Selection of CMIP6 representative climate models and downscaling in Northern Nigeria

Abstract Using multiple General Circulation Models (GCMs) for impact assessment improves the credibility of an analysis and circumvents limit-based uncertainties. Therefore, the selection of the most suitable models for regional/local studies is crucial prior to impact studies and forcing hydrological models. This study included the selection of GCM models from 100 ensembles each for SSP4.5 and SSP8.5 scenarios from the CMIP6 archive using advanced envelop-based selection approach for Northern Nigeria. We used (2021-2050) as the short-term and (2051-2080) as the long-term periods. The approach found that, CanESM5 models are skilful in simulating the warm and wet season, HadGEM3-GC31-LL in the warm and dry season, whereas MPI-ESM1-2-HR and MPI-ESM1-2-LR are skilful in the cold and dry season. We selected three skilled models for each scenario and study period. The skilled models were downscaled for adequate futuristic representation of the temperature and precipitation variables in the region. The downscaled variables predicted an increase in the future. Considering SSP4.5, the average annual precipitation predicted a 13% and 20% increase for the short-term and long-term period respectively. Similarly, for SSP8.5, the predictions are 23% and 41% increment in short-term and long-term periods respectively. For temperature, considering SSP4.5, the average annual temperature predicted an increase of 1.1⁰C and 2.5⁰C for short-term and long-term respectively. Similarly, an increment of 1.2⁰C and 2.7⁰C is expected for SSP8.5 during the short-term and long-term respectively. The selection also suggests that the skilled model does not vary with climate scenario and future periods.