Assessment of climate change impacts on floods with a hybrid data-driven and conceptual model across a data-scarce region

Abstract

In an era of unprecedented environmental shifts, assessing climate change's repercussions is paramount. This study investigates the impact of climate change on floods in the Kashkan watershed, western Iran, for the near future (2030–2059) and far future (2060–2089). A HEC–HMS model was developed and calibrated from January 1997 to December 2012 and validated from January 2013 to August 2019. To enhance simulation precision, we employed long short-term memory (LSTM), reducing the annual average error peak flow from 23.62 to 9.49%. We selected eight general circulation models (GCMs) using a Taylor diagram for cumulative daily precipitation, maximum daily temperature, and minimum daily temperature, bias-corrected using the quantile mapping method. Under SSP126 and SSP585 scenarios, significant increases in annual maximum discharge are projected, with maximum discharge increasing by 45.08 and 37.59% for SSP126, and by 54.34 and 73.27% for SSP585 in the near and far future, respectively. Most years will see flows exceeding baseline values, especially in autumn. A three-way analysis of variance indicated that the primary uncertainty driver is the interaction among model, GCM, and SSP factors. These findings will aid policymakers in implementing measures to mitigate financial and human losses due to climate change.