Spatiotemporal variations of wetland backscatter: The role of water depth and vegetation characteristics in Sentinel-1 dual-polarization SAR observations

Abstract

Synthetic Aperture Radar (SAR) backscatter observations are sensitive to hydrologic conditions and vegetation characteristics of land cover. This study conducted a high spatial-resolution investigation (30-m) on the response of dual-polarization C-band (5.6 cm wavelength) SAR backscatter coefficients (σ°) to temporal changes of surface water depth (dw) and spatial variations of vegetation characteristics in the south Florida Everglades wetlands. We investigated (1) linear relationships between σ° and dw values, and (2) the effects of vegetation density and morphology on σ°-dw relationships. We developed a new method to classify pixels with significant linear relationships of multi-temporal σ° and dw (R2 > 0.5 and p-value <0.04), which were termed “Reliable Scatterer” (RS). RS included positive, negative, and a combination of both positive and negative relationships (corresponding to RS+, RS−, RS±, respectively). Our analysis revealed spatially varying vegetation densities and morphologies had a significant impact on RS types, where we found RS+ type pixels for woody vegetation, RS± for a mix of medium- and high-density herbaceous vegetation using C-band VV (C-VV) data, and RS− for sparse herbaceous vegetation using C-VH data. Overall, our study indicates that C-band dual-polarization backscatter is sensitive to water-depth variations for some vegetation types, and this sensitivity has the potential to serve as a reliable indicator for monitoring water depth in wetland environments.