Flood anticipation, reality, and uncertainty, the 2019 flood in Khuzestan, Iran

Despite the application of various methods to calculate uncertainty in flood vulnerability assessments, the challenge of uncertainty remains. The main purpose of this study was to establish the impact of uncertainty on flood risk anticipation. To this end, the real flood zone in Khuzestan region, southwestern Iran, was initially determined in 2019, and then compared with the 100-year flood, forecasted by competent authorities and modeled using the analytic hierarchy process (AHP) and the ArcGIS software. In this study, uncertainty in flood-induced vulnerability was also assessed via the AHP and the ArcGIS software to reveal its conformity with reality. Afterward, the digital elevation models (DEMs) and satellite imagery data coupled with the overlay of vulnerable/invulnerable areas were determined in the real flood. The results were subsequently compared with the 2019 real flood and the 100-year return, anticipated by the AHP-GIS model. The output revealed that 53% of the invulnerable areas had been inundated in total, leading to huge economic damage and surprising local people. Furthermore, the findings demonstrated that despite much effort, the applied models did not always correspond to reality. To predict the future and to minimize uncertainty, a scenario-based robust optimization approach or regret theory would be suggested to solve such problems.

Automated summary

This study aims to investigate the impact of uncertainty on flood risk anticipation and evaluate the accuracy of the models by comparing the actual flood with the model predictions. The findings reveal that the existing models have certain uncertainties in prediction, and other methods should be employed to reduce uncertainty.