The impact of varying spatial resolution of climate models on future rainfall simulations in the Pra River Basin (Ghana)

This work compares future projections of rainfall over the Pra River Basin (Ghana) using data from five climate models for the period 2020-2049, as referenced to the control period 1981-2010. Bias-correction methods were applied where necessary and models' performances were evaluated with Nash-Sutcliffe Efficiency, root-mean-square error and coefficient of determination. Standardised Anomaly Index (SAI) was used to determine variability. The onset and cessation dates and length of the rainy season were determined by modifying the Walter-Olaniran method. The ensemble means of the models projected a 1.77% decrease in rainfall. The SAI showed that there would be drier than normal years with the likelihood of drought occurrence in 2021, 2023, 2031 and 2036. The findings showed that high-resolution models (<= 25 km) were more capable of simulating rainfall at the basin scale than mid-resolution models (26-150 km) and projected a 20.13% increase. Therefore, the rainfall amount is expected to increase in the future. However, the projected increase in the length of the dry season by the ensemble of the models suggested that alternative sources of water would be necessary to supplement rainfed crop production for food security.