Implications of error and uncertainty for an environmental planning scenario: A sensitivity analysis of GIS-based variables in a reserve design exercise

The creation of a protected area network is used as an example to demonstrate the potential effects of uncertainty and error in geographic information systems (GIS) data on our ability to make reliable land planning decisions. A graph-based model of the landscape was employed where nodes represent habitat patches, edges represent inter-node paths, and inter-node distance is measured as the least-cost path according to a resistance surface. Optimal reserve networks were generated by considering the trade-offs between area and connectivity, as measured according to correlation length. A “base case” network was first created and then varied according to the land cover categorical definitions used to extract the patch map, the resistance surface, and the cell size used to represent the raster patch and resistance surfaces. The largest recommended network was over 101% larger than the smallest recommended network. The identity of the small patches predicted to be critical stepping stones and the projected effectiveness of the alternative network configurations was widely disparate. Our results support the need for a precautionary approach to compensate for data uncertainties. Through sensitivity analysis, alternative scenarios can be created for decision-makers that highlight the most conservative options and emphasize the potential uncertainties surrounding the outputs.