Slope failures caused by excessive rainfall are a common occurrence in residual soils, and climate change is expected to increase precipitation. To prevent slope failures, identifying high-risk areas using slope susceptibility maps is crucial. To identify high-risk areas, regional seepage and slope stability analyses are necessary. However, due to the spatial variability of soil properties, a soil database is required before conducting distributed or GIS-based analyses. This study developed a comprehensive soil database containing both saturated and unsaturated hydraulic and mechanical soil properties using machine learning to predict unknown soil properties. The developed database is divided into 97 zones for Singapore Island, each with similar soil properties. Additionally, the study emphasizes the significance of unsaturated soil properties in slope stability analyses, and highlights the effectiveness of machine learning in predicting soil properties for accurate and reliable slope stability assessments. The developed database can be applied in regional GIS-based water balance and slope stability analyses to account for spatial heterogeneity and improve the accuracy of slope stability assessments.

unsaturated soils;slope susceptibility mapping;machine learning;GIS;Regional analyses