Digital Elevation Models for eight study areas in coastal Oregon and Washington, 2012

To assess the current topography of tidal marsh at the study sites we conducted survey-grade global positioning system (GPS) surveys between 2009 and 2014 using a Leica RX1200 Real Time Kinematic (RTK) rover (±1 cm horizontal, ±2 cm vertical accuracy; Leica Geosystems Inc., Norcross, GA; Figure 4). At sites with RTK GPS network coverage (Padilla, Port Susan, Nisqually, Siletz, Bull Island, and Bandon), rover positions were received in real time from the Leica Smartnet system via a CDMA modem (www.lecia-geosystems.com). At sites without network coverage (Skokomish, Grays Harbor, and Willapa), rover positions were received in real time from a Leica GS10 antenna base station via radio link. At sites where we used the base station, we adjusted all elevation measurements using an OPUS correction (www.ngs.noaa.gov/OPUS). We used the WGS84 ellipsoid model for vertical and horizontal positioning and referenced positions to a local National Geodetic Survey (NGS) benchmark or a benchmark established by a surveyor (Figure 4). Average measured vertical errors at benchmarks were 1-9 cm throughout the study, comparable to the stated error of the GPS. To measure topographic variation at each site, we surveyed marsh surface elevation along transects perpendicular to the major tidal sediment source, with a survey point taken every 12.5 m; 50 m separated transect lines (Appendix Figs. A1 – I1). We used the Geoid09 model to calculate orthometric heights from ellipsoid measurements (m, NAVD88; North American Vertical Datum of 1988) and projected all points to NAD83 UTM zone 10 using Leica GeoOffice v7.0.1 (Leica Geosystems Inc, Norcross, GA).In ArcGIS 10.2.1 Spatial Analyst (ESRI 2013, Redlands, CA), we created a digital elevation model (DEM) for each site using each sites survey elevation data points. We processed the elevation point data with exponential ordinary kriging methods (5 x 5 m cell size) while adjusting model parameters to minimize the root-mean-square (RMS) error to create the best model fit for the DEM (Table 2). We used elevation models as the baseline conditions for subsequent analyses including tidal inundation patterns, SLR response modeling, and mapping of sites by specific elevation (flooding) zones.