Performance evaluation of GPM-IMERG early and late rainfall estimates over Lake Hawassa catchment, Rift Valley Basin, Ethiopia

High resolutions of satellite rainfall products have been widely used for hydrometeorological and hydroclimatological studies over the globe. However, the performance of satellite rainfall estimates varies and is affected by topography and atmospheric characteristics. The assessment of satellite rainfall products is important over different regions. In this study, Integrated Multi-SatellitE Retrievals’ performance for the Global Precipitation Mission version 6 (GPM-IMERG v6) was evaluated before and after bias correction, over the Lake Hawassa catchment. A linear scaling bias correction approach was used to correct the bias of GPM-IMERG early and late rainfall products. The satellite rainfall products were also compared with ground observed rainfall data in the Lake Hawassa catchment. Statistical performance assessing methods were used to evaluate both raw and bias-corrected IMERG early and late rainfall products. The percentage of bias (PBIAS) for early and late rainfall estimates was 91.54 and 77.03, respectively, for the entire periods before bias correction. It indicates that GPM-IMERG overestimated rainfall relative to ground-gauged rainfall. The results show that IMERG rainfall products are in good agreement with ground observed rainfall after bias correction. The correlation values (R) for IMERG early and late is 0.86 and 0.85, respectively, indicating a good correlation between IMERG’s estimated rainfall and observed rainfall after bias correction. The performance of IMERG rainfall estimates varies with the seasons. The bias correction for only rainy season shows a good match with observed rainfall  compared to all seasons. Bias-correction resulted in a good match between estimated and observed rainfall. Generally, evaluation of GPM-IMERG satellite rainfall products is essential prior to use for hydrological modeling and forecasting in data-scarce areas.