Ecosystems and Spatial Patterns - Ecological Systems - page 102

Because spatial pattern analysis is often interested in inferring the processes that
created them, it is important to recognize that any single observed pattern represents
but one realization of the stochastic process(es) that generated it [ 7 , 9 ]. By
acknowledging that an observed pattern is but a single “snapshot,” its temporal
dimension is recognized and that under different circumstances, the patterns seen
may not be exactly the same. Hence, a main objective of studying spatial pattern is
to try to tease apart stochastic processes and the patterns they create from their
spatiotemporal conditionalities.
In addition to being driven by processes that are stochastic, patterns emerge as
a result of multiple processes that operate at different spatial and temporal scales
[ 10 , 11 ]. These processes can be biotic or abiotic and are usually interconnected
through dynamic, and occasionally nonlinear, feedback loops. For example, emer-
gent spatial pattern following forest fires is conditional on the initial distribution of
forest fuels as well as fire-weather conditions [ 12 ]. Similarly, patterns in forest
vegetation composition are often related to patterns in abiotic factors such as
moisture, drainage, and soil conditions. Patterns in the genetic composition in
animal populations have also been shown to be influenced by the environmental
variation (e.g., suitable vs. unsuitable habitat) between sampled populations [ 13 ].
These relationships are often nonlinear as the patterns that result from the
interactions among pattern-generating processes tend to be different than any one
process on its own [ 14 , 15 ].
Novel spatial patterns created by contemporary anthropogenic processes have
uncertain consequences for natural ecosystem dynamics. Anthropogenic processes
including deforestation, development, land use change, and climate change do not
replace natural processes, but have the capacity to interact with and alter them. As
such, a significant question in modern ecology and ecosystem science is that of
what are the effects of such novel patterns and processes on natural, or historical,
system dynamics [ 16 - 18 ]. Not only do new sources of spatial variability influence
natural dynamics through changing the patterns to which natural processes respond,
but they also can alter the processes themselves. For example, with regard to forest
fire dynamics, this is true where forest composition has been changed due to fire
suppression and management (pattern change) and fire frequencies are increased
due to increased ignitions near roads or changes in local weather patterns (process
change). Similarly, with regard to animal population dynamics, movement and
dispersal may be impeded through habitat loss and fragmentation (pattern change)
and habitat loss can have an absolute effect on effective population size, rates of
dispersal, and genetic variability (process change). Sophisticated spatial statistical
analyses are required to begin to disentangle the contributions of different processes
to observed spatial patterns to understand how best to manage natural systems to
safeguard against further habitat-related losses to biodiversity [ 19 ].
Here the causes and consequences of spatial patterns in terrestrial forest
ecosystems are reviewed with particular emphasis on patterns of forest vegetation
generated through landscape level disturbance processes. Spatial patterns in forest
vegetation are both ecologically and economically important in that they are directly
relevant to wildlife habitat supply, timber supply, future disturbance dynamics,