LB 36-301 - Inundation mapping of the lower Illinois river valley using synthetic aperture radar and optical satellite imagery for wetland conservation and restoration prioritization efforts

Wetlands provide critical habitat for fish and wildlife, as well as vital ecosystem services like sediment abatement, carbon sequestration, water filtration, and stormwater control. The wetlands of the Lower Illinois River Valley (LIRV) are no exception and serve as key drivers of the local terrestrial water cycle. Much of the LIRV has been degraded by years of urban development and leveed to protect its rich agricultural lands from inundation, but there remains great potential for restoration in the area. The Great Rivers Land Trust (GRLT), the National Great Rivers Research & Education Center, Principia College, and the American Geophysical Union’s (AGU) Thriving Earth Exchange seek to incorporate inundation and surface water extent layers into their process for identifying priority areas for wetland restoration in the LIRV. This project aimed to determine the feasibility of detecting inundation extent and duration along the LIRV using synthetic aperture radar (SAR).Wetlands are, in part, characterized by their inundation dynamics. The U.S. Fish and Wildlife Service (USFWS) uses high-resolution optical imagery and in situ measurements to map wetlands by detecting inundation. The low spatial resolution of in situ data and the inability of passive optical imagery to penetrate cloud and vegetation cover, though, severely limits these wetland classification methods. SAR can penetrate cloud cover and sparse vegetation, making it useful for validating water inundation metrics and detecting inundated vegetation. We used Sentinel-1 C-SAR data in the Google Earth Engine JavaScript API to classify open water and inundated vegetation along the LIRV in 2019 and 2020. The open water classification was then validated against the Dynamic Surface Water Extent (DSWE) product derived from Landsat 8 Operational Land Imager data. We successfully created layers of monthly and yearly inundation minimum and maximum extent, as well as inundation duration. The open water classification exhibited an overall accuracy of 86% when validated against DSWE classifications. The resultant classified maps also lined up with qualitative observations provided by our partners, including instances of levee breaches.These analyses will help the end users identify high-priority areas along the LIRV best suited for future land conversion. This work also further bolsters the high applicability of SAR data to wetland conservation efforts, particularly leading up to the release of L-band SAR data products from the 2024 NISAR mission.