Modeling temporal variations in multipolarized radar scattering from intertidal coastal wetlands

Coastal wetlands form critical ecosystems that are very sensitive to natural and anthropogenic environmental factors. These dynamic environments can exhibit both seasonal and year-to-year variability causing significant temporal variations in remotely sensed data. Understanding these variations is necessary if time series of remotely sensed data are to be used to monitor and manage coastal environments. Coastal areas can also experience persistent cloud cover, thereby making the use of microwave sensors, such as radar, potentially attractive for monitoring. Data acquired by polarimetric Synthetic Aperture Radar (SAR) and Laser Altimeter (LIDAR) sensors are analyzed to evaluate the response of a coastal marsh complex to climatic and tidal processes. The utility of L-band SAR data is investigated for detecting the presence of inundation and quantifying seasonal variations in plant phenology in several wetland environments. A coherent microwave scattering model is used to simulate radar backscatter from the marsh under different phenological and inundation conditions. l marshes can exhibit scattering behavior that is different from other environments, such as forests and inland grasslands, for which scattering phenomena have been more thoroughly studied. Using a high-resolution Digital Elevation Model (DEM) derived from LIDAR data and ancillary measurements of climatic and tidal conditions, changes in the multi-polarization L-band SAR responses over a three-year period are related to specific scattering mechanisms in the marsh. Results indicate that variations in polarization preference of herbaceous marshes are often more pronounced than those observed over upland areas. Succulent halophytic plants in the high marsh respond strongly to inundation but only mildly to climatic factors, such as average daily temperature. Conversely, the response of non-succulent halophytic plants in the low marsh is strongly dependent on climate, which in turn affects their response to inundation. Results indicate that some knowledge of the vegetation phenology is required, along with scattering simulations, for meaningful interpretation of multi-temporal SAR data over coastal wetlands.