Assessing the potential of rainwater harvesting for sustaining small scale irrigated coffee farming in the Bigasha watershed, Uganda2023

Abstract

This study used remote sensing, Geographical Information System (GIS), the Quantum GIS Soil and Water Assessment Tool (QSWAT+) and Crop Water and Irrigation Requirements Program (CROPWAT) to assess the potential of rainwater harvesting (RWH) for sustaining small scale irrigated coffee farming in the Bigasha watershed of Isingiro District in South-western Uganda. The study further demonstrated the benefits of geospatial analysis of large areas where nonspatially based methods have limitations by utilising the many capabilities and strengths found in the four tools selected above. Despite the irrigation potential assessment studies conducted in the Bigasha watershed in recent years, which found the soils in this area to be suitable for coffee farming, production of the crop has been hampered by recurrent drought conditions (Droogers et al., 2012). Farmers were unable to implement irrigated coffee farming due to the lack of a reliable irrigation water source. Therefore, this study was carried out to assess the potential of RWH in the study area to supply the required irrigation water. To identify potential RWH sites in the Bigasha watershed, four RWH site suitability analysis criteria were used: topography (slope), soils, land use and rainfall and runoff depth. Using ArcGIS 10.7 software, the raster maps for the slope and soil feature layers for the study area were created. The study made use of three temporal land use land cover (LULC) maps of 1999, 2010 and 2022. These maps were created by a supervised classification method using Landsat 7 (ETM+), Landsat 5 (TM) and Sentinel 2A (MSI) satellite imageries, acquired from the United States Geological Survey (USGS) and European Space Agency (ESA) databases. The QSWAT+ software interface was used to integrate the input datasets of aforementioned criteria layers and further simulate the surface runoff from the Bigasha watershed. Subsequently, 62, 125 and 114 Hydrologic Response Units (HRUs), derived from the 1999, 2010 and 2022 LULC maps, respectively, were selected as potential RWH sites. The selected HRUs had the potential to generate an annual average surface runoff depth of at least 92 mm and require relatively small catchment area to harvest adequate amount of rainwater. The surface runoff volumes that could potentially be harvested from the three HRUs categories mentioned above were 1.61, 2.68 and 1.39 million cubic meters (MCMs), respectively. vi Finally, the computed annual average gross irrigation water requirement of the coffee crop was 995 mm. This implies that the 1.39 MCM of rainwater in the Bigasha watershed that is currently potentially harvestable could irrigate up to 145 hectares (101.8 per cent) of small-scale coffee fields in the study area annually. The prediction accuracy of the Kagera river basin model was very good with NSE(R2 ) values of 0.81(0.82) for calibration and 0.87(0.88) for validation. The Kagera watershed model fitted parameters were further used to calibrate the Bigasha watershed model. This was done because the Bigasha is a sub-basin of Kagera and does not have its own gauged outlet