Optimizing wildlife habitat mitigation with a habitat defragmentation algorithm

Habitat fragmentation is being increasingly recognized as a serious problem for a variety of wildlife species. While it is possible to approximately determine by eye where on a map a habitat alteration could be used to decrease fragmentation, this is a slow and imprecise method that is impractical for large maps. An algorithm is presented that automates this task. The method is based on concepts of diffusion-based chemical signaling. Many organisms use chemical signals spread by diffusion to detect prey or to find conspecifics. Based on this concept, a ‘scent' is artificially generated for each unit of wildlife habitat and allowed to diffuse randomly. This creates a gradient around all habitat patches. A grid square located between two habitat patches that are close together will have a high concentration of ‘scent' and will be a candidate for converting to wildlife habitat to increase connectivity or to decrease edge. The algorithm chooses for conversion those squares with the highest concentration of ‘scent' that are not already habitat. The algorithm is shown to produce least-cost corridors to connect two patches, to fill in holes in an existing patch, and to decrease edge/interior ratios.