Optimized Geo-Environmental Modeling and Deep Learning for Predictive Sinkhole Susceptibility Mapping Along the Southeastern Coast of the Dead Sea

Abstract

The quick drop of the water level in the Dead Sea has expedited the formation process of dissolution depression (DDs) along the southeastern coast, which brings about serious dangers with it. The objective of this research is to create a model that will allow us to forecast the risks of DD formation by means of a unique framework based on deep learning-segmented imaging of the earth in combination with multi-criteria geospatial susceptibility modeling. The research created a dependable geospatial database containing information on 668 identified DDs. Seven geo-environmental conditions were studied using the Analytic Hierarchy Process (AHP) and were combined through weighted overlay analysis. The results show that almost 10% of the area under study falls into zones with a very high risk of DD occurrence; thus, the appropriate data were derived at the end of the study that can be used in forming an effective danger prediction tool.