Hossein Etemadfard - Assistant Professor, Civil Engineering Department, Ferdowsi University of Mashhad, Iran
Ali Hamid Imran Huseeni - M.Sc. Student, Civil Engineering Department, Ferdowsi University of Mashhad, Iran
Rouzbeh Shad - Associate Professor, Civil Engineering Department, Ferdowsi University of Mashhad, Iran
Marjan Ghaemi - Visiting Professor, Civil Engineering Department, Ferdowsi University of Mashhad, Iran

This study aimed to use remote sensing datasets and machine learning approaches to assess risk of floods in Al-Kut, Wasit governorate, Iraq. The flood event that happened on November ۲۶, ۲۰۱۸ in Al-Kut city used as the case study for this study. First, an inventory flood map with a total of ۱۴۴ samples was generated based on a flood reference map given by SERTIT for the research area. Second, six flood conditioning factors, including altitude, aspect, slope, curvature, land cover, and distance from water bodies, were prepared for flood susceptibility assessment. Third, the relationship between flood conditioning factors and flood events in the research area was established using a logistic regression (LR) for flood spatial prediction. Then, monthly rainfall data from ۲۰۱۳ to ۲۰۱۷ were collected and analyzed using effective accumulation maximum for flood hazard assessment. Flood vulnerability was first evaluated using vulnerability indicators such as population density and land use for flood risk assessment. Finally, the Receiver Operating Characteristic (ROC) curve and the area under the ROC curve, known as the AUROC, were used to evaluate flood susceptibility models. The flood susceptibility model's validation accuracy using LR was ۰.۶۳ Precision, ۰.۶۲ Recall, ۰.۶۲ F۱-score, and ۰.۶۲۵ AUROC. When it comes to risk assessments, residential and agricultural areas are the most vulnerable (bare land and water body). The majority of the study area is at extremely low risk of flooding, according to the results (۳۴.۸۵ percent). Only ۹.۸۷ percent and ۱.۷۹ percent of the area, respectively, are at high and extremely high risk.

Flood susceptibility, Logistic regression, Sentinel, Wasit province