Geoprocessing solutions developed while calculating Human Footprint™ statistics for zones representing protected areas and adjacent lands at the continent scale

Περίληψη

We calculated the mean Human Footprint™ (HF) for 196,498 polygons representing state and federally administrated protected areas (e.g., National Forests, National Parks, State and Provincial Parks, etc.) of Canada, Mexico, and the Continental United States. Separate sets of calculations were made for (1) the area in each protected area which ranged in size from less than one to over 11 million hectares and (2) the area outside and within 10 km of each protected area. We used Last of the Wild version 2 (2005) for North America as the source of data for HF values. This paper concerns the technical problems we encountered using ArcGIS 9.3 and Spatial Analyst to accomplish this task in a timely manner. We developed several scripts to automate processes and address overlapping polygons resulting from zone calculations of 10 km around each protected area (doughnut-shaped polygons defining the zones from which to calculate average HF values adjacent to protected areas). We learned that Spatial Analyst does not honor the object integrity of overlapping polygons when using them to define zones for calculating zonal statistics from a raster database. We tried alternative solutions, including the use of Hawth's Analysis Tools version 3.27 (Zonal Statistics ++) and writing scripts in Visual Basic 6.0 (VBA) to separate overlapping polygons and to calculate zonal statistics both as a table and output raster database. One of the four scripts resulting from this project was developed to calculate the 10 km zone around each protected area polygon. This script can be used to calculate a separate doughnut polygon for any distance outside of any size polygon, even if it shares boundaries with other polygons. We also discovered that the Zonal Statistics function in Spatial Analyst does not calculate all of the zones in a large database even if the polygons do not overlap. Our solution for this problem is described in this paper as an iterative process ending with another custom script to define the raster value located under the label point of each polygon in a vector database. Ultimately, we successfully calculated the mean HF from a spatially defined raster database both inside and outside the nearly 200,000 polygons defining the boundaries of Protected Areas in North America (http://cec.org/atlas).