(1) Focal Patch Analysis – Under the patch mosaic model of landscape structure the focus of the
investigation may be on individual patches (instead of the aggregate properties of patches);
specifically, the spatial character and/or context of individual focal patches. This is a ‘patch-centric’;
perspective on landscape patterns in which the scope of analysis is restricted to the characterization
of individual focal patches. In this case, each focal patch is characterized according to one or more
patch-level metrics (see below). The results of a focal patch analysis is typically given in the form of
a table, where each row represents a separate patch and each column represents a separate patch
metric.
Local Neighborhood Structure – In many applications it may be appropriate to assume that organisms
experience landscape structure as local pattern gradients that vary through space according to the
perception and influence distance of the particular organism or process. Thus, instead of analyzing
global landscape patterns, e.g., as measured by conventional landscape metrics for the entire
landscape (see below), we would be better served by quantifying the local landscape pattern across
space as it may be experienced by the organism of interest, given their perceptual abilities. The local
landscape structure can be examined by passing a ‘moving window’; of fixed or variable size across
the landscape one cell at a time. The window size and form should be selected such that it reflects
14
the scale and manner in which the organism perceives or responds to pattern. If this is unknown,
the user can vary the size of the window over several runs and empirically determine which scale the
organism is most responsive to. The window moves over the landscape one cell at a time, calculating
the selected metric within the window and returning that value to the center cell. The result is a
continuous surface which reflects how an organism of that perceptual ability would perceive the
structure of the landscape as measured by that metric. The surface then would be available for
combination with other such surfaces in multivariate models to predict, for example, the distribution
and abundance of an organism continuously across the landscape.