Ecosystems and Spatial Patterns - Ecological Systems - page 107

challenge resides in the fact that the single observed pattern is an amalgamation of
these multiple processes interacting with existing spatial structure and historical
legacies; the functional relationships that connect these contributors to pattern are
largely uncertain ( Fig. 7.2 ). When analyzing the spatial structure of sampled data, it
is not easy to disentangle the key spatial scales, and therefore processes, that act on
the data. However, in the last decade, hierarchical decomposition methods (multi-
scale ordination [ 47 ]; PCNM [ 48 ]; wavelets, [ 49 ] more detailed below) have been
developed to identify the spatial scales at which data are most strongly structured
and to decompose the data on the basis of scale-specific variances. Beyond simply
describing patterns and the scales at which they are structured, it is also important to
have a priori hypotheses about which scales and processes are the most relevant for
the questions under study as these methods could reveal many patterns and spatial
scales, many of which may not be of relevance [ 8 ]. Spatial pattern analysis will be
more effective at describing underlying processes when used in an explicit and
informed hypothesis testing framework.
Ecological Consequences of Spatial Heterogeneity
The consequences of changes in spatial pattern in forest landscapes are easily
confounded by absolute losses in wildlife habitat [ 19 ]. That is, although both forest
composition and configuration are important, issues related to configuration are
only relevant below a critical threshold of forest amount (usually 20-30% of area;
[ 50 ]). Above the critical threshold, the landscape generally remains “connected”
and organisms or disturbances can spread in the landscape [ 51 ]. Below such
a critical threshold, species' response to the amount of habitat area is nonlinear as
most species do not have enough habitat to meet their needs. Surrounding habitat
quality (composition) and configuration become more important for local popula-
tion persistence in this case. Moreover, fragmented landscapes with various
landcover types can impede species abilities to move from habitat patches to
another [ 52 ]. For example, nesting birds do not cross forest gaps larger than 25 m
[ 53 ]. Impediments to movement across landscapes can influence population
dynamics [ 54 ] as well as genetic heterogeneity [ 55 ], both of which affect the
probability of population persistence.
Spatial Analyses
There are three main approaches to investigating the different aspects and
consequences of spatial heterogeneity. Spatial statistics, landscape metrics, and
statistical modeling, all approach the question of identifying spatial pattern in
ecosystems in a slightly different way [ 5 ]. Owing to the varied history of
approaches to studying spatial patterns including methods and concepts drawn