| Summary: | Reliable rainfall data are critical for managing hydrometeorological hazards in West Africa, yet they are often sparse and temporally inconsistent. The current study assessed the accuracy of four near real-time satellite-based rainfall data, namely IMERGv7 Late, IMERGv6 Early, GSMAP-NRT and PERSIANN-DIR Now, for rainfall estimation and hydrological modeling in the Ouémé basin. These datasets were compared with ground-based rainfall data, bias-corrected and used to calibrate and validate the hydrological model HBV light. While they demonstrated qualitative accuracy, their quantitative estimation shows obvious discrepancies on a daily scale, varying across subdomains. The original IMERGv7 product outperforms others in capturing the rainfall pattern and amount (KGE > 0.6), while GSMAP performs moderately (KGE ≈ 0.51) and IMERGv6 and PERSIANN show lower reliability with KGE < 0.5. Quantile mapping emerges as the most effective bias-correction method, improving the performance of all satellite products, with RMSE reductions ≤ 15%. The results of hydrological simulations demonstrate the potential of satellite-based rainfall, particularly IMERGv7 and corrected IMERGv6 (NSE > 0.75), for near real-time flood monitoring and water management in the study area. This study underscores their suitability as valuable alternatives to ground-based data for flood management decision making in the Ouémé basin.
|
|---|
