University of Sistan and Baluchestan Journal of Natural Environmental Hazards 2676-4377 15 48 2026 06 22 Assessing the flood potential of the Khiaochay watershed using the ‎MFFPI model and spectral indices Assessing the flood potential of the Khiaochay watershed using the ‎MFFPI model and spectral indices 73 92 9333 10.22111/jneh.2025.51600.2109 FA Sayyad Asghari Saraskanrood Professor, Department of Physical Geography, Faculty of Social Sciences, University ‎of ‎Mohaghegh Ardabili, Iran Mahdi Frotan Ph.D. Student of Climatology, Department of Physical Geography, Faculty of Social Sciences, University ‎of ‎Mohaghegh Ardabili, Ardabil, Iran Journal Article 2025 04 07 Flash floods are important hazards in water resource and ‎environmental ‎management. This study aimed ‎to identify flash ‎flood-prone areas in the Khiaochay ‎watershed and investigate ‎the relationship between ‎spectral indices and flood risk. In ‎this ‎regard, the MFFPI model was used as the main tool to ‎determine flood-prone ‎areas. The parameters used included ‎slope, flow density, slope curvature, rock type, ‎soil texture, and ‎land ‎use, which were extracted from sources such as ‎digital ‎elevation models (DEM), geological maps, and remote ‎sensing ‎data. After ‎classification and weighting based on the modified ‎version of the MFFPI model, ‎these ‎parameters were processed ‎in a GIS environment, and a final flood hazard ‎map was ‎produced. To evaluate the ‎model's performance, two floods ‎recorded in ‎‎2020 were examined, and ROC analysis was ‎performed to measure ‎the accuracy of ‎the model. In addition, ‎the relationship between the spectral indices MNDWI, ‎NDMI, ‎AWEI, and LSM ‎ and the MFFPI values was examined using ‎Spearman's ‎correlation test. The results showed that the ‎northern ‎areas, parts of the center, and ‎some southern areas of ‎the basin have the highest risk of flooding. The ‎evaluation ‎of the ‎parameters showed that factors such as low slope, high ‎flow ‎density, low-permeability clay soils, concave ‎slopes, urban land ‎use, and hard ‎igneous rocks are effective in increasing runoff ‎and flooding. The analysis of ‎the ‎spectral indices also showed ‎that the LSM index has a positive and significant ‎relationship ‎with the MFFPI ‎model and can be effectively used to identify ‎flood-prone areas. The AUC values for the two floods ‎studied ‎were 0.73 and 0.72, ‎respectively, which indicates the ‎acceptable performance of the model in predicting ‎flood ‎risk. Flash floods are important hazards in water resource and ‎environmental ‎management. This study aimed ‎to identify flash ‎flood-prone areas in the Khiaochay ‎watershed and investigate ‎the relationship between ‎spectral indices and flood risk. In ‎this ‎regard, the MFFPI model was used as the main tool to ‎determine flood-prone ‎areas. The parameters used included ‎slope, flow density, slope curvature, rock type, ‎soil texture, and ‎land ‎use, which were extracted from sources such as ‎digital ‎elevation models (DEM), geological maps, and remote ‎sensing ‎data. After ‎classification and weighting based on the modified ‎version of the MFFPI model, ‎these ‎parameters were processed ‎in a GIS environment, and a final flood hazard ‎map was ‎produced. To evaluate the ‎model's performance, two floods ‎recorded in ‎‎2020 were examined, and ROC analysis was ‎performed to measure ‎the accuracy of ‎the model. In addition, ‎the relationship between the spectral indices MNDWI, ‎NDMI, ‎AWEI, and LSM ‎ and the MFFPI values was examined using ‎Spearman's ‎correlation test. The results showed that the ‎northern ‎areas, parts of the center, and ‎some southern areas of ‎the basin have the highest risk of flooding. The ‎evaluation ‎of the ‎parameters showed that factors such as low slope, high ‎flow ‎density, low-permeability clay soils, concave ‎slopes, urban land ‎use, and hard ‎igneous rocks are effective in increasing runoff ‎and flooding. The analysis of ‎the ‎spectral indices also showed ‎that the LSM index has a positive and significant ‎relationship ‎with the MFFPI ‎model and can be effectively used to identify ‎flood-prone areas. The AUC values for the two floods ‎studied ‎were 0.73 and 0.72, ‎respectively, which indicates the ‎acceptable performance of the model in predicting ‎flood ‎risk. https://jneh.usb.ac.ir/article_9333_bfb48dc6b7489d19ce165cf87617d4ff.pdf