چكيده لاتين
Flood, as one of the powerful and dangerous natural phenomena, occurs in various regions including mountainous and low-lying areas. Due to the presence of hills and steep areas with flood potential, mountainous regions are prone to soil and rock erosion due to the occurrence of this phenomenon. In the event of a flood in mountainous areas, water, soil, and large rocks quickly move downhill and can cause serious damage to homes, roads, bridges, and other infrastructure in these areas. Floods also occur in low-lying areas due to excessive rainfall, snow melting, and the presence of rivers. In the event of a flood, water quickly flows down from the riverbank and may destroy low-lying areas along the waterway. The objective of the present study is to classify the flood potential using a FUZZY Model and predict the occurrence of floods in flood-prone areas using a Random Forest Model in the Zohreh-Jarahi watershed. To achieve this goal, 13 parameters including land use, NDVI, geology, rainfall, runoff curve number, stream power index, drainage density, distance from the river, landform, plan curvature, elevation, and slope were considered. The fuzzy map of the considered parameters was prepared, and their weights were determined based on the AHP method and their impact on flood occurrence. Using the fuzzy maps, the overlapping maps of the layers were prepared using fuzzy operators such as GAMMA, and, and or. Among these operators, the 0.9 GAMMA operator had the highest conformity with reality, and the potential flood zoning map in the study watershed was prepared based on this fuzzy operator. According to the correlation table of the layers in this fuzzy operator, SPI with a coefficient of 0.6, NDVI -0.4, rainfall 0.38, geology -0.34, land use 0.3, and drainage density 0.001 had the highest and lowest correlation coefficients, respectively, with the flood potential of the area. The potential flood zoning map of the watershed shows that the Seyedoun, Baghmalek, Heydarabad, orokan, and Behbahan sub-watersheds are among the highly hazardous areas in terms of floods, and overall, the very high-risk zone covers 6.86% of the watershed area. 53% of the watershed area is located in the very low-risk and low-risk zones. To prepare the flood hazard zoning map of the areas in the watershed, the Random Forest Model was used based on the 13 mentioned parameters. First, the map of flood-prone points in the watershed was prepared using Google Earth and LandSat satellite images. The layer coding was done using R software. Finally, the model was implemented in the SPM software, which showed that the distance from the river with a score of 64, land use 52, rainfall 49, and elevation 38 had the highest scores, while the runoff curve number with a score of 0.3 and Plan curvature with a score of 0.1 had the lowest scores and values in the sub-watersheds prone to floods. According to the results of this model, the Hindijan, Ramhormoz, and Choram sub-watersheds are among the highly hazardous areas in the Zohreh-Jarahi watershed, and the very high-risk flood-prone zones cover 7% of the watershed area. After examining and comparing the two models, it was found that the Random Forest Model has a higher accuracy in predicting flood-prone areas in the studied watershed compared to the Fuzzy Model in introducing flood-prone zones of the watershed.
