Sinkhole susceptibility mapping using logistic regression in KarapA±nar (Konya, Turkey)

Globally, sinkholes cause hundreds of millions of dollars in damage and hundreds of deaths or injuries each year. To mitigate the damage caused by sinkholes, it is necessary to determine the susceptible or hazardous areas. The purpose of this study is to produce a sinkhole susceptibility map based on a logistic regression method within a geographic information system environment. A field survey for this investigation identified the locations of 182 sinkholes in the study area. Many geologic, geomorphologic, hydrogeological and anthropogenic factors that influence sinkhole development were identified in the KarapA +/- nar Region. In this study, 30 sinkhole-influencing factors were selected and used in the analysis. The coefficients of the predictor variables were estimated using binary logistic regression analysis and were used to calculate the sinkhole susceptibility for the entire study area. The area value of the receiver operating characteristics curve model was 0.814. The final map indicates that most of the observed sinkholes are predicted in the high or very high sinkhole susceptibility classes. These results indicate that this model is a good estimator of sinkhole susceptibility in the study area. The sinkhole susceptibility map shows that areas with no or very low, low, moderate, high and very high sinkhole susceptibility classes are 605 km(2) (25.6 %), 310.8 km(2) (13.1 %), 531.2 km(2) (22.5 %), 487.7 km(2) (20.6 %), and 429.0 km(2) (18.1 %), respectively. Interpretation of the susceptibility map shows that sinkhole formation decreased with increasing slope angle, cover thickness, electrical conductivity, and the concentration of calcium, magnesium, sodium, and potassium ions in groundwater. However sinkhole formation increased with drainage density, fault density, upper levels of karstic formations, decline in groundwater level, and well density. This map will serve to help citizens, urban planners and design engineers prevent damage caused by existing sinkholes as well as sinkholes that develop in the future.