Term of Award
Fall 2025
Degree Name
Master of Science, Applied Geography
Document Type and Release Option
Thesis (open access)
Copyright Statement / License for Reuse
This work is licensed under a Creative Commons Attribution 4.0 License.
Department
Department of Geology and Geography
Committee Chair
Christine Hladik
Committee Member 1
Jacque Kelly
Committee Member 2
Risa Cohen
Abstract
Salt marshes, essential for coastal protection and carbon sequestration, are increasingly vulnerable to dieback events, threatening ecosystem resilience and vital services. Detecting these shifts early is essential for timely intervention. However, few studies have applied Early Warning Signals (EWS) to coastal marsh systems. Most EWS research has focused on lakes, forests, or climate tipping points, with limited application to salt marsh dieback in the southeastern U.S. Site-specific, long-term spatial analyses are also lacking, as prior work often examines short time frames or single disturbance events. This study investigates the spatiotemporal patterns of salt marsh dieback on the Georgia coast from 1984 to 2024, utilizing satellite remote sensing data from Landsats 5, 7, 8, and 9. By analyzing multiple vegetation indices, including Normalized Difference Vegetation Index, Near Infrared Reflectance of Vegetation, Modified Soil Adjusted Vegetation Index, Visible Atmospherically Resistance Index, Normalized Difference Water Index, and Automated Water Extraction Index, and validating findings with field data from Kelly and Hladik (2017) as well as Google Earth imagery, this research identified early indicators of dieback using the spatial warnings package in R. Variability in these indicators was used to detect marsh resilience and shifts to pinpoint critical temporal thresholds indicating dieback onset and recovery. Additionally, climatic factors such as drought, precipitation, temperature, and river discharge were assessed to determine their role in marsh vulnerability to dieback. Results show that variance and spatial autocorrelation (Moran's I) were the most sensitive and consistent EWS indicators across all sites, with peaks in the early 1990s, 2009, and during the 2012-2014 period aligning with known dieback events. Minimum temperature emerged as the most consistent climate predictor across all sites, with drought severity also strongly associated with EWS shifts, suggesting that warming and water stress are primary triggers of marsh vulnerability. This work enhances early detection and monitoring methods for marsh health, supporting conservation efforts and improving resilience under environmental conditions.
Recommended Citation
Obeng, E. B. (2025). Evaluating spatiotemporal vegetation index variation to detect salt marsh dieback on the Georgia coast (Master’s thesis, Georgia Southern University). Georgia Southern University Electronic Theses and Dissertations.
Research Data and Supplementary Material
Yes
Included in
Environmental Health and Protection Commons, Environmental Monitoring Commons, Geographic Information Sciences Commons, Natural Resources and Conservation Commons, Remote Sensing Commons, Spatial Science Commons, Sustainability Commons
