Process-Based Flood Type Separation for Partial Duration Series (Sample Study: The Gorganrood River Sub-Basins)

Determining flood generation processes and incorporating them in hydrological studies can help make more reliable estimates. Focusing merely on mathematical computations in flood studies sometimes lead to paying less attention to the basic physical processes. Therefore, this research aims to present a process-based flood separation study for 20 hydrometric stations in the Gorganrood River Basin attributed as a flood prone area in the northern part of Iran. For this purpose, partial series of flood events at each station were extracted using the Peak-Over-Threshold (POT) approach selecting an optimal threshold value to determine flood events and considering the independence of the series values. Subsequent to characterizing flood events using some separation indicators including time of occurrence, rainfall duration, flood duration, air temperature, and antecedent soil moisture condition, different types of floods (i.e., floods caused by frontal precipitation, convective precipitation, snowmelt and rain on snow) were identified for each station. The results showed that floods caused by frontal precipitation and rain on snow have the highest and lowest frequencies in the basin, respectively. The research findings indicated that using POT approach and the indicators for classifying the different flood types can lead to a more accurate statistical analysis in flood studies and management.