Sinkhole Monitoring and Early Warning Detection in the Karst Landscapes of Guimaras Island, Philippines Using Interferometric Synthetic Aperture Radar (InSAR) Satellite Data

Abstract. Guimaras Island in the Philippines is underlain by karst landscapes that pose a risk of ground subsidence and potential sinkhole formation, particularly in the municipalities of Buenavista, Jordan, and Nueva Valencia. This study explores the use of Interferometric Synthetic Aperture Radar (InSAR), specifically the Persistent Scatterer Interferometry (PSI) technique, to monitor ground displacement and identify early signs of karst-related subsidence. Sentinel-1 data from 2024 to 2025 was processed using the StaMPS workflow, generating 26,529 persistent scatterer (PS) points primarily over built-up and urban areas. A total of 3,807 subsiding scatterers were identified, with mean vertical displacement velocities reaching up to –36 mm/year in Nueva Valencia. Bivariate Ripley’s K-function analysis revealed that subsiding PS points deviate from spatial randomness, but they exhibit negative spatial association with known sinkhole locations. This is attributed to the limitations of PSI, as a huge amount of points in the sinkhole inventory are located in vegetated areas where persistent scatterers are sparse which limits PSI coherence and detection capability. This emphasizes the need for ground validation to effectively detect and distinguish sinkhole hazards from other subsidence-related events in the study area. Through DBSCAN clustering analysis, areas with high localized subsidence were observed near infrastructure such as schools, barangay halls, and coastal residential areas. Generally, while the findings exhibited limitations in sinkhole detection capabilities of PSI, the methods were effective in detecting signs of localized subsidence in built-up areas in Guimaras. This provides a foundation for developing early warning and risk mitigation strategies against subsidence-related hazards in the area.