Using integrated multi-criteria decision analysis, GIS-based applications for selecting suitable sites for small irrigation reservoirs : A case study of the Savannah region, Ghana

Water scarcity affects people living in semi-arid and arid regions due to extreme rainfall variability and unpredictable droughts or floods, leading to livelihood vulnerability. The study area, the Savannah region of Ghana, is situated in the country's north-western corner with a surface area of 35,862 km2, a semi-arid area with yearly average rainfall ranging from 890 – 1300 mm. Considering the climatic condition of the region, one of the strategies to overcome water scarcity in these regions is the construction of small dams and reservoirs to harness rainfall. The engineering design of Rainwater Harvesting (RWH) structures and the identification of suitable sites are critical to their success. Several studies have developed different criteria and approaches for selecting suitable RWH sites. Finding an appropriate dam/reservoir site selection approach that replaces the traditional iterative, costly and time-consuming process is key to solving agricultural water problems and improving productivity. This research aimed to develop a procedure for locating probable sites for small dams and reservoirs in the Savannah region of Ghana using remote sensing (RS) and geographic information systems (GIS) in combination with multi-criteria decision analysis (MCDA) tools in ArcGIS 10.7.1 software. The methodology considered various biophysical and socioeconomic factors such as slope, rainfall, roads, drainage networks, geology, soils, settlements, land use land cover, and protected areas. Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) was used to create maps of elevation, slope, contour and drainage network density. Rock structure and precipitation distribution were generated using geologic maps and precipitation maps from the United States Geological Survey (USGS) website. In addition, soil data from the Food and Agriculture Organisation (FAO) website were used to show how the major soil classes are distributed. Suitability maps were generated for each criterion using ArcGIS, with weighting based on the most influential factor. Analysis of a weighted overlay was used to create a final suitability map indicating suitable locations for dam construction. The findings show that 25% of the study region is designated as a conservation area. The remainder of the area is divided into five suitability levels: poor, fair, moderate, good, and very good, covering 11%, 19%, 20%, 15%, and 10% of the potential areas, respectively. Dam characteristics such as height, surface area, and storage capacity were calculated for six potential sites from a combination of suitability maps, drainage network density maps, and cross-sectional profiles generated from SRTM DEM. The maximum height and length of the dams were determined to be 15 metres and 1931 metres, respectively. The highest storage capacity was estimated at 89,619,168 cubic metres. The result is validated using existing reservoirs. The findings showed that most existing reservoirs were not identified in good and very good suitability areas. However, the final suitability map showed good sites for constructing water harvesting structures. Planners and decision-makers should be able to use the suitability map to quickly identify sites where rainwater harvesting is most likely to succeed