The map below shows the locations of observed and predicted fish biomass and densities. When multiple observations are present on a segment, the median is shown. By default, the observed values are shown, and model predictions can be shown from the layers menu.

Predictions of observed populations are available across all tributaries to the Great Lakes and St. Lawrence River. Predictions to simulated 'Reference' landscapes are available as well. Reference landscapes were approximated by removing urban and agricultural landcovers, and proportionally increasing all natural landcovers in the catchment to account for those removed. Additionally, the LDI was set to nearly 0 for urban and agriculture landcovers in the simulated 'reference' landscapes. Finally, the difference between observed and reference communities are mapped, and expressed as (Current - Reference; i.e., positive values indicate present day predictions are higher than simulated reference, and negative values indicate present day predictions are lower than simulated reference).

Selecting a stream segment with either the 'Current' or 'Reference' layers selected will show the predictor values associated with that segment for the 'Current' or 'Reference' predictions respectively. Selecting a stream segment in the 'Current' or 'Reference' predicted layers will also present a 'SHAP Breakdown' figure showing the contributions of each predictor variable to the final predicted outcome. A positive effect on the SHAP score of the mean suggests that predictor value is increasing the mean prediction, whereas a positive effect on the SHAP score of the presence/absence suggests that the predictor values is increasing the likelihood of a presence. Once a reach is selected, that reach will also be highlighted in the 'Predictor Response' tab.