Empirical Evaluation of the Performance of Global Reanalysis Products for Rainfall Estimation Over the Niger Delta Region

Global atmospheric reanalysis products are increasingly used as substitutes for sparse in-situ meteorological observations in data-limited regions, yet their performance over the complex meteorological environment of the Niger Delta remains insufficiently validated. This study empirically evaluates the performance of four major reanalysis products, ERA5, MERRA-2, NCEP-CFSR, and JRA-55, in estimating daily, monthly, and seasonal rainfall over the Niger Delta region, comprising Rivers, Delta, Bayelsa, Akwa Ibom, and Cross River states. Validation was conducted against quality-controlled daily rainfall records from 14 NiMet stations for the period 2010 to 2020 using standard verification metrics including bias, root mean square error, Pearson correlation, and equitable threat score for wet day occurrence. Results reveal that ERA5 consistently outperforms other products, achieving monthly rainfall correlations of 0.78 to 0.91 and bias of minus 8 to plus 14 percent across stations. All products substantially underestimate extreme daily rainfall events exceeding 80 millimetres, with bias values exceeding minus 40 percent for the upper decile of daily rainfall. MERRA-2 exhibits a systematic dry bias exceeding 20 percent during the August rainfall minimum period, limiting its utility for hydrological applications in the Niger Delta. Performance is notably inferior over coastal stations compared to inland stations, reflecting the challenge of resolving mesoscale moisture convergence processes over the Gulf of Guinea coastline. Recommendations include bias-correction protocols before using any reanalysis product for operational applications in the Niger Delta. Keywords: reanalysis, rainfall validation, Niger Delta, ERA5, bias correction.