GIS-Based Risk Assessment to Karst Susceptibility of Shenzhen Universiade Center (China)

Abstract

Abstract Shenzhen Universiade Center is located in Longgang District, Shenzhen city, Guangdong province. There are various size karst cave and soil cave on site and their buried depth are from 2m to 50m. With the application of GIS technology and statistical techniques, three assessment models were established to karst spatial distribution of Shenzhen Universiade Center. Model I was the Logistic model based on geophysical information with dependent variables as anomalous areas revealed by geophysical prospecting. Model II was the Logistic model based on drilling information with dependent variables as karst caves and soil caves revealed by drilling. The independent variables of both models were the geological factors affecting karst susceptibility, including lithology, groundwater depth and the distance away from tectonic lines. Model III was the Bayesian model with integrated information of Model I and Model II. The correct probability of the model I is 58.17% when karst occurs; when does not occur karst, the correct probability is 86.75%. The fitness rate is 90.75%. The correct probability of the model II is 60.34% when karst occurs; when does not occur karst, the correct probability is 79.13%. The fitness rate is 88.38%. The result shows that Bayesian model can establish integrated assessment model including multi- source data to a certain extent, which can be employed in karst susceptibility assessment. The results of Logistic model and Bayesian model can be used as evidence for karst ground collapse vulnerability in the site.