Detection and Monitoring of Inundation with Polarimetric L-Band SAR

Abstract:

It has been known for decades that at wavelengths L-band or longer, SAR is a sensitive indicator of inundation underneath forest canopies. The high resolution detection of below-canopy inundation is difficult to accomplish at regional to continental scales using other types of remote sensing sensors, making it a compelling SAR measurement especially useful for studying wetland inundation dynamics, particularly in difficult-to-reach access, canopy-covered tropical forest environments. Most results have utilized spaceborne SAR observations with less than fully polarimetric data. Since one of the objectives of the NISAR mission is to characterize and understand the fundamental process that drives changes to ecosystems such as wetland inundated areas, we will discuss the sensitivity of L-band SAR to inundation.

We will illustrate the detection of inundation using fully polarimetric L-band SAR data from UAVSAR, NASA’s airborne SAR, over a tropical forest region in Ecuador and Peru. At the same time as the data collection, measurements were made on the ground to characterize vegetation and inundation characteristics. The field data were used to validate the results of classifying the vanZyl decomposition of the polarimetric data. We compare this classification with that possible with a reduced subset of the polarimetric observations.