Researchers at UMD and USGS, led by Paul Marban (M.S. MEES), and including Drs. Jennifer Murrow Mullinax (ENST), Jonathan Resop (GEOG), and Diann Prosser (USGS Patuxent Wildlife Research Center) have published a paper titled "Assessing beach and island habitat loss in the Chesapeake Bay and Delmarva coastal bay region, USA, through processing of Landsat imagery: A case study" in the journal Remote Sensing Applications: Society and Environment. The manuscript documents island loss in the Chesapeake Bay region using Landsat imagery, showing island area declining over 1200 ha from 1986 to 2016.
Beaches and islands provide economic and social value to humans and contribute critical habitat for breeding and foraging wildlife. These ecosystems, however, are being severely impacted by global climate change and sea level rise through increased erosion and frequency of inundation. The case study presented here aimed to document island loss in the Chesapeake Bay and Delmarva coastal bay region of the United States using image processing techniques within a GIS from 1986 to 2016. Satellite imagery from Landsat Thematic Mapper (TM) and Operational Land Imager (OLI) sensors were processed within ArcMap 10.5 to determine spatial and temporal trends in island and beach habitat. Calculation of unweighted Cohen's Kappa showed that classified scenes were, on average, within the range of moderate agreement between the classified Landsat scenes and the validation imagery within Google Earth Pro (0.539). From 1986 to 2016, island area declined by over 1200 hectare (ha) with agriculture/open field (all open vegetated spaces) declining by nearly 82% and beach, surprisingly, increasing nearly 2%. This study was the first to document Chesapeake Bay region-wide island loss beyond the mid-2000s. The accuracy of this study was limited slightly by the 30 m spatial resolution of the imagery. Therefore, this technique may be best suited for documenting trends on large islands and along the mainland coastline.