Subsidence Hazard Assessment at the Samcheok Coalfield, South Korea: A Case Study Using GIS

Abstract

This paper presents a case study of subsidence hazard assessment using geographic information systems (GIS) in an abandoned coal mine area in South Korea. A spatial database was constructed using mine drift maps, topographic maps, geologic maps, borehole data, and subsidence inventory maps representing the locations of past subsidence occurrences. Eight factor layers (drift depth, drift density, distance from nearest drift, distance from nearest railroad, rock mass rating, groundwater depth, slope, and surface runoff accumulation) were extracted from the spatial database to examine the relationships between the factors and past subsidence occurrences (training data set). A frequency ratio (FR) model was used to establish rating classes for each factor, and an analytic hierarchy process (AHP) model was used to establish weightings for the factors. The two models were integrated to combine multiple factor layers into a subsidence hazard map. When the area under the curve technique was used to verify the subsidence hazard map, by comparing the determined hazard rankings with past subsidence occurrences (validation data set), the FR model and FR-AHP integrated model showed 97 percent and 94 percent accuracies, respectively, in predicting subsidence occurrences. Finally, the subsidence hazard map based on the FR model was overlain with exposure intensity and vulnerability maps to generate a priority setting map, representing the relative risk of mine subsidence to buildings. The priority setting map can be used by planners and developers to identify and prioritize areas requiring more detailed investigations of mine subsidence hazards and risks.