Document Type : Research Article
Authors
Department of Geography, Faculty of Humanities, University of Mazandaran, Babolsar, Iran
10.22067/geoeh.2025.92683.1560
Abstract
Introduction
Two of the eight main soil degradation processes with which soils worldwide are confronted
are soil erosion and landslides. Landslides, as environmental hazards, have significant negative impacts on ecosystems. In Iran, studies on landslides have primarily focused on hazard zonation, with less attention paid to their role in exacerbating erosion and sediment yield in basins. Furthermore, most methods and models developed for assessing erosion and sediment yield in basins have not considered the role of landslides. Additionally, studies evaluating the impact of landslides on sediment yield have often relied on data from hydrometric stations. However, not all material displaced by landslides is transported into the drainage network; some remains on the slopes. Moreover, sediment measurement data is often unavailable for many basins.
The study area, Darabkola watershed in Mazandaran Province, is highly susceptible to landslides due to its weak and tectonically active formations, as well as significant land-use changes. Therefore, landslide hazard zonation and its relationship with sediment yield are essential for better watershed management. The aim of this study is to evaluate the relationship between landslides and erosion intensity, as well as their impact on increasing sediment yield in the Darabkola watershed.
Method
In this study, landslides in the Darabkola watershed were first investigated, and their frequency was determined by integrating factor maps, including lithology, elevation, slope, aspect, land use, distance from roads, and distance from streams. For landslide zonation and determining the weight of influencing factors, the Area Density Model was used, and its evaluation was performed using the Empirical Probability (P) equation.
Next, to determine the sediment yield of landslides, erosion class maps of the watershed were prepared using the MPSIAC and RUSLE models. The accuracy of these models was evaluated using the BLM model, which is based on field observations, as a ground reality map. Additionally, a point layer with a regular grid was created for sampling at 1,204 points from the maps generated by these models. Based on these sampling points, statistical indices such as RMSE, MAE, MSE, and NSEC were calculated. Subsequently, considering the accuracy of the models, the landslide zonation map was overlaid with the sediment yield map, and the data for this study were obtained. The collected data were analyzed using one-way ANOVA.
Results and Discussion
The results of overlaying the landslide distribution map with the factor maps indicate that conglomerate, marl, sandstone, and siltstone (Plcm) with a landslide area density of 0.634, slopes of 15-25% with 0.127, southwest aspect with 0.117, elevation of 150-300 meters with 0.228, distance of 150-300 meters from streams with 0.165, distance of 150-300 meters from roads with 0.273, and rainfed agricultural land use with an area density of 0.304 have the highest landslide susceptibility. After calculating the weight of each factor class influencing landslides in the study watershed, the weighted maps were combined, and a landslide susceptibility map was prepared using the Area Density Model. Its accuracy evaluation using the Empirical Probability equation showed that the landslide zonation map has high accuracy (p = 0.99).
To determine the sediment yield of landslides, the MPSIAC and RUSLE models were first used to prepare erosion class maps for the watershed. Their accuracy was evaluated using observational data from the BLM model and statistical indices. The evaluation of model accuracy using the base model and comparison of RMSE, MAE, MSE, and NSEC indices showed that the MPSIAC model has higher efficiency. Accordingly, the sediment yield map derived from this model was overlaid with the landslide susceptibility map, and the sediment yield of landslide susceptibility classes was determined. The highest sediment yield in the Darabkola watershed was observed in the very high and high landslide susceptibility classes, with values of 61.35 and 53.42, respectively. To examine the significance of sediment yield variations across landslide susceptibility classes, one-way ANOVA was used. The results indicated a significant relationship between sediment yield and landslide susceptibility classes in the watershed. In other words, erosion and sediment yield vary significantly across landslide susceptibility classes in the study area.
Conclusion
In this study, the distribution of landslides in the Darabkola watershed was evaluated by integrating maps of lithology, elevation, slope, aspect, land use, distance from roads, and distance from streams with the existing landslide distribution map. The weighted values of each factor class were calculated using the Area Density Model. The evaluation of the landslide susceptibility map confirmed the high accuracy of the zonation map prepared with this model. Subsequently, the erosion intensity map of the watershed was prepared using the MPSIAC and RUSLE models, and their accuracy was evaluated using the BLM model and statistical indices. Given the higher accuracy of the MPSIAC model, it was used to determine the sediment yield of landslide susceptibility classes in the region. The results of the ANOVA test showed that sediment yield varies significantly with landslide susceptibility, and areas with high landslide susceptibility have greater potential for sediment production. The highest sediment yield was observed in areas with very high landslide susceptibility. The significant contribution of high landslide susceptibility classes to sediment yield underscores the urgent need for targeted conservation measures in the study watershed. It is also recommended that future erosion and sediment models directly incorporate the role of landslides.
Keywords
)
Send comment about this article