Predictive modeling of cave entrance locations: relationships between surface and subsurface morphology

Cave entrances directly connect the surface and subsurface geomorphology in karst landscapes. Understanding spatial distribution of these features can help identify areas on landscape that are critical to flow groundwater system. Sinkholes springs major locations inflow outflow from system, respectively, however not all sinkholes equally connected main conduit Predicting where zones high connectivity exist is a challenge because cave difficult detect imperfectly documented. Wildlife research has similar issue understanding complexities given species likely incomplete information presence-only data. Species models address some issues create accurate predictions or event occurrence across landscape. Here we apply model, MaxEnt, predict entrance three geomorphic regions Kentucky. We built with Kentucky Speleological Survey database predictor variables, including distance sinkholes, springs, faults, elevation, lithology, slope, aspect. All regional well most important variables for predicting consistent between models. Throughout models, had largest influence likelihood presence. This unique use modeling techniques shows they potentially valuable tools understand patterns other either ephemeral using standard techniques.