Projects & Initiatives

Frequently Asked Questions

Find information, guidance and troubleshooting tips.

Jeope Wolfe

Using the arrow to the left of the search bar, make sure that both options are selected (ArcGIS World Geocoding Service allows searching by city name or address; Quarter Sections allows searching by quarter section (legal land description))

It displays the predicted number of terrestrial vertebrate species (amphibians, birds, mammals, and reptiles) based on land cover and climate, at sites that are 6.44 x 6.44 kilometers (4 x 4 miles).

Anyone is welcome to use the public version of the tool. See our terms of use.

The predictions are based on individual species distribution models using species observations, habitat data, and climate data.

This tool was developed using data from species in the Prairie Ecozone of Canada. The project team is developing a similar tool for Eastern Canada.

We built distribution models for 329 species in Prairie Canada. Many species have geographic ranges smaller than the total study area so the maximum number of species at a site is lower than 329.

Urban areas in Prairie Canada typically have major rivers running through them, which are areas with high predicted biodiversity. Urban areas can include many different habitats in a small area, so contain species that use many different habitat types, like forests, rivers, and wetlands.

There are two main differences: (1) the predictions are coarser in the public tool and (2) we updated the tool predictions to use 2020 land cover, whereas the models in the paper were fit using 2015 land cover.

Land cover categories were simplified to focus on distinct habitat types that affect biodiversity patterns. We separated: wetland, open water, native grassland, tame grass, annual cropland, forest, and human settlement.

The tool is built based on land cover; for example, the tool predicts different numbers of species in fields that are annual cropland versus fields that are native grassland. The tool does not account for differences in management activities on the same habitat type, such as grazing intensity in different grassland fields.

There is inadequate information across large scales to be able to estimate and map effects of management on biodiversity.

A portion of the data from the tool was held back and used to test its accuracy. We used the mean Area Under the Receiver-Operator Curve (AUC) as an indication of model performance, and most of our species’ models are well above the AUC benchmark for performance. We also compared a subset of the data to independent wetland species presence data collected using Autonomous Recording Units (ARUs).

You can submit feedback to the research team by contacting DUC’s Institute for Wetland and Waterfowl Research.