Evaluation and Error Correction of Satellite Rainfall Using Dense Network of Rain Gauges Over Mumbai

This study evaluated satellite-based rainfall estimates from the Global Precipitation Measurement (GPM) Integrated Multi-satellitE Retrievals for GPM (IMERG) -Late product over the Mumbai region, India, using a dense rain gauge network. In the initial evaluation of the southwest monsoon seasons (JJAS of 2019–2023), the IMERG-Late rainfall product consistently underestimated rainfall relative to gauge observations. To correct these biases, three quantile-mapping techniques were applied: basic, normal, and modified quantile methods. Among all three, the basic method showed the best performance, reducing bias by ~ 75% and also reducing root mean square and mean errors, and enhancing the correlation with gauges from 0.39 to ~ 0.8. The Normal quantile method reduced bias from approximately 4 mm/day to 1 mm/day, while the modified quantile method yielded comparatively weaker improvements. The probability of detection increased to nearly 1.0 for rainfall rates exceeding 1 mm/day at most stations, while the false alarm ratio remained low. This study demonstrated that integrating dense gauge networks with satellite data can produce high-resolution rainfall estimates with better accuracy in complex urban environments. Enhanced estimates have the potential to improve flood forecasting and water resource management in flood-prone regions like Mumbai. This methodology can be applied to other urban areas to improve satellite-based precipitation estimates, though further research is needed to refine bias-correction techniques for different rainfall intensities and seasons.