to ensure that the catchment produces enough runoff to meet the crop water demand (Mati et al. ). The design rainfall is determined through probability analysis of annual rainfall series that occur with desired frequency. It can be determined using the empirical formula in Equation (4) (Dirk ); where P is the frequency of rainfall in percentage or the probability %, m is the rank order of rainfall series sorted from the lowest to the highest and N is the number of years of the rainfall series.
For optimal design in tropical humid regions such as Kenya, the use of long-term series of rainfall is recommended with a rainfall magnitude that occurs with 90% frequency. If the rainfall is less than the design rainfall, the stored water cannot meet the crop water requirement. If the design rainfall exceeds the expected amount, the system could be damaged and at times crops may be lost as a result of flood (FAO a, b).
The catchment area in this context refers to the surface area which collects runoff for use in smallholder irrigation.
The surface may be a small watershed, rooftops of houses, where S v is the storage in m 3 , Q is runoff depth in mm and A is the catchment area in m 2 .
The runoff depth Q can be calculated from the following relationship equation (Rozaki et al. ): where P is the accumulated rainfall in (mm), and S (mm) is the potential maximum moisture retention after the beginning of runoff calculated by the following equation (FAO ): where z ¼ 254 for metric units, and RCN is the runoff curve number (estimated from tables).
For optimal design and operations, there is need for careful selection of the design rainfall for the catchment. The RWH system design should take into consideration the maximum potential rainfall (P) that is experienced at 90% frequency. The design must also take into account careful consideration of the appropriate runoff curve numbers that should be an accurate representation of the land-cover type and hydrologic condition of the catchment area. The maximum volume of the pond (V ) can therefore be determined that can hold the desirable quantity of runoff (Q) from the catchment area and satisfies the crop water requirement (D t ).

Optimization of RWH system design for techno-economic feasibility
RWH systems design can also be optimized for technical feasibility. Under this consideration, the design should try to answer the question 'can it be done with the prevailing resources?' (Senkondo et al. ). For RWH systems, technical feasibility is largely a factor of the precipitation that is experienced within a given catchment (Matiti ). If the rainfall experienced within an area is considered adequate, there exists a potential for harvesting and storage. The adequacy of rainfall for harvesting and storage is determined by establishing the design rainfall of a given catchment area (Mati et al. ).
To optimize design for maximum benefits of RWH systems, it must be proved to be technically and economically viable (Rozaki et al. ). Investment analysis can be used to provide useful information for decision-making before constructing RWH systems. Economic analysis of the net present value (NPV), the internal rate of return (IRR) and cost-benefit ratio (B/C) can be used to determine if the RWH system is economically viable (Senkondo et al. ). NPV aims to determine the current value of the project on forecasted net flows which present the future benefits. NPV can be calculated from the following equation (Badiru & Omitaomu ): where t is the time of cash flow, r is the discount rate, n is the depreciation period, C t is the net cash inflow at time t and C o is cash outflow.
The RWH system is feasible when the NPV is positive (Stec & Martina ).
IRR is the discount rate that makes the NPV of all cash inflows of a project equal to zero. A feasible project is one whose IRR is positive and higher than the decided discount rate. If the IRR calculated is greater than the reference rate, then the adoption of RWH system in the area is economically viable. If the IRR determined is less than the chosen reference rate, then the construction of RWH system in the area is economically unattractive (Rozaki et al. ).
B/C is also used to analyze the economic viability of Benefits of RWH systems can be estimated both onsite and offsite. Onsite benefits in this context involve the quantity of crops harvested and their associated value, while offsite benefits may be approximated based on the opportunity cost such as flood protection and soil conservation (Senkondo et al. ). B/C is the ratio of discounted benefits to costs. If the value is more than 1, then the project is acceptable from the financial perspective; while a value less than 1 shows that the project has a negative financial return hence not acceptable (Rozaki et al. ).
B/C can be approximated based on the following equation (Badiru & Omitaomu ): where B t is the benefit at time t, r is the discount rate and C t is the initial cost at time t.
The design of RWH systems therefore calls for the optimization of design parameters to achieve maximum economic benefits.
Optimization of RWH system design for water-saving efficiency RWH system design can also be optimized for the maximum water-saving efficiency. The water-saving efficiency refers to the amount of water that has been conserved against the overall demand of water in the area (Matiti  where E is water-saving efficiency in %, R i is the total volume of rainwater collected per given period in m 3 , D i is the water demand for crop growth in m 3 and T is the total time under consideration.
Maximizing on water-saving efficiency of RWH system can be achieved through a number of techniques including optimal siting of the RWH system, appropriate design of the pond capacity, seepage and evaporation losses control as well as effective management of the stored water (Matiti 

CONCLUSION AND RECOMMENDATIONS
The potential benefits of RWH in Kenya cannot be overemphasized. It can provide the needed additional water resource to bridge the gap in water scarcity; support rainfed agriculture through supplemental irrigation as well as reduce the adverse effects of climate change such as drought and flood. The Kenyan government and other development partners have been active in promoting these RWH technologies in the recent past. Furthermore, individual smallholder farmers have also shown a keen interest in adopting the RWH system techniques as a solution to irrigation water challenges in the last decade. This has seen a rise in the construction of water pans, small earth dams, rock catchments, shallow wells and ponds in different parts of the country.
However, these interventions have achieved little impact owing to consistent poor performance in many areas. From this review, a number of factors were noted to contribute to this low performance and include inadequate design of RWH systems, seepage and evaporation losses, siltation, lack of effective operation and maintenance program, ineffective management as well as inefficient utilization of the stored water. Proper design and implementation are therefore essential for better performance and adoption of RWH systems in the region. This will ensure that RWH systems are reliable, technically and economically feasible as well as possess a desirable water-saving efficiency.
Research on ways of improving the RWH system performance to maximize benefits is thus necessary. RWH system designers should therefore explore techniques of optimizing the design parameters including geological,