Owl Research

McNemar's Tests

McNemar's tests for paired, binary response data used to test the null hypothesis that the proportion of records or detection points within vs. outside 2017 mapped modeled habitat does not differ between current and projected climate-induced conditions. The link to these tests can be found here: McNemar's tests

Model Validation

Link to simulation program can be found here: ArcGIS simulation

Description

The Owl Counts code was used to validate the statistics from the paper "Climate Change Habitat Model Forecasts for Eight Owl Species in the Southwestern U.S." (Cartron et al.). It utilizes several ArcGIS layers and model functions to generate a specified number of points and calculate the number of intersections with our target habitats.

ArcGIS Layers

• ERU_Boundary
  • This is the layer of the owl species as created in ArcGIS Pro 3.1.x. It is a single polygon of the Ecoregion for the given species. It was built using a combination of Select (Analysis Tool), Clip (Analysis Tool), and Dissolve (Data Management Tool)
• ERU_Layer
  • This is a layer consisting of all potential ERU zones as polygons. Each polygon contains the corresponding ERU values.

Other Variables

• Num_Owls
  • This is an integer value corresponding with the number of observed owls for the selected species. Using the function CreateRandomPoints, we can generate these points across our constraining layer, ERU_Boundary.
• Habitat_Values
  • This is a query statement checking our expected habitat for the given owl species. From our ERU_Layer, we select these ERU values and count the number of intersecting points from the points we generated.

Model Workflow - ArcGIS & Python (ArcPy)

• Using Select (Analysis Tool) and Clip (Analysis Tool), create a layer to be used for the boundary layer.
• Dissolve (Data Management Tool) the created layer into the ERU_Boundary layer.
• Within the .py, adjust the number of observations using Num_Owls, and set the target habitats in Habitat_Values.
• Ensure the write file, and the Workspace/Geodatabase are pathed correctly.
• Run the program and record the results

One Sample Proportion Z-tests (results)

• One-sided Z-tests were performed to test whether the habitat model performed better than the simulations, or random model. The simulation randomly generates points across eco regions (regions in which a given owl species could be found) relevant for the given owl species. The link to the Z-tests can be found here: Z-tests for each owl species