Comparative Assessment of Sinkhole Susceptibility Mapping in Mexico City: Weight of Evidence versus Weighted Linear Summation

Sinkholes are a geological phenomenon that appears as a closed funnel-shaped surface depression, where water can stagnate and drain into the subsoil. This phenomenon occurs mainly in karst environments, however it can also occur in multiple geological environments, generated by natural and anthropogenic processes, such as subsurface erosion, changes in groundwater levels and groundwater extraction, among others. The main distinctive feature of sinkholes is that their presence is not detectable until a sudden collapse of the surface layer of soil occurs, generating a significant risk for infrastructure and population in urban areas. Mexico City presents a critical situation due to the presence of sinkholes, since from 2017 to 2020 more than 500 sinkholes have been registered throughout the city, exposing to serious risks to the structures, roads and safety of the people who live and transit daily in the city. The aim of this study is to compare the sinkhole susceptibility maps of two methodologies: Weights of Evidence and Weighted Linear Sum. The accuracy of both methodologies will be obtained by comparing the values obtained using the Receiver Operating Characteristic (ROC) curve and the Area Under the Curve (AUC). The maps were elaborated using a GIS database made up of 18 conditioning factors (groundwater depletion, land elevation, density of waterlogging, density of faults, density of fractures, density of leaks, density of mines, density of water wells, density of natural drainage, distance to faults, distance to fractures, distance to subway lines, distance to mines, distance to roads, sinking speed, slope, lithology and land use) and the record of damage caused by sinkholes in Mexico City. Both maps show a good identification of areas susceptible to the presence of sinkholes, with the central-northern and eastern parts of the city having the greatest potential for sinkhole formation. The convergence of the results underlines the importance of the conditioning factors that contribute to the formation of sinkholes, highlighting the factors of anthropogenic origin as the main forming factors. The findings emphasize the potential of both methods to generate good urban planning and elaborate adequate risk mitigation strategies in the identified areas.