assessment of re-forecast data in the modeling of extreme rainfall-runoff events (case study: floods in the bakhtiari basin, iran, march-april 2019)

Predicting inflow into reservoirs is essential for their operation during floods, particularly in mountainous watersheds characterized by snow-rain regimes. the objective of this research is to evaluate the gefsv12 re-forecast data as an input of the hec-hms model for forecasting floods due to the extreme precipitation in march/april 2019 in the reservoir of bakhtiari dam in southwestern iran. so, ensemble flood forecasting (control and ensemble members) was conducted using extracted precipitation and temperature data with the lead-time up to 10 days. a sequence of predictions for flood warnings was analyzed when 50% of the members exceeded the threshold inflows of 1000 and 1500 m³/s. the relative volume error values for the control member and the ensemble mean for five days ahead were -15% and -22%, respectively. while previous studies in catchments with snow-rain regimes anticipated challenges in flood forecasting at mid-lead times, this research demonstrated that the proposed probabilistic framework could effectively issue flood warnings for events with a lead time of five days. to address and predict flooding at the bakhtiari dam with a threshold of 1500 m³/s, flood warnings are issued with a lead time of 5 to 8 days.