Pacific Canadian Coast Shallow Substrate Model at 20-Meter Resolution

A 20-meter resolution raster model classifies shallow substrate bottom types along the Pacific Canadian coast to support near-shore habitat modeling. The model predicts four categories (Rock, Mixed, Sand, Mud) using a weighted random forest algorithm trained on observed bottom types and environmental predictors like oceanographic layers, fetch, and bathymetry. Its spatial domain is restricted to within 5 kilometers of the 50-meter depth contour.

Use Cases

• Train a geospatial classifier using the four predicted substrate categories (Rock, Mixed, Sand, Mud) to map benthic habitats.
• Analyze the relationship between bathymetry derivatives and the classified substrate type for coastal geomorphology studies.
• Integrate the substrate classification raster with oceanographic layers to model species distribution in near-shore marine environments.

Strengths

• Model provides a four-category classification (Rock, Mixed, Sand, Mud) for substrate analysis.
• Spatial resolution is specified at 20 meters for detailed coastal mapping.
• Derived from environmental predictor layers including oceanographic data, fetch, and bathymetry.

Limitations

• Spatial coverage is limited to within 5 kilometers of the 50-meter depth contour, excluding deeper offshore areas.
• Model accuracy is dependent on the quality and availability of input observation and predictor layers.

Provenance

Source

Fisheries and Oceans Canada

Collection Method

Created using weighted random forest classification from the ranger R package on observed bottom types and environmental predictors.

Freshness

Last updated on March 24, 2026.

Geography

Pacific Canadian coast, within 5 km of the 50 m depth contour.