Skip to main content
U.S. flag

An official website of the United States government

Evaluation of Offshore Wave Climates along the Alaskan North Slope using K-means Clustering

Presentation Date
Thursday, December 16, 2021 at 3:02pm
Convention Center - Room 225-227



Erosion and flooding impacts to Arctic coastal environments are intensifying due to increasingly energetic and warm ocean conditions coupling with warming terrestrial permafrost. Nearshore wave propagation models that simulate wave transformation processes to establish onshore conditions from deep water require sufficient boundary conditions. The present study interprets wave, water level, wind, temperature, and salinity data from multiple offshore sites along the North Slope of Alaska to determine a representative subset of boundary conditions in a set of location-independent typologies. We used WAVEWATCH III and Delft3D Flexible Mesh model output from six oceanographic sites located along a constant ~50m bathymetric line spanning the Chukchi to Beaufort Seas to develop offshore, location-independent typologies. K-means clustering was applied to the energy-weighted joint-probability distribution of significant wave height (Hs) and peak period (Tp) from each site to identify location-dependent cluster centroids for six sea states. Location-independent centroids were determined from the location-dependent spread and subsequent membership to the location-independent centroids was found for each site. Distributions of wave and wind direction, wind speed, and water level associated with these sea states at all sites were assessed to assign single values to describe a simplified, typological rendition of Arctic sea states during both Historic (2007-2019) and Future (2020-2040) timespans. Reanalysis data (e.g. ASRv2, ERA5, and GOFS) grounded the historic simulations while projected conditions were obtained from downscaled GFDL-CM3 forced under RCP8.5 conditions as provided by Scenarios Network for Alaska & Arctic Planning (SNAP). This method reestablishes location dependence for each simulated typology at each site by weighting nearshore results to match their unique statistical occurrence in an average year as defined by centroid membership in the occurrence joint-probability distribution. Increasingly energetic ocean conditions are found in the Future typologies compared to the Historic. These location-independent typologies can serve as offshore boundary conditions to enable computationally efficient simulation of the nearshore along the North Slope of Alaska.

Funding Program Area(s)