The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Share
Publications Copernicus
Download
Citation
Share
Articles | Volume XLVIII-5/W3-2025
https://doi.org/10.5194/isprs-archives-XLVIII-5-W3-2025-125-2025
https://doi.org/10.5194/isprs-archives-XLVIII-5-W3-2025-125-2025
12 Nov 2025
 | 12 Nov 2025

Spatial-temporal Analysis of Land Subsidence in Jizan Province for 2019–2024

Roman Shults and Esubalew Adem

Keywords: Land Subsidence, Spatial-temporal Model, Prediction, SBAS, InSAR, Displacement

Abstract. This study provides a vertical displacement analysis of Jizan Province, Saudi Arabia, using InSAR-derived time-series data processed with the Short Baseline Subset (SBAS) technique. Vertical displacements were examined through a simple trajectory model and a spatiotemporal model based on distributed scatterer observations. The research focused solely on remote sensing displacement measurements from 2019 to 2024. Displacement time series were analyzed for 80 spatially distributed points across areas with varying landslide susceptibility. Since the area was not affected by earthquakes during the study period, the simple trajectory model only indicates a general trend. Additionally, the time series was very noisy, making the standard extended trajectory model highly sensitive to local variations. Unlike the spatial-temporal analysis, forecasting with the extended trajectory model was unstable and unreliable. The spatial-temporal model successfully captured patterns of settlement and uplift. Estimated vertical velocities ranged from −13 ± 2 mm/year for sediments to +9 ± 2 mm/year for uplift, observed in regions with medium and high landslide activity. The remaining areas are stable. Temporal trends, obtained without external variables, show that displacements are mostly consistent with slow deformation typical of distributed scatterers in arid terrain. This displacement-only assessment provides fundamental insight into the kinematic behavior and serves as a baseline for future data-fusion models that incorporate environmental or human factors.

Share