Anthropogenic landscape disturbances, including industrial development, can have significant impacts on wildlife populations. In Canada, federal, territorial, and provincial laws require major industrial development projects to submit detailed environmental impact assessments (EIA) reports as part of the project application process. These assessments are meant to establish baseline habitat conditions and predict which landscape components will be altered by the project and to what degree. Based on these changes, indirect predictions for wildlife impacts are made using a variety of models, which can vary in validation adequacy and often rely heavily on expert opinion. In the oil sands region of Canada, wildlife species and habitat types used to make predictions are not comprehensive nor standardized between EIAs, despite a high degree of landscape similarity between projects. We extracted habitat model parameters, projected impacts, and anticipated mitigation effectiveness from 30 project EIAs. Despite all these projects occurring in the same natural region, we found very little agreement in the species used to assess wildlife impacts as well as the parameters used to model impacts on those species. Relative to unvalidated habitat models, we found that models receiving independent validation required half the habitat amount for proponents to conclude that the project will have an adverse effect. Our analyses have exposed many areas where policy could improve the efficiency of the EIA process as well as the scientific rigour underlying regulatory decisions.
Read full abstract