Cifaldi, R., J.D. Allan, J.D. Duh, and Daniel G. Brown. 2004. "Spatial Patterns in Land Cover of Exurbanizing Watersheds in Southeastern Michigan." Landscape and Urban Planning, 66(2): 107-123.
Recent research into landscape composition and configuration, or pattern, seeks to identify a core set of metrics and determine whether these describe unique gradients or dimensions of pattern across diverse settings. Prior work generally has examined relatively large units, and it is uncertain whether this approach will prove useful with small (50 to 100 km2) landscape units such as the sub-catchment of headwater streams. We estimated 25 pattern variables for the 109 sub-catchment of the Huron and Raisin river basins in southeastern Michigan, which are similar in terrain but represent, respectively, urbanizing and agricultural conditions. Three principal components analyses (PCA) performed on sub-watersheds within the combined area, and for each basin separately, identified five axes that explained about 80% of the variation in landscape pattern. The first and strongest component described a fragmentation gradient ranging from landscapes dominated by a single land cover type to more diverse, patchy landscapes, and was similar in all three analyses. Variables quantifying variation in patch size were related to the second component in each analysis. Components three through five quantified different gradients in land cover pattern among the analyses, suggesting that gradients of variation in land cover spatial patterns quantified by later components are unique to each landscape. Pattern metrics were correlated with proportion of land in a land cover class, especially for proportion agricultural and proportion urban land, which exhibited the broadest land cover gradients in the study area. Moreover, a number of relationships were non-linear, indicating that the same value for a variable could occur in two different landscapes. Overall, we find that a suite of commonly used landscape metrics typically applied to large landscape units provides a similar basis for the quantitative description of the major gradients of variation in land cover spatial patterns when applied to small landscape units.