TITLE:
Evaluating Geological Disaster Susceptibility Using Information Value and Neural Network Models: A Case Study of Xi Ao Town, Guangdong Province
AUTHORS:
Yinzhong Chen, Bo Tang, Jinan Qiu, Aoyang Li
KEYWORDS:
Geohazard, Susceptibility, Informativeness Method, Artificial Neural Network, Xi Ao Town
JOURNAL NAME:
Journal of Geoscience and Environment Protection,
Vol.13 No.9,
September
30,
2025
ABSTRACT: Geological disaster susceptibility assessment provides a critical foundation for disaster prevention, mitigation, and formulation of governmental reduction policies and emergency management strategies, constituting a vital component of regional sustainable development. This study evaluated geological disaster susceptibility in the mountainous area of Xi Ao Town, northern Guangdong Province, China. We integrate the Information Value Model (IVM) with a Radial Basis Function Neural Network (RBF-NN) within a geographic information systems (GIS) framework. The assessment utilizes a comprehensive susceptibility evaluation system and field survey data of potential disaster sites to identify the key influencing factors and perform spatial zoning. The results indicate: 1) Geological disasters predominantly cluster on slopes with gradients of 55˚ - 75˚, sunny slopes, at elevations between 156.7 and 321.1 m, within lithological units dominated by Early Jurassic formations, within 100 m of fault zones, within 50 m of roads, 50 - 100 m of water systems, and within 50 m of residential areas. 2) Elevation, distance to residential areas, and slope gradient were the dominant factors influencing susceptibility within the study area. 3) The majority (85.39%) of the town’s area was classified as low-susceptibility or non-susceptible. Medium-susceptibility zones accounted for 13.86% of the area, while high-susceptibility zones accounted for only 0.76%.