Volume 11, Issue 4 (2-2025) Journal of Spatial Analysis Environmental Hazards 2025, 11(4): 0-0 | Back to browse issues page

Mendeley

Zotero

RefWorks

Asgari S, shadfar S. Landslide risk zoning using artificial neural network (ANN) in Mishkhas watershed of Ilam. Journal of Spatial Analysis Environmental Hazards 2025; 11 (4)
URL: http://jsaeh.khu.ac.ir/article-1-3453-en.html

Landslide risk zoning using artificial neural network (ANN) in Mishkhas watershed of Ilam

Shamsollah Asgari

¹, Samad Shadfar²

1- Assistant Prof, Soil Conservation and Watershed Management Research Department, Ilam Agricultural and Natural Resources Research and Education Center, Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran. , Shamsasgari@yahoo.com
2- Associate Prof, Soil Conservation and Watershed Management Research Institute (SCWMRI), Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran.

Abstract: (178 Views)

Landslides are one of the natural hazards that threaten human life and property. A landslide may destroy tens, hundreds and maybe thousands of hectares of land in a short time. For years, this hazard has destroyed orchard lands, fields, forest areas and pastures, communication roads, and rural settlements in the Mishkhas mountainous region of Ilam province. Landslide risk zoning is necessary to control this risk in this basin. The main goal of this research is the zoning of landslide risk areas in this watershed. One of the new methods to investigate the risk of landslides is the artificial neural network method. This method has advantages over other methods, the statistical distribution of the data is independent and does not require special statistical variables. In this research, first, a landslide distribution map was prepared in the selected basin. Then, the relationship between independent variables such as slope, lithology, distance from fault, land use, distance from road network, distance from waterways, direction of slope with areas affected by landslides was investigated. After preparing the weighted maps, these layers were converted into numerical information in the ArcGIS software environment, and after standardization, they were entered into the MATLAB software, and a program with a perceptron structure was written with the learning algorithm after the error propagation. After determining the structure of the artificial neural network and its training and testing, the evaluated results and the output of the network in the geographic information systems environment became a landslide risk map. The resulting risk map was calculated into different risk zones, classification and amount of landslide in each zone. The results of the analysis of the factors showed that in the Mishkhas basin of Ilam, Asmari Formation, the slope is 10-20%, the distance from the fault is more than 500 meters, the northeast direction, the distance from waterways is more than 100 meters, fruit orchards are the most sensitive land uses and the distance from the road is more than 200 meters are the most sensitive classes to the occurrence of landslides and have the highest rate of occurrence of landslides in the basin. On the other hand, the results of landslide risk zoning using artificial neural network method showed that in Mishkhas Basin of Ilam, about 80% of landslides are in high and very high-risk zones.

Keywords: zoning, artificial neural network, perceptron, error back propagation, landslide

Type of Study: Research | Subject: Special
Received: 2024/06/20 | Accepted: 2025/03/11 | Published: 2025/03/11

Send email to the article author

Rights and permissions
	This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.