A Comparative Evaluation of MABAC and CODAS Multi-Criteria Decision Algorithms in Landslide Risk Zoning (Case Study: Kowsar County)

The landslide as one of the mass movements consists of rapid movement or the slow motion of rock, earth, or the sum of both, on the downstream side under the stress of gravity. Landslides are considered dangerous and catastrophic phenomena so that the identification of landslide prone areas and the production of accurate zoning maps of landslide sensitivity are important topics for risk management studies. One of the most important and widely used methods of studying landslides is zoning. The division of lands into separate areas, as well as the ranking of areas according to the degree of landslide risk, is called zoning. In the present study, the study area is Kowsar County, which has a high potential for landslides in terms of the specific situation of the region, such as topography (mountainous and high basin), high slope, the presence of loose and unstable surface materials on resistant structures and climatic conditions.

The present study is of applied type and its research method is based on remote sensing techniques, GIS, and the use of multi-criteria analysis techniques. ENVI, Ecognition, Arc GIS, Idrisi, and Excel software have been used for image processing and data analysis. Criteria required in this research are slope, slope direction, elevation classes of lithology, land use, soil, precipitation, distance from communication road, distance from the waterway, and distance from the fault, according to the natural and human conditions of the region. Information layers of communication routes and waterway networks were obtained using the map of communication lines and rivers of Ardabil province. Slope layers for slope direction were prepared using the digital elevation model. Using the geological map of Givi (scale 1: 100000), layers related to lithology (rock resistance) as well as faults were prepared. In order to extract the land use of the city, the object-oriented method, the nearest neighborhood algorithm, and Sentinel 2 satellite images were used. To prepare the soil map, the soil map of Ardabil province with a scale of 1:50,000 was used. The rainfall map of the city was also drawn using meteorological and rainfall station data and the IDW interpolation method. In order to perform the steps of MABAC and CODAS algorithms, fuzzy membership functions were used for standardizing the criteria. Then, in order to weight the criteria, the critic method was used.

In this study, distance from the road, distance from the waterway, digital elevation model, distance from the fault, precipitation, land use, slope, aspect slope, soil, and lithology were identified as factors affecting the area. Then, fuzzy membership functions, the above-mentioned elements became fuzzy maps. The following results were obtained with respect to each element:Height: height changes of each region as an effective factor in the creation of mass movements.Slope: the slope is mainly determined in spatial distribution and landslide intensity.Slope direction: gradient orientation plays an undeniable role in the amount and type of weathering of rocks, absorption, receiving solar energy, vegetation concentration, and moisture content.Lithology: the kind of geological formation of each region plays an important role in the development of sliding zones.Distance from fault: the path of potential faults is on unstable points of the earth.Soil: the soil of each region is an effective factor in the creation of domain movements.پPrecipitation: precipitation has an effect on pore water pressure and groundwater-surface, increasing amplitude and displacement and elimination of lateral stresses as a driving factor in landslide occurrence.Distance from waterway: water flowing through the surface (runoff) and water that penetrates the material.Distance from the road: the road is the natural state of the domain and causes vertical gashes in the slope.Land use criteria: certainly, the type of land use always plays a decisive role in environmental changes, especially the creation and intensification of range movements.Based on the zoning done using the MABAC method, 211.99 and 298.85 km2 of the area are in very high-risk and high-risk classes, respectively. Based on the results of using the CODAS method, 106.90 and 222.02 km2 of the area were in high-risk and high-risk categories, respectively. In order to validate the slip potential zoning maps, using CODAS and MABAC methods, a sample of selected high-risk pixels was examined according to the set criteria. Examination of high-risk points showed that these areas were mainly in the slope of 20 to 35%. The road factor plays a very important role in creating slippery movements due to under-passing and removing the heel of the slope and changing the slope of the slopes.

The results of the study indicated the high power of Kosar city in terms of the occurrence of sliding movements. Slope factors, land use, soil, and lithology were the most important factors involved in creating landslides in the study area. Also, the results of comparing the models showed that the MABAC model provided a larger area of landslide risk than the CODAS model. It is expected that the results of this study will pave the way for better and more scientific management of competent managers and planners in this field.