Introduction

Flooding is one of the most significant natural hazards worldwide, causing substantial damage annually to infrastructure, agricultural lands, communities, and natural resources (Rentschler et al., 2022). Due to the increasing frequency of extreme precipitation events and ongoing climate change, both the frequency and intensity of floods have risen in many regions of the world (Costache et al., 2020). Higher surface runoff and increased flood occurrence are also driven by human-induced factors such as land use change, unplanned urban expansion, and vegetation degradation (Rogger et al., 2017). Since flash floods are often triggered by steep slopes and short concentration times, this issue is particularly critical in hilly and semi-mountainous watersheds.

Method

In this study, thirteen factors were employed for flood susceptibility mapping, including elevation, slope, curvature, distance from streams, drainage density, Topographic Wetness Index (TWI), Stream Power Index (SPI), Sediment Transport Index (STI), Normalized Difference Vegetation Index (NDVI), land use, precipitation, geology, and soil. A Digital Elevation Model (DEM) was used to extract topographic parameters such as elevation, slope, and curvature. Geological and soil layers were incorporated to account for their influence on infiltration capacity, erodibility, and water retention characteristics. The methodology involved the extraction, preparation, and standardization of spatial datasets relevant to flood susceptibility assessment in the Tajyar Sarab Watershed. A Digital Elevation Model (DEM) was processed in ArcMap and clipped to the watershed boundary to derive elevation. Topographic derivatives including slope and curvature were generated from the DEM, given their influence on runoff velocity, infiltration, and flow convergence. Hydrological conditioning of the DEM was performed to extract flow direction and flow accumulation, which were subsequently used to delineate the drainage network. Distance to streams was calculated using Euclidean distance, while drainage density was derived from the channel network. Lower distances to streams and higher drainage density were considered indicative of higher flood susceptibility. Geomorphometric and hydrological indices, including the Topographic Wetness Index (TWI), Stream Power Index (SPI), and Sediment Transport Index (STI), were computed to represent soil moisture accumulation, flow energy, and sediment transport capacity, respectively. Vegetation conditions were represented using NDVI, where lower values were associated with higher flood susceptibility due to reduced infiltration and surface roughness. Land use, precipitation, soil, and geology were also incorporated based on their influence on runoff generation, permeability, and water retention capacity. All spatial layers were standardized in terms of coordinate system (WGS 1984 UTM Zone 38N), spatial resolution (30 m), extent, and cell alignment using a snap raster to ensure spatial consistency. The criteria were classified into five susceptibility classes and normalized to a 0–1 scale to enable integration within multi-criteria decision-making (MCDM) models, where higher values represent higher flood susceptibility.Objective weights were assigned using the entropy method based on information content and spatial variability of each criterion. Among the evaluated factors, SPI and STI obtained the highest weights (0.4686 and 0.3108, respectively), indicating their strong discriminatory power in the study area. Flood susceptibility mapping was then performed using two MCDM techniques: SAW and VIKOR. The SAW model produced a weighted linear aggregation of standardized criteria, while the VIKOR model generated a compromise ranking based on proximity to an ideal solution. The outputs of both models were integrated to produce a final composite flood susceptibility map and a spatial agreement map identifying areas consistently classified as highly susceptible.Finally, area statistics were computed for each susceptibility class based on pixel counts. Comparative analysis of SAW, VIKOR, and the integrated map enabled quantitative evaluation of high-risk zones and model consistency, providing a robust framework for flood susceptibility assessment in data-scarce watersheds.

Results

Entropy-based weighting revealed that flood susceptibility in the Tajiar-Sarab watershed is primarily governed by hydrological and geomorphometric factors, with Stream Power Index (SPI) and Sediment Transport Index (STI) receiving the highest weights of 0.4686 and 0.3108, respectively. This indicates that flow power, runoff energy, and sediment transport capacity are the main drivers of spatial differentiation of flood-prone areas. Other factors, including geology, Topographic Wetness Index (TWI), precipitation, and slope, contributed less but remained influential in shaping the watershed’s susceptibility patterns. Both SAW and VIKOR models identified flood-prone areas mainly in the southern and southwestern parts of the watershed along primary drainage pathways, while the northern and northeastern regions exhibited lower susceptibility. Quantitative comparison indicated that approximately one-third of the watershed falls into high and very high susceptibility classes (35.06% in SAW and 36% in VIKOR). The spatial agreement map highlights regions consistently classified as highly susceptible by both models, offering a practical tool for prioritizing land-use planning and runoff management interventions. These findings demonstrate that integrating topographic, hydrological, and environmental criteria through multi-criteria decision-making models, combined with entropy-based weighting, provides a reliable approach for identifying and managing flood-prone areas in watersheds lacking historical flood records.

Conclusions

This study demonstrated the effectiveness of entropy-based weighting combined with SAW and VIKOR multi-criteria decision-making models for flood susceptibility mapping in the Tajyar-Sarab watershed. Among the thirteen selected criteria, the Stream Power Index (SPI) and Sediment Transport Index (STI) played the most dominant roles in spatial differentiation, highlighting the critical influence of flow energy and runoff transport capacity on flood susceptibility patterns.The results of both SAW and VIKOR models indicated that approximately one-third of the watershed area (35.06% in SAW and 36% in VIKOR) falls within high and very high susceptibility classes. These areas are mainly concentrated along the primary drainage network and in the southern and southwestern parts of the watershed. The spatial agreement analysis revealed that nearly 35% of the watershed was consistently classified as highly susceptible by both models, indicating higher reliability of these zones for management prioritization. In contrast, about 65% of the watershed showed spatial disagreement between the two models, reflecting differences in their underlying decision-making logic. The final integrated map provided a more stable and consistent representation of flood susceptibility, identifying the southern, southwestern, and main drainage corridors as critical zones. Overall, the findings confirm that integrating entropy weighting with SAW and VIKOR models provides a robust framework for identifying flood-prone areas in data-scarce watersheds.