Prediction of present and future flood discharges in catchments with sparse data coverage

In this study, we developed a novel and practical approach for curve number estimation that provides a unique solution to represent the infiltration process and address the limitations of input characteristics required by hydrologic models used for flood prediction. Geographic information system based hydrological models were constructed to predict flood discharges for the three future time levels in the basins where the highest annual average precipitation over 2300 mm is observed across the country. Our results indicated a decreasing trend in flood discharges across the study area continuing until 2069. Furthermore, an increasing trend was observed across the study area at the end of the century; however, it remains lower than present condition. Our results also showed that runoff curve numbers of the basins ranged from 72.9 to 81.8, which can be safely used in flow prediction in other locations with similar characteristics. Peak flow elasticity analysis revealed that approximately 1.5% variation in peak discharge may be expected from 1% precipitation change, suggesting considerable basin sensitivity to climate shifts. Most basins demonstrated U-shaped elasticity trends across projection periods, potentially reflecting evolving water retention mechanisms. A noteworthy finding is that 100 year floods exhibited greater sensitivity to precipitation variability than 500 year floods, indicating that moderate flood events may be more responsive to climate variability and thus more informative for detecting climate change impacts.