Landslide hazard zonation in Taleghan watershed using Shannon entropy index

Introduction
Landslide is one of the most important and disastrous natural hazards which can bring many financial losses and heavy casualties worldwide every year.
Entropy means a quantity of disorder between causes, results or decisions taken in different situations, it can be highly effective particularly in geomorphological studies such as landslide, where existing data are confused with uncertainties.
Material and Methods
Topographic maps 1: 25000 scaled, geological map of Makran and Marzanabad scale map 1: 100000 and ASAR images from ENVISAT since 2003 till 2009 in the region were used as the main data in this model. Active landslide also detected in the basin. Then, 17 zones were selected via field visiting and Google earth software images. Also location of landslides recorded using GPS. Analyzing the data and providing the required maps were done using Arc Gis10 and SAGA.
A total number of 13 effective parameters were selected based on condition of studied areas for next step. Then slide layer, slide direction, elevation, geology, land use, distance to fault, distance from the river and distance from the road obtained by Arc Gis10 analysis, vegetation layer (NDVI) measured by ENVI4.2 software and the layers of surface area ratio (ASR), topography index (TPI), Topographic Wetness Index (TWI) and slide length index (SLS ) were analyzed by SAGA software. After providing the information layers and importing the landslide locations, the properties of each slip extracted due to the mentioned layers and scoring was regarding to the role of each factor in the occurrence of slide. And the primary matrix was formed according to the entropy method. The decision matrix contains information which can be evaluated by entropy as a criterion. Then by calculating the entropy matrix and weight of 13 factors (WJ), the (HI) index will be obtained as landslide risk hazard
 
Results and discussions
At first due to the characteristics of the occurred landslide, each class was scored from the information layers which were obtained by examining the region after two field visits, study of aerial photographs and satellite imagery and identification of the affecting factors and their roles to create slip. Then the information layers used in the research were categorized and scored as a raster data and utilized as the main data in the formation of an entropy matrix for further analysis.
To convert qualitative values into quantitative a bipolar scale is used. The bipolar interval scale is a general method for ranking quantitative and qualitative indexes. Measurement in this technique is based on a 10-point scale so that zero specifies the lowest possible value which is practically comprehensible and ten represents the maximum possible value of the intended index. The middle point is also the point of dividing the scale into favorable and unfavorable data.
 
Conclusion
Prioritizing the effective factors using the Shannon entropy index indicates that data layers such as slides direction, land use, elevation, slope, normalized vegetation index and the distance from the river had the greatest effect on the landslide occurrence in the area. And the topography, moisture index, geology, distance from fault and road, and the rest of the information layers had the least effect.
Based on landslide hazard map of Taleghan watershed, very low to low risks regions are some areas in the northeast, central and southwest regions, while most of the studied areas have a moderate to very high risks. In general, site locations which could be at risk in the event of a landslide are limited to residential areas, roads, rivers, lakes and power lines. There are various land use in the region and Residential and demographic areas are at the top of the priority list. One the other hand there is Taleghan dam which can create a much greater potential risk if landslide happens at its location.