Before the arrival of European settlers, wolves inhabited the majority of the United States and played an important role in the ecosystem. As humans spread across the country, the wolf population dropped to a point of near extinction. The Eastern US lacks a significant predator since most, like the wolves, were driven out. This project was created to determine the optimal locations for grey wolf reintroduction into the Adirondack Park based on habitat suitability.

The project used the raster calculator tool in ArcGIS Pro to combine weighted values from four separate layers into a suitability index. The layers were land cover, land cover diversity, distance to roads, and food source. Previous studies and wildlife resources were used to rank the land cover types based on preferred wolf habitat.  Land cover diversity was created using the focal statistics tool in ArcGIS Pro to count the unique number of land cover types in a seven by seven window. A higher number of land cover types was more preferable for wolf habitats. The distance to roads layer was created by taking a NY state roads shapefile, and inputting it into the Euclidean Distance tool in ArcGIS Pro, resulting in a raster layer where each pixel contained its own proximity to the nearest road. Pixels closer to roads were deemed less suitable. The final layer was food source, which consisted of where deer were likely or unlikely to be found in the park. Deer population was predicted using a generalized linear regression model in ArcGIS Pro. Deer presence was the dependent variable; a binary variable that contains whether a deer was sighted somewhere (field generated data) or if it was a randomly generated point. The independent variables were elevation, biodiversity, and land cover diversity. The resulting layer provided the percentage of deer being present at each pixel based on the independent variables. The layers were each given a weight based on their importance to the suitability, and every pixel’s value from the separate layers were added together to form the habitat suitability index. 

The regression model had an accuracy of 75.25%, which was acceptable for this project, but needs improvement. The raster calculator results found that about 20% of the land, 1.15 million acres, were deemed either suitable or ideal wolf habitat. The largest of these areas were located in the southwestern region of the park. This area is less populated, contains large wilderness areas, and is at lower elevations leading to a prevalence of deer. With more time and resources this project could be improved by expanding upon the food source model, and adding a layer that includes human population data.

Data sourced from NYS GIS Clearinghouse and ESRI Living Atlas