Satellite and Street Imagery-Informed Calibration of Physics-based Slope Failure Susceptibility Maps

Physics-based predictive models for rainfall-induced slope failures consider slope geometry, soil type and properties, and precipitation to assess the susceptibility of slopes to failure. However, these models do not account for the impact of land cover (e.g., retaining structures, vegetation, wetlands, and water bodies) on slope failure susceptibility. The presence of retaining structures and vegetation can improve slope stability, while exposure of slope to water bodies can adversely degrade slope stability. Existing data sources, such as the US Department of Agriculture (USDA) and Texas Natural Resources Information System (TNRIS) databases lack sufficient land cover information, making it challenging to derive the necessary data. The objective of this research is to develop a satellite and street imagery-informed calibration method and integrate it with physics-based models to incorporate the effects of land cover on slope stability. The methodology comprises creating a formal workflow for collecting images along highway corridors through high-resolution satellite and street images, classifying the condition of land cover through assessment of these images, and developing the polygon feature layer representing the extent of different land covers. This method is integrated with the existing physics-based model by transforming the polygon feature layer into a raster file and conducting a raster overlay analysis of the slope failure susceptibility maps to update the susceptibility levels of slopes. This integration method was implemented along the highway corridors in the Houston District, Texas. Over 10,000 slope assessments were calibrated utilizing satellite and street imagery to identify instances of water-eroded and mechanically stabilized slopes. Consequently, the susceptibility levels in these locations were adjusted to account for the impact of the observed land cover conditions. This integration method enhances the accuracy of slope failure susceptibility maps, enabling transportation agencies to identify critical slope segments and plan proactive slope maintenance initiatives more effectively.