Fetch is the unimpeded distance over which wind-driven waves can build. The methods used to create fetch are described here: https://gis-hub.ca/dataset/gridded-nearshore-fetch, but generally they are the distance in metres from a point on the ocean to the first land barrier along a given heading (in our case, 72 headings, every 5 degrees).

Effective fetch was computed from the gridded nearshore fetch source data in R code (developed by John O'Brien https://github.com/obrienjm25/REI-WaveExp/tree/main/Code ) for each site and 8 wind directions (N, NE, E, SE, S, SW, W, NW). For a given compass heading (e.g., 0 degrees), all fetch lengths within 45 degrees on either side are considered. In our case, this means that the fetch length for the given bearing, along with 9 lines on either side (every 5 degrees) are used by averaging the product of those lines multiplied by the cosine of the angle of departure from the heading. The weighting of many fetch values is meant to address shoreline geometry irregularities and follows the work of Fonseca and Bell (1998).

Modelled wind data (2015-2022) obtained from the High Resolution Deterministic Prediction System (HRDPS) - nested limited-area model (LAM) forecast grids from the Global Environmental Multiscale (GEM) model were processed using Climate Data Operators (CDO) software, R, and Python. CDO was used to calculate wind speed and direction from the zonal and meridional components of the wind (u and v). The data were binned into wind directions and subset geographically and filtered to remove data where wind speed is below the 95th percentile for a given site. Wind frequency was then determined for each of the binned wind directions. Finally, the grand mean (mean of monthly means of daily maximum values) was calculated. The end result is 8 records (binned wind directions) per location with the grand mean and frequency at that side. The frequency and grand mean values were then interpolated and resampled to raster surfaces that were used to calculate the REI values.

For each fetch/REI location, the interpolated grand mean and frequency values were extracted. Then, the relative exposure index (REI) was calculated by summing the product of the wind frequency, maximum wind speed (grand mean), and effective fetch for each direction. The resulting REI values and site geometries were saved as a REI.csv file (provided in the data package).

The REI values were converted to raster format and resampled to 20 m resolution before having a 5 cell circular focal mean applied. Next, bathymetric data was used to attenuate the REI layers by depth following the methods described by Bekkby et al. (2008). Essentially, the depth attenuation is applied by using an exponential decay function to the REI value, based on the depth value at the site location. Finally, the layers were normalized between regions from 0 to 1.