EP042-08
Space-based monitoring of water depth and soil salinity over collapsing peat marshes using Sentinel-1 SAR amplitude observations

Friday, 11 December 2020: 11:08
Virtual
Boya Zhang1, Lukas Lamb-Wotton2, Khandker S. Ishtiaq2, Daniel Gann3, Shimon Wdowinski1, Tiffany Troxler2 and Eric D Swain4, (1)Florida International University, Earth & Environment Department, Miami, FL, United States, (2)Florida International University, Miami, FL, United States, (3)Florida International University, Biological Sciences, Miami, United States, (4)USGS Florida Water Science Center Davie, Davie, FL, United States
Recorded Presentation
Abstract:
Peatland stores a large amount of terrestrial carbon. In coastal peatland environments, saltwater intrusions in combination of droughts are associated with peat collapse leading to peatland degradation and loss of carbon. Peatlands are often inaccessible and best monitored by remote sensing observation. We evaluated the ability of spaceborne Synthetic Aperture Radar (SAR) to monitor the effects of peat collapse on water depth and soil salinity in coastal marsh.

We studied a brackish peat marsh in the Florida Coastal Everglades experiencing collapse and salinization, using remote sensing observations and ground-based model products (Fig. 1a). The remote sensing data consisted of 32 Sentinel-1 SAR amplitude (σ° in dB unit) images from April 2016 to December 2017. We obtained ground data of daily average water depths (dw) and groundwater salinity (S) products (a proxy for soil salinity) from the Biscayne and Southern Everglades Coastal Transport (BISECT) model. We used WorldView-2 data to map three land covers: sawgrass, submerged aquatic vegetation, and open water (Fig. 1b). All products were re-sampled and aligned at a 30-m grid resolution.

We investigated σ° changes in response to dw and S using correlation analyses. For the majority of the areas, σ° responded differently to dw changes, depending on the location of water table relative to the ground surface. Under inundated conditions, σ° decreased with increasing dw, which reduced emergent vegetation volume. During soil exposure (negative dw) σ° varied in a large range regardless of dw (Fig. 1c). Correlation between σ° and S was conducted only during soil exposure, when SAR signals can reach the soil surfaces. We detected the locations where σ° were negatively correlated with S, and the locations were characterized by submerged vegetation or water land covers (Fig. 1d, 1e). Overall, our results demonstrate the potential of SAR σ° in combination with vegetation maps for monitoring changes in peat marsh environments.

<< Previous Abstract | Next Abstract