This dataset was compiled using data on bathymetry, substrate type, geomorphic features on the sea floor, and data on the distribution of biogenic features such as kelp, eelgrass, and sponge reefs.

Bathymetry: The method used to create the bathymetry dataset was compiled using the 3-arc second (~80m) bathymetry Digital Elevation Model (DEM) from NOAA (Carignan et al., 2013), the 100m bathymetry dataset from the BCMCA Atlas (2011), and a 250m resolution subset of the GMRT v3 global bathymetric dataset (Ryan et al., 2009b). All three datasets were projected to BC Albers NAD 83, clipped to a box slightly bigger than the BC EEZ. and resampled to 80m. The data were then combined using the ArcGIS Spatial Analyst Con (conditional) tool and Raster Calculator to mosaic the datasets together, so that the NOAA 80m overwrites the BCMCA dataset, which in turn overwrites GMRTv3. Once combined, the mosaic was resampled to 100m to reduce any sharp edges between the source layers. A Canadian Pacific EEZ feature class was buffered by 100m, the final 100m mosaic was then clipped to the buffered Canadian Pacific EEZ feature class using the ArcGIS Extract by Mask Spatial Analyst Tool. The 100m mosaic was reclassified into the different depth classes (shallow <30m, shelf 30m-200m, slope 200m-2000m, deep >2000m), and converted into polygons. The final polygon dataset was then clipped with the Pacific Ocean polygon feature class to limit the dataset to the Canadian EEZ.

Substrate: The substrate dataset was compiled using the BCMCA Atlas benthic classes (2011), the DFO 100m substrate model (Fields et al., 2018 ; Gregr et al. 2021), the Natural Resources Canada (NRCan) surficial geology polygons for Kitimat and Douglas Channel (Shaw & Lintern, 2016), intertidal polygons created by Ian Murfitt (DFO, n.d.) based on CHS charts, and data on the distributions of kelp and eelgrass (British Columbia Marine Conservation Analysis (BCMCA) Project Team, 2011; Proudfoot & Robb, 2021). The 100m substrate model was reclassified to three substrate classes: Soft, Mixed and Hard substrates, and converted to polygons. In order to smooth out the edges of the zones, the raster was resampled to a higher resolution of 50m and run through the Spatial Analyst Boundary Clean Tool. Mixed substrates were generally smaller in area, so the Spatial Analyst Expand tool was applied, increasing Mixed substrate zones by 2 cells to ensure that these smaller zones did not disappear into the larger zones on conversion to polygons. The raster was then converted to polygons (not simplified) using the Raster to Polygon tool and smoothed using the Smooth Shared Edges tool with the PAEK smoothing algorithm and a smoothing tolerance of 1000m. Once the DFO 100m substrates were converted into smoothed polygons, the dataset was combined with substrate information from the BCMCA benthic classes, the NRCan surficial polygons, and substrate information from the Murfitt intertidal polygons.

The NRCan surficial polygons were classified as soft/mixed/hard based on text descriptions provided (see Appendix for details), and used to refine the substrate model. In areas where substrate is undefined in the NRCan surficial polygons (e.g. dredged areas and bioherms), the data from the DFO 100m substrate model was prioritized. The Murfitt intertidal classes were also classified as soft/mixed/hard based on FCODES (see Table 4), and used to define substrate for areas not covered by the NRCan surficial morphology polygons. Biogenic features such as kelp, eelgrass, and sponge reefs were also used to refine substrate information. Areas where kelp were present were assumed to be hard substrate, while areas with eelgrass and sponge reefs were present were assumed to be soft substrate. Areas where eelgrass and kelp were found to overlap were assumed to be mixed substrate.

Special features: The seamounts data were available from 2 datasets generated by DFO. Seamounts are underwater mountains with peaks >1 km above the base. The polygon boundaries from the latest seamount modeling were provided by Cherisse Du Preez (DFO, 2021). Underwater ridges and knolls were taken from the Manson Underwater Features dataset (2009), but were reclassified using the following depth classifications – Knolls 500m-100m, Hills <500m (Cherisse Du Preez, DFO, pers. comms.). One exception were the Dellwood Knolls, which were classified as Knolls even though the recorded height in the dataset was > 1000 m. Features identified as Ridges in the Manson Underwater Features dataset (2009) were kept as Ridges. The canyon features were taken from the global geomorphic features dataset generated by Harris et al. (2014). All canyon features (blind, and shelf-incising) within the Pacific EEZ boundary were included in the dataset. The canyon features from the Manson Underwater Features dataset (2009) were not used because those features included only canyons at the shelf break, and not beyond.

Biogenic Features: Canopy forming kelp – Bull Kelp (Nereocystis luetkeana) and Giant Kelp (Macrocystis pyrifera) – features were obtained from the BCMCA Atlas (2011). These were compiled from various sources from 1897-2008. Survey effort is not consistent across the entire study area. Some areas may be over/underrepresented and some areas with no data may not have been surveyed. Areas where kelp is present are assumed to be shallow (<30m deep), and have hard substrates. Eelgrass features (Zostera marina) were obtained from a dataset representing eelgrass beds in the Pacific Region compiled from various sources (Proudfoot & Robb, 2021). The eelgrass beds feature class originally contained survey features and buffer polygons of different sizes to represent eelgrass areas where only point and line data were available. In some areas, there was a significant amount of overlap between survey polygons and buffer polygons. To preserve the higher level detail where survey polygons were available, we used only the features representing survey polygons and included only buffer polygons that did not overlap with survey polygons. This dataset was built on prior efforts to compile a coastwide eelgrass dataset (BCMCA, 2011) and includes data from various studies and surveys. Some areas may be over/underrepresented and some areas with no data may not have been surveyed. Areas where eelgrass is present are assumed to be shallow (<30m deep) and have soft substrates. The sponge reef features were compiled by Dunham (2018) from features mapped by NRCan and Fisheries and Oceans Canada. The sponges are known to trap fine sediments, forming large bioherms or reef mounds over the span of centuries (Dunham 2018); therefore, sponge reef areas are assumed to have soft substrate.