Using Dempster–Shafer theory to model earthquake events

• Published in: Natural hazards (Dordrecht) 2020-09, Vol.103 (2), p.1943-1959
• Main Authors: Mokarram, Marzieh, Pourghasemi, Hamid Reza, Tiefenbacher, John P.
• Format: Article
• Language: English
• Subjects: Civil Engineering; Confidence intervals; Dempster-Shafer Method; Distance; Earth and Environmental Science; Earth Sciences; Earthquake damage; Earthquake prediction; Earthquakes; Economic conditions; Economics; Elevation; Environmental Management; Geographic information systems; Geographical information systems; Geophysics/Geodesy; Geotechnical Engineering & Applied Earth Sciences; Hydrogeology; Information systems; Land use; Lithology; Model accuracy; Natural Hazards; Original Paper; Remote sensing; Seismic activity; Vulnerability; Zoning
• ISSN: 0921-030X; 1573-0840
• DOI: 10.1007/s11069-020-04066-w

In this study, Dempster–Shafer theory (DST) is integrated into a geographic information system to model vulnerability of the land surface to earthquake events in northwestern Kermanshah Province, Iran, to predict where damage is most likely to occur. DST has never been used to spatially model earthquake vulnerability. To achieve this, data layers for several environmental attributes—aspect, elevation, lithology, slope angle, land use, distance from river courses, distance from roads, and distance from faults—were compiled in ArcGIS 10.2.2 software. Using membership functions, fuzzy maps were generated for each parameter. These fuzzy maps provided input data for the DST model. The predicted values were analyzed and compared at three confidence levels to determine the effectiveness of the model. The results are that 11.14%, 14.14%, and 17.18% (95%, 99%, and 99.5% confidence levels, respectively) of the study area are predicted to be susceptible to earthquakes based on receiver operating characteristic curves. The results also show that, according to the area under the curve (AUC) values (0.967, 0.828, and 0.849 for 95%, 99%, and 99.5% confidence levels, respectively), DST model generates earthquake zoning maps with high accuracy. Therefore, this model can be used for generating earthquake zoning maps with confidence levels that best suit the economic conditions and significance of the region.