Research Article: Mine landslide susceptibility assessment using IVM, ANN and SVM models considering the contribution of affecting factors

Date Published: April 11, 2019

Publisher: Public Library of Science

Author(s): Xiangang Luo, Feikai Lin, Shuang Zhu, Mengliang Yu, Zhuo Zhang, Lingsheng Meng, Jing Peng, Claudionor Ribeiro da Silva.


The fragile ecological environment near mines provide advantageous conditions for the development of landslides. Mine landslide susceptibility mapping is of great importance for mine geo-environment control and restoration planning. In this paper, a total of 493 landslides in Shangli County, China were collected through historical landslide inventory. 16 spectral, geomorphic and hydrological predictive factors, mainly derived from Landsat 8 imagery and Global Digital Elevation Model (ASTER GDEM), were prepared initially for landslide susceptibility assessment. Predictive capability of these factors was evaluated by using the value of variance inflation factor and information gain ratio. Three models, namely artificial neural network (ANN), support vector machine (SVM) and information value model (IVM), were applied to assess the mine landslide sensitivity. The receiver operating characteristic curve (ROC) and rank probability score were used to validate and compare the comprehensive predictive capabilities of three models involving uncertainty. Results showed that ANN model achieved higher prediction capability, proving its advantage of solve nonlinear and complex problems. Comparing the estimated landslide susceptibility map with the ground-truth one, the high-prone area tends to be located in the middle area with multiple fault distributions and the steeply sloped hill.

Partial Text

Mine landslides are common geological hazards that have caused huge loss of life and property worldwide. The loose accumulation of waste slag and lack of stable engineering facilities provide good conditions for development of mine landslides. China is most likely the country with the largest number of heavy mine tailings ponds, and mining activities have produced 20,000 km2 of mine tailing wastelands [1]. Hence, the restoration and management of mines are particularly important. Landslide susceptibility modeling (LSM) is considered as a first procedure towards susceptibility assessment, which is a spatial distribution of probabilities of landslide occurrences in a given area based on local geo-environmental factors [2]. Predicting the occurrence of landslide can avoid potential hazards and is helpful for the sustainable development of society [3].

Shangli County was selected as the study area in this paper. This area is located in the middle of China, within longitude 113°43′E−114°04′E and latitude 27°38′N−28°01′N, as shown in Fig 1. It belongs to Jiangxi Province, and the total area is about 720 km2. Besides the mountains in the central region, Shangli has rolling hills and valley plain on north and south sides. The area has a distinct four seasons and abundant rainfall. The average annual rainfall is 1300–1700 mm, and the average temperature is 4.8°C in January and 28.7°C in July. The study area is rich in mineral resources and has documented more than 26 minerals. Pingshui river, Lishui river and other tributaries originate from mountains, run across plains and hills, moisten fertile lands, and finally flow into the Xiangjiang River. Topographically, the highest elevation of the study area is 947 m and approximately 45% of the area has a slope angle of less than 20°. Geologically, the main lithology includes sandstone, shale, and limestone rocks.

In this study, landslide susceptibility assessment has been carried out in five steps (Fig 3): (1) collecting and processing data, (2) selecting suitable landslide affecting factors by feature selection method, (3) using K-folder cross validation to divide dataset for model training and testing, (4) constructing and comparing three landslide models, (5) developing landslide susceptibility map of study area. The evaluation measure system contains the receiver operating characteristic curve (ROC), rank probability score and the area percentage of each landslide susceptibility mapping. The objective is to evaluate and compare the comprehensive performance of feature selection arithmetic and three assessment methods for mine landslide susceptibility assessment. The area with high-prone landslide will be identified and the causes will be discussed as a supplement.

Mine landslide susceptibility assessment is a key step in reducing disaster risk in landslide-prone areas, especially for the restoration of abandoned mines. In this study, Shangli county was taken as a case study where more landslide disasters occurred. This study applied three widely used models including ANN, SVM and IVM to mine landslide susceptibility mapping under totally thirteen affecting factors.