Flood

Panahi, M., Rahmati, O., Kalantari, Z., Darabi, H., Rezaie, F., Moghaddam, D. D., ... & Lee, S. (2022). Large-scale dynamic flood monitoring in an arid-zone floodplain using SAR data and hybrid machine-learning models. Journal of Hydrology, 128001. Click (a) Inundation mapping at peak flooding: Maximum flood extent on April 14, 2019, analyzed via the SVM-GWO model. (b) Suitable locations for flood protection: Nine candidate sites for embankments derived from spatio-temporal flood patterns. (c) Performance metrics: Comparison results showing the superior predictive capability of the proposed hybrid model over the standalone SVM. This study proposes a robust framework for monitoring large-scale spatio-temporal flood dynamics in an arid-zone floodplain (Ahvaz, Iran) by integrating multi-temporal SAR data with hybrid machine-learning models. Using Sentinel-1 SAR images acquired between March and May 2019, the Support Vector Machine (SVM) algorithm was hybridized with three metaheuristic optimization procedures, including Grey Wolf Optimization (GWO).The results demonstrated that the SVM-GWO approach yielded the highest predictive capability, achieving a validation overall accuracy of 93.39%. Time-series analysis revealed that wetlands were the last land-use type to return to normal conditions due to previously dry oxbow lakes trapping water. Furthermore, by analyzing the spatial distribution of floodwaters, nine highly suitable locations for flood protection structures were identified, providing a data-parsimonious approach to optimize resource allocation for flood mitigation in data-scarce regions. Continue reading...