Use of satellite and reanalysis data in integrated rainfall-runoff and flood inundation analysis

The loss of life, displacement of populations, and destruction of infrastructure resulting from floods, which have emerged as a significant global issue, underscore the critical importance of scientific research in this domain. This study aims to analyze flood inundation areas within a river basin utilizing various datasets and to validate the findings through independent data sources. The methodology integrates three principal platforms: a grid-based hydraulic model (HEC-RAS version 6.5), remote sensing tools, and geographic information systems (Google Earth Engine and ArcGIS). In the context of this study, an integrated rainfall-runoff and flood inundation modeling approach was utilized, incorporating observational data, satellite-based precipitation data (specifically IMERG-30min and CHIRPS), and reanalysis data (ERA5-Land) as inputs to the model. The resulting streamflows, flood inundation extents, and water depths were subsequently analyzed and compared with one another. Furthermore, model outputs are validated against spatial flood observations obtained from Sentinel-1 imagery, which serve as an external reference. Flood simulations were conducted for five events that occurred in 2018, 2019, and 2024. Among the scenarios evaluated, the combination of IMERG-30mn and observed precipitation data yielded the most consistent results, with an average POD (Probability of Detection) score of 80.20%. The highest agreement with Sentinel-1 data was observed in the analysis of the 2024 flood event. This research underscores the significance of integrating multiple data sources in flood modeling and demonstrates their potential applications in flood risk management and planning. By evaluating their effectiveness, the study contributes to the refinement of strategies aimed at mitigating flood hazards.