The data provide an estimate of flood extent following Hurricane Matthew (2016) across the Piedmont and Coastal Plain of North Carolina. Flooded and non-flooded regions were delineated using a random forest classification model leveraging pre- and post-storm synthetic aperture radar from the European Space Agency's Sentinel-1 sensor, in addition to topography, floodplain, and landcover data. The classification model was trained with USGS and NCDEMS high-water marks, in addition to flooded and non-flooded regions delineated from high-resolution NOAA aerial photography; the model achieved 92% accuracy against an independent withheld sample. This effort was aimed at identifying inland flooding and not storm surge. This dataset is not intended to replace North Carolina's Floodplain Mapping Program hazard projections. For additional details regarding the methods, please see the peer-reviewed publication and data and code archives referenced in the Credits below.

The data provide an estimate of flood extent following Hurricane Matthew (2016) across the Piedmont and Coastal Plain of North Carolina. Flooded and non-flooded regions were delineated using a random forest classification model leveraging pre- and post-storm synthetic aperture radar from the European Space Agency's Sentinel-1 sensor, in addition to topography, floodplain, and landcover data. The classification model was trained with USGS and NCDEMS high-water marks, in addition to flooded and non-flooded regions delineated from high-resolution NOAA aerial photography; the model achieved 92% accuracy against an independent withheld sample. This effort was aimed at identifying inland flooding and not storm surge. This dataset is not intended to replace North Carolina's Floodplain Mapping Program hazard projections. For additional details regarding the methods, please see the peer-reviewed publication and data and code archives referenced in the Credits below.