324
Elements of Ecological Land Classifications for Ecological Assessments
(Bailey, 1976, 1995; Austin and Yapp, 1978; Walter, 1979; Rowe and Sheard, 1981; Omernik, 1987;
Zonneveld, 1989; ECOMAP, 1993; Bailey et aI.,
1994; Maxwell et aI., 1995; Jensen et aI., 1997).
The main weakness of this approach is that the
hypothesized ecological relationships between the
biotic components and the indirect variables are not
always explicitly stated and tested. For example,
landform patterns have often been used to stratify
areas into natural landscape units to represent
natural levels of ecosystem integration with respect
to environmental regimes and key ecosystem
processes, and indeed there is compelling evidence
for such relationships (Swanson et aI., 1988; Ligon
et aI., 1995). Geomorphic pattern, through erosion-sedimentation processes, has been shown to
control carbon, nitrogen, and phosphorus cycles in
soils of riparian forests in southern France (Pinay
et aI., 1992). However, Kovalchik and Chitwood
(1990) used geomorphology in addition to a floristic classification of the vegetation of riparian zones
in central Oregon, but did not explicitly test the purported vegetation-geomorphological process relationship.
On the other hand, the strength of this approach
is that it allows a direct analysis of the spatial and
temporal scales of landscape features (Urban et aI.,
1987; De1court and De1court, 1988; Zonneveld,
1989), which is necessary to match patterns and
processes (Levin, 1992; see also Chapter 3). A terrestrial example is provided in a mountainous landscape in northwestern Montana by Lathrop and Peterson (1992). They established that watershed
morphological characteristics exhibited the same
basic properties at various spatial scales (selfsimilarity).
A third approach is based on the argument that,
to be meaningful, ELCs should be based on pattern-process relations (Hutchinson, 1959; Whittaker, 1972; Nix, 1982; Brown, 1984). Therefore,
the need is to estimate a pattern or group of patterns' responses to a limited set of dominant direct
variables (Nix, 1982). Direct variables are defined
as factors that have a direct influence on the distribution of the pattern(s) (definition after Austin et
aI., 1984; Austin, 1985; Austin and Smith, 1989).
For example, direct variables for species comprise
their primary niche dimensions, such as temperature, radiation, moisture, and nutrient regimes for
plants (Nix, 1982). Direct variables (Figure 22.1e)
have been used to generate ELUs (Mackey et aI.,
1989; DeVelice et aI., 1994). The definition of a
direct variable depends on the pattern of interest.
Regardless of the approach taken, the accuracy
and utility of describing and mapping ELUs are
functions of the variables selected, the strength of
the purported relationship between the biota and
these variables, the relationship between indirect
and direct variables, and the estimation procedures
and mapping scales used. Ideally, the interactions
between landscape features, climatic factors, and
ecologically meaningful variables should be obtained using a combination of geographical information systems (GIS) and simple process models.
In practice during an EA, testing biotic-abiotic relationships and the interactions between direct and
indirect variables is generally difficult, time consurning, and costly. Consequently, until recently,
indirect variables have served, largely untested, as
surrogates for direct variables in the delineation of
ELUs. An example of an ELC framework that incorporates the elements described in this section,
including testing, is presented in Perera et al. (1996)
for Ontario, Canada. A close examination of ELCs,
including evaluation of their performance, is necessary for validating their accuracy and usefulness
(see Section 22.4).
22.3 General Applications of
Ecological Land
Classifications in
Ecological Assessments
Table 22.1 summarizes some of the general areas
of application of ELCs in EAs. The specific application implemented depends on the EA objectives,
the spatiotemporal scale of the data (see Chapters
TABLE 22.1. Example of potential areas of application
of ecological land classifications in ecological
assessments.
Delineating geographic areas
Assessment area
Characterization area
Analysis area
Cumulative impact area
Reporting unit
Basic characterization unit
Integrated response unit
Determining geographic-spatial context for defining ecosystem properties
Characterizing historic range of variability
Assessing ecological and socioeconomic conditions and pattern and process persistence
Assessing potential for restoration
Assessing suitability for specific purposes
Providing environmental stratification for strategic surveys
Setting priorities for monitoring
Selecting monitoring sites
Elements of Ecological Land Classifications for Ecological Assessments
(Bailey, 1976, 1995; Austin and Yapp, 1978; Walter, 1979; Rowe and Sheard, 1981; Omernik, 1987;
Zonneveld, 1989; ECOMAP, 1993; Bailey et aI.,
1994; Maxwell et aI., 1995; Jensen et aI., 1997).
The main weakness of this approach is that the
hypothesized ecological relationships between the
biotic components and the indirect variables are not
always explicitly stated and tested. For example,
landform patterns have often been used to stratify
areas into natural landscape units to represent
natural levels of ecosystem integration with respect
to environmental regimes and key ecosystem
processes, and indeed there is compelling evidence
for such relationships (Swanson et aI., 1988; Ligon
et aI., 1995). Geomorphic pattern, through erosion-sedimentation processes, has been shown to
control carbon, nitrogen, and phosphorus cycles in
soils of riparian forests in southern France (Pinay
et aI., 1992). However, Kovalchik and Chitwood
(1990) used geomorphology in addition to a floristic classification of the vegetation of riparian zones
in central Oregon, but did not explicitly test the purported vegetation-geomorphological process relationship.
On the other hand, the strength of this approach
is that it allows a direct analysis of the spatial and
temporal scales of landscape features (Urban et aI.,
1987; De1court and De1court, 1988; Zonneveld,
1989), which is necessary to match patterns and
processes (Levin, 1992; see also Chapter 3). A terrestrial example is provided in a mountainous landscape in northwestern Montana by Lathrop and Peterson (1992). They established that watershed
morphological characteristics exhibited the same
basic properties at various spatial scales (selfsimilarity).
A third approach is based on the argument that,
to be meaningful, ELCs should be based on pattern-process relations (Hutchinson, 1959; Whittaker, 1972; Nix, 1982; Brown, 1984). Therefore,
the need is to estimate a pattern or group of patterns' responses to a limited set of dominant direct
variables (Nix, 1982). Direct variables are defined
as factors that have a direct influence on the distribution of the pattern(s) (definition after Austin et
aI., 1984; Austin, 1985; Austin and Smith, 1989).
For example, direct variables for species comprise
their primary niche dimensions, such as temperature, radiation, moisture, and nutrient regimes for
plants (Nix, 1982). Direct variables (Figure 22.1e)
have been used to generate ELUs (Mackey et aI.,
1989; DeVelice et aI., 1994). The definition of a
direct variable depends on the pattern of interest.
Regardless of the approach taken, the accuracy
and utility of describing and mapping ELUs are
functions of the variables selected, the strength of
the purported relationship between the biota and
these variables, the relationship between indirect
and direct variables, and the estimation procedures
and mapping scales used. Ideally, the interactions
between landscape features, climatic factors, and
ecologically meaningful variables should be obtained using a combination of geographical information systems (GIS) and simple process models.
In practice during an EA, testing biotic-abiotic relationships and the interactions between direct and
indirect variables is generally difficult, time consurning, and costly. Consequently, until recently,
indirect variables have served, largely untested, as
surrogates for direct variables in the delineation of
ELUs. An example of an ELC framework that incorporates the elements described in this section,
including testing, is presented in Perera et al. (1996)
for Ontario, Canada. A close examination of ELCs,
including evaluation of their performance, is necessary for validating their accuracy and usefulness
(see Section 22.4).
22.3 General Applications of
Ecological Land
Classifications in
Ecological Assessments
Table 22.1 summarizes some of the general areas
of application of ELCs in EAs. The specific application implemented depends on the EA objectives,
the spatiotemporal scale of the data (see Chapters
TABLE 22.1. Example of potential areas of application
of ecological land classifications in ecological
assessments.
Delineating geographic areas
Assessment area
Characterization area
Analysis area
Cumulative impact area
Reporting unit
Basic characterization unit
Integrated response unit
Determining geographic-spatial context for defining ecosystem properties
Characterizing historic range of variability
Assessing ecological and socioeconomic conditions and pattern and process persistence
Assessing potential for restoration
Assessing suitability for specific purposes
Providing environmental stratification for strategic surveys
Setting priorities for monitoring
Selecting monitoring sites