Patch Analyst 3.1

A major revision to Patch Analyst has been made in version 3 (and continued in version 4) in that the analysis of regions, as well as the ability to relate a response variable (e.g., an interpolated surface of bird density) to landscape structure, has been added. This is more than a technical enhancement, in that it overcomes a major limitation of traditional landscape analyses.

An example: If the goal of the analysis is to relate landscape structure to the ability of an area to sustain viable wildlife populations, then the simple calculation of patch metrics has several fundamental problems. To begin, the characterization of average conditions across an unspecified scale has limited interpretation value. Consider the calculation of mean cutover (patch) size in a forest management unit. If the unit contains only very large or very small patches, then the mean patch size will be intermediate between these, and might not even occur on the landscape. A management unit with only medium sized patches would give the same result, even though the texture of the landscape was completely different.

For species that are edge loving, or edge avoiding, a principal issue is the interspersion of young and old age classes of forest in the proximity of their home-range. The size of the clearcut is essentially irrelevant. The calculation of only two statistics, the proportion of young forest and the density of edge, within a spatially delineated area, sufficiently describes this condition. If these calculations are conducted for every possible home-range on the landscape, then a spatial surface representing degree of interspersion can be created.

This example begs the question of at what scale the calculations should be conducted. The answer to this cannot be known a priori, and for this reason, the correlation space must be explored. By conducting calculations over a range of scales, the researcher can determine the relevant spatial at which to characterize interspersion.