Evaluating Future Temperature and Precipitation Trends in Northern Nigeria: A Multi-Model Approach to Downscaling and Forecasting

Abstract

This study assessed the effects of climate change on hydroclimatological parameters in Northern Nigeria with a focus on temperature and precipitation. The stations considered were Sokoto and Gusau. Ten variables were chosen each from GCMs under IPCC-AR5 and PalMod2 in order to perform statistical downscaling for the years 1990-2022 and forecast temperatures and precipitation for Sokoto and Gusau for the years 2067-2099 using the Hammerstein-Wiener (HW), Nonlinear Autoregressive Exogenous (NARX) and Autoregressive Integrated Moving Average (ARIMA) models. In order to identify the best variables that have a stronger correlation with the observed temperature and precipitation, the correlation coefficient feature extraction method was used. The models were developed based on the 2, 3 and 4 most dominant input variables designated as M1, M2, and M3. The downscaling results showed good performance from the models, although HW and NARX showed a more consistent performance than ARIMA in downscaling of temperature. The predictions for temperature show that both stations experience a general increase from November to February, while a decrease occurs from March to June. Precipitation in both stations and under both projects show a decreasing trend especially in the month of August.