Environmental Engineering Reference
In-Depth Information
Fig. 5.8 Direct gradient
analysis in which species
distributions are plotted
along an environmental
gradient (Published with kind
permission of
©
M. Kuchta
2014. All Rights Reserved)
data sets that have linear relationships with other normally-distributed data sets.
Ecological data sets rarely have these characteristics. Species frequently respond to
environmental gradients in a non-linear manner (Fig.
5.8
), with low abundances at
the extremes of their tolerances and high abundances in the center at their ideal
conditions (called a Gaussian distribution). In addition, response curves can be
solid: even in the most favorable environments, species may not be present due to
dispersal restrictions or other factors and species abundances may range from zero
to very abundant in the most favorable conditions. In addition, we have no infor-
mation on species response to conditions beyond the range of their tolerances. Our
data sets are truncated at zero, because it is impossible for a species to have a
negative abundance (Fig.
5.8
). Additional complications arise when species
distributions are more skewed or peaked than normal. Finally, many species will
exhibit shared absence in numerous sites, creating a species by site matrix that
contains numerous zeroes. Just because two species are not present in the same site
does not mean that they respond similarly to the same environmental factors.
However, this mutual absence may produce a correlation artifact in the data.
These characteristics of ecological data can be dealt with by with data preparation
and transformation strategies that will minimize variation and maximize expressed
data structure. These strategies are beyond the scope of this chapter, but are
described in McCune and Grace (
2002
).
5.4.4 Classification of Wetland Plant Communities
Classifying wetlands, or putting them into categories, is important to effectively
manage and restore them. It is typically a first step in any study of a novel system,
essential to description. Classification facilitates conservation, and predicting future
behavior in response to environmental change. There are numerous methods of
classifying wetlands, such as the hydrogeomorphic classification system (Brinson
1993
; see Chap. 2 in Vol. 3). Wetlands, or communities within wetlands, can also
be classified on the basis of their vegetation. Classifying based upon vegetation can
