University of Twente Student Theses
Developing a Spatio-Temporal Model to Predict InSAR-derived Hillslope Deformation
Gururani, Chakshu (2024) Developing a Spatio-Temporal Model to Predict InSAR-derived Hillslope Deformation.
|
PDF
7MB |
| Abstract: | Interferometric Synthetic Aperture Radar (InSAR) technology has revolutionized the monitoring of surface deformation, providing detailed and high-resolution data critical for assessing hillslope stability. Despite these advancements, integrating InSAR data with environmental parameters to predict deformation occurrences remains underexplored. This study aimed to bridge this gap by developing a novel spatio-temporal model to predict hillslope deformation by combining InSAR data from Sentinel-1, made available by the European Ground Motion Service (EGMS), with dynamic environmental variables obtained from ERA-5 dataset and static terrain properties from the British Geological Survey (BGS). Utilizing InSAR's capability for precise surface deformation monitoring, a model using Graph Attention Networks (GAT) and Gated Recurrent Units (GRU) was developed, capturing both spatial and temporal dependencies in the dataset. The study was conducted for central England region, including the Peak District National Park, featuring significant surface deformation activity. A correlation analysis was performed to explore relationships between dynamic variables—such as precipitation, temperature, Leaf Area Index (LAI), and terrestrial water storage (TWS)—and hillslope deformation, revealing varying correlations ranging from -0.4 to 0.6. The proposed GAT-GRU model achieved a final training loss of 0.00015. The predictions were evaluated using Mean Squared Error (MSE) and Root Mean Squared Error (RMSE). The analysis indicated that the model captured the overall spatial and temporal dynamics. However, it struggled with extreme values and regions of high variability, with significant errors in regions of rapid surface movement beyond ±90 mm. The findings demonstrate the model's ability to capture general trends and spatial patterns of deformation, including slow movements that are often overlooked in traditional landslide inventories and can lead to catastrophic failures. However, challenges remain in predicting extreme values and highly variable regions. This research advances the development of predictive tools for surface deformation monitoring and early warning systems. Future work should focus on improving model performance with higher resolution environmental data and exploring causal relationships between variables. |
| Item Type: | Essay (Master) |
| Faculty: | ITC: Faculty of Geo-information Science and Earth Observation |
| Subject: | 38 earth sciences, 43 environmental science, 54 computer science |
| Programme: | Geoinformation Science and Earth Observation MSc (75014) |
| Link to this item: | https://purl.utwente.nl/essays/100495 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page