Mapping Low Coastal Areas

UPDATE: Within the U.S., the Risk Zone Map and Risk Finder analyses derived from it now utilize lidar-based elevation data almost exclusively, courtesy of NOAA, but otherwise follow the methods described below. Outside the U.S., the Risk Zone Map follows analogous methods but employs Shuttle Radar Topography Mission (SRTM) elevation data from NASA, with roughly 90-meter (300-foot) horizontal resolution; a SRTM water body mask; and global tidal model data supplied by Mark Merrifield of the University of Hawaii.

For the elevation data behind our maps, we used the National Elevation Dataset (NED), a product of the U.S. Geological Survey. The NED divides the contiguous United States into a grid of tiny, roughly square cells covering its full area.  For each of the millions of trillions of cells, NED provides coordinates and an estimate of average elevation. We used the highest-resolution edition of NED that has full coverage of the coastal contiguous U.S. Cells are approximately thirty feet (ten meters) on a side; this is the finest resolution data publicly available with such extensive coverage. For discussions of the accuracy of the elevation data and what it means for our maps and statistics, see our published paper and this further explanation.

We began our process by classifying all cells as ocean (ocean, bay, estuary or saltwater wetland) or land (land or freshwater wetland), because ocean or saltwater marsh misclassified as land would lead to overestimates of susceptible total land area. We admitted NED cells as land according to a conservative consensus of three independent data sets. First, the cells had to be designated as land within the NED itself. Second, we included only cells with centers landward of NOAA’s Medium Resolution Digital Vector Shoreline. Finally, we eliminated NED cells with centers inside areas classified in the National Wetlands Inventory (NWI) as estuarine or marine wetland or deepwater. In computing total land area susceptible, we included NWI freshwater wetlands. We also computed dry land area susceptible, excluding NED cells with centers inside areas classified as freshwater wetlands. True water chemistry may vary from NWI classification.

Next, we adjusted the elevation of each cell to be in reference to the nearest average high tide line, instead of a standard zero. For example, if a cell’s elevation were five feet, but the local high tide reached three feet, then we would compute an elevation of two feet relative to the tide line. Clearly, sea level rise or a storm surge would need to reach only two feet above high tide to threaten this cell with inundation. Sea level and tidal amplitude vary sometimes widely from place to place, and therefore also the average height of high tide. For local high tide elevations, we used values of “Mean High Water” from VDatum, a NOAA data product and tidal model.

Finally, we identified the set of cells beneath each water level threshold from one to ten feet above local high tide, and drew maps of each area. We excluded the far-inland areas of Death Valley and the Salton Sea in California, which include land below sea level, but are separated from the ocean by a mountain range.

For a more detailed description of our methods, see our published paper. To read how we translated these maps into statistics of threatened land, homes and population, see this description.