Evaluating the Accuracy of CHIRPS Satellite Rainfall Estimates Over Jos Plateau

Satellite-derived rainfall products have become indispensable tools for hydrological and agricultural applications in data-sparse regions of Nigeria, yet their accuracy must be evaluated against ground-based observations before use in decision-making contexts. This study evaluates the accuracy of the Climate Hazards Group InfraRed Precipitation with Stations (CHIRPS) product over the Jos Plateau region in Plateau State, using daily and monthly gauge observations from four NiMet stations (Jos, Shendam, Pankshin, and Wase) for the period 2019 to 2023. Verification metrics including Pearson correlation coefficient, bias, root mean square error, and probability of detection for wet days were computed at daily, monthly, and seasonal timescales. Results show that CHIRPS performs well at monthly and seasonal aggregations, achieving correlations of 0.87 to 0.93 with gauge observations during the main rainy season, but exhibits significant overestimation of light rainfall events at daily timescales. Performance is notably poorer at Wase station, attributed to complex topographic influences that the satellite algorithm inadequately captures. CHIRPS consistently underestimates extreme rainfall events exceeding 60 millimetres per day across all stations, a limitation with significant implications for flood early warning applications. The study recommends bias-correction procedures calibrated with local gauge data before using CHIRPS for operational hydrological applications in the Jos Plateau area. Keywords: CHIRPS, satellite rainfall, validation, Jos Plateau, bias correction.