Hindcast (1981-2010) and projected (2011-2100) coastal storm events, including duration, wave conditions, and storm surges in the vicinity of Arey Lagoon and Barter Island, Alaska

See Erikson others (2020) for detailed information on model development. Numerical models were used to quantify the frequency and magnitude of wave conditions, storm surge, and coastal flooding. Time- and space-varying wind and atmospheric sea-level pressure fields, generated with regional model and reanalysis products and global climate models (GCMs), were obtained from outside sources (see Source Information). Ten-meter-elevation winds were applied to two-dimensional surfaces of WaveWatchIII (WW3) and SWAN model grids. Both WW3 and SWAN models are third-generation, physics-based spectral wave averaging models; these models were used to simulate the generation and propagation of waves in the Beaufort Sea to the study site. The WW3 global grid (resolution 1 degree by 1.25 degree, maximum northern latitude of 78degrees) was used to simulate hourly time-series of deepwater waves. Waves generated with the WW3 global model were incorporated into the open boundaries (edges of the model domain) of the SWAN model to ensure that long-component waves generated from distant storms were included in the simulations. The SWAN model used three nested, curvilinear grids with a resolution ranging from 1 kilometer at the continental shelf to less than 30 meters in the area of Arey Island and Lagoon. Bathymetry and elevation data (Arey Lagoon DEM) were used to populate the grids near the study site. Storm surge and hydrodynamics were modeled with the Deltares Delft3D (D3D). Three D3D FLOW grids were used to generate storm surge. The D3D model was run in a depth-averaged mode and dynamically coupled with the SWAN model to simulate surf zone longshore currents and cross-shore wave setup. The D3D module was forced with the same wind-fields used in the SWAN simulations as well as space and time-varying sea-level pressure fields covering the entire domains. D3D grid bathymetry was developed from the same depth data sources as used for the SWAN model. Wind, sea-level pressure, and sea-ice extent from reanalysis products and GCMs were used to provide boundary conditions and force the wave and storm surge model simulations. The National Centers for Environmental Prediction (NCEP) / National Center for Atmospheric Research (NCAR) North American Regional Reanalysis (NARR) dataset provided spatial- and time-varying wind, sea-level pressure, and sea-ice extent for the years 1981 through 2010. NARR is an atmosphere and land surface hydrology dataset of the North American domain developed from a reanalysis of NCEP/NCAR’s global models. The reanalysis is based on the addition and assimilation of observational data from monitoring stations, radiosondes, satellites, aircraft reports, and other sources. The spatial and temporal resolutions of the NARR dataset are 32 kilometers and 3 hours, respectively. Projected coastal storm events were derived by downscaling atmospheric conditions of the RCP 4.5 climate scenario with the MIROC5 global climate model (GCM). Sea ice retreat distances were estimated from 4 separate GCMs: MIROC5, BCC-CSM1.1, INM-CM4, GFDL-CM3. Sea-ice extent for the hindcast period 1981-2010 were obtained from NOAA OAR ESRL PSD sea ice maps. Sea-ice extents from 2011 to the end of the 21st century were obtained from four GCMs (MIROC5, BCC-CSM1.1, INM-CM4, GFDL-CM3; Taylor and others 2012).

Models were run using atmospheric forcings described above. The 25-year return period and maximum water levels in the nearshore were calculated for the time-periods 1981-2010, 2000-2050, 2051-2100, and decadal intervals, and applied at the open boundaries of a 10mx10m resolution model to simulate overland flow. The maximum depths and extents of the modeled overland flows are presented as extreme flood maps. The vertical resolution of the model simulations is +/-1cm; the uncertainty is less than +/-30 cm for the DEM and +/-15 cm for the modeled water levels as estimated from comparisons to measured water levels in Arey Lagoon.

Decadal flood maps were produced that show the area that is flooded at least two times per month due to astronomic tides, storm surges, and sea level rise; these maps were generated to provide a sense of decadal changes in monthly flood extents that might be expected to maintain halophytic vegetation. The maps were generated by extracting the second highest nearshore modeled water level of each month that had more than 21 days (~70 percent) of open water. Median values of all such water levels for each decade were calculated and applied at the open boundary of a 10mx10m resolution model to simulate overland flow. The depths and extents of the modeled overland flows are presented as decadal flood maps. The vertical resolution of the model simulations are +/-1cm; the uncertainty is less than +/-30 cm for the DEM and +/-15 cm for the modeled.

Performed minor edits to the metadata to correct typos. No data were changed