| Description |
Provided here are the required input files to run a standalone wave model (Simulating Waves WAves Nearshore [SWAN]; Booij and others, 1999) on eleven model domains from the Canada-U.S. border to Norton Sound, Alaska to create a downscaled wave database (DWDB). The DWDB, in turn, can be used to reconstruct hindcast (1979-2019) and projected (2020-2050) time series at each point in the model domains see Engelstad and others, 2023 for further information on reconstruction of time-series. The model forcing files consist of reduced sets of binned wind and wave parameter combinations, hereafter termed ‘sea states’. The use of representative sea states allows for lower computational costs and follows modified methods outlined in for example Camus and others, 2011, Reguero and others, 2013, and Lucero and others, 2017. For these sea states, hourly time-series of wave heights, mean wave periods, and mean incident wave directions, wind speed and wind direction were first extracted from the ERA5 reanalysis (Hersbach and others, 2020; https://cds.climate.copernicus.eu/cdsapp#!/home) for the hindcast period (1979–2019) and from WAVEWATCHIII wave model runs (Erikson and others, 2022) driven by winds and sea ice fields from the 6th generation Coupled Model Inter-comparison Projects (CMIP6 Haarsma and others, 2016). Sea states were selected using a multivariant maximum-dissimilarity algorithm (MDA) which determined representative combinations of significant wave heights, mean wave periods, mean wave directions, wind speeds, and wind directions (for more details see Engelstad and others, 2024). The model output can be used to reconstruct time series at each grid point. The extent of each model domain can be inferred from the browse graphic. Model input files are described in the Entity and Attribute Overview section. [More]
|
