| Autor: |
Ahmed Abdalla, Siavash Shami, Mohammad Amin Shahriari, Mahdi Khoshlahjeh Azar |
| Jazyk: |
angličtina |
| Rok vydání: |
2024 |
| Předmět: |
|
| Zdroj: |
Egyptian Journal of Remote Sensing and Space Sciences, Vol 27, Iss 2, Pp 204-215 (2024) |
| Druh dokumentu: |
article |
| ISSN: |
1110-9823 |
| DOI: |
10.1016/j.ejrs.2024.02.008 |
| Popis: |
Subsidence in southeastern Louisiana is a significant geological issue caused by natural and human-induced factors like low-lying topography and groundwater pumping. Human activities also led to coastal land loss and reduced sediment supply. Satellite-based technologies such as Global Navigation Satellite Systems (GNSS) and Interferometric Synthetic Aperture Radar (InSAR) are used to monitor subsidence. Louisiana has about 130 continuously operating reference stations (CORS) monitoring subsidence statewide. GNSS provides accurate point measurements but limited spatial coverage. InSAR, however, detects ground deformation over large areas using satellite-based radar imagery. In response to this advantage, we employed Sentinel-1 SAR images from 2017 to 2021 to estimate the vertical displacement in East Baton Rouge (EBR) Parish. Significant displacement is found in urban and industrial areas, particularly in high- and medium-density residential areas. The significant subsidence area is between Denham Spring and Baton Rouge faults, where residential areas experience displacement of -0.7 to -1 cm/year. The displacement variation in land use indicates significant annual subsidence in some buildings and infrastructure. Three strategic facilities in Baton Rouge Downtown experienced displacement, with -6.1 mm/yr in Downtown, -2.99 mm/yr at Horace Wilkinson Bridge, and -4.94 mm/yr at central railway station. In addition, machine learning is employed to estimate the vertical displacement in the study area. The K-Nearest Neighbors (KNN) model provides a comprehensive understanding of subsidence estimation among the GBR (Gradient Boosting Regression), RFR (Random Forest Regression), and KNN models. Machine learning models revealed that proximity to fault lines and precipitation are the most influential factors in displacement. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
