The FGS used a modeling technique called Weights of Evidence (WofE) that involves the combination of diverse spatial data to describe and analyze interactions and generate predictive models from which a map of favorability can be produced. The project began with a one year pilot study in Columbia, Hamilton, and Suwannee Counties, during which methodologies were developed in preparation to model the entire state. To train and validate the model, locations of sinkholes were required. Over two-and-a-half-years, field teams traversed the state investigating over 3,600 points of interest (potential sinkholes) and mapped 654 sinkholes. After evaluating fourteen different spatial data types, the three statistically strong spatial data layers were used to model the favorability of the State’s geology to sinkhole formation. The resulting map depicts four classes representing areas where the geology is least favorable to most favorable to sinkhole formation.

The FGS used a modeling technique called Weights of Evidence (WofE) that involves the combination of diverse spatial data to describe and analyze interactions and generate predictive models from which a map of favorability can be produced. The project began with a one year pilot study in Columbia, Hamilton, and Suwannee Counties, during which methodologies were developed in preparation to model the entire state. To train and validate the model, locations of sinkholes were required. Over two-and-a-half-years, field teams traversed the state investigating over 3,600 points of interest (potential sinkholes) and mapped 654 sinkholes. After evaluating fourteen different spatial data types, the three statistically strong spatial data layers were used to model the favorability of the State’s geology to sinkhole formation. The resulting map depicts four classes representing areas where the geology is least favorable to most favorable to sinkhole formation.