346
were used in conjunction with written records, a
detailed analysis of anthropogenic disturbance history could be constructed. Identifying the sequence
of historic disturbance events, including fire, logging, and conversion to farmland, makes it possible to separate these events from natural disturbances, such as windthrows, lightning-set fires, ice
damage, and floods.
A fmal limit to using historical data pertains to
the nature of the site and the disturbance regime itself. In areas subject to infrequent and catastrophic
disturbance, not only may the historical record be
too limited in duration to capture the disturbance
pattern, but more recent disturbances may erase all
trace of previous disturbances. In areas where human influence has been particularly high, evidence
of a natural disturbance regime may be lacking.
Ecosystems dominated by nonwoody vegetation,
lacking tree rings, fire scars, and structural evidence of release events, are also difficult to assess
from the historical record (Morgan et al., 1994).
Sorting out anthropogenic from natural disturbance
in some landscapes can be difficult, particularly
when past human effects alter the occurrence of
natural disturbances. A particular problem is assessing the impact of Native Americans on the natural fire regime, which, until recently, has been
generally underestimated (White et al., 1999). Only
the results of fires can be observed, and differentiating between fires ignited by lightning, Native
Americans, or Euro-American settlers can be challenging.
Limitations in availability or quality of on-site
historical data can be augmented by using more
complete data from intact similar systems. Areas
believed to have similar vegetation, topography,
climate characteristics, and unaltered disturbance
regimes can be used as models for systems for
which data are unavailable. Large wilderness areas
and public lands are often used in this manner. Intact sites are rare, however, because the effects of
fire suppression, climate alteration, air pollution,
and exotic invasion are widespread. In many parts
of the western United States, for example, the extirpations of native herbivores (e.g., bison) or the
introduction of nonnative herbivores (cattle and
sheep) have permanently altered vegetation structure on a regional scale (Billings, 1990). An additional impediment to substituting one site for another is that protected areas are often limited in their
range of ecological conditions (Morgan et al.,
1994). Most large wilderness areas in the United
States, for example, are limited to mountainous regions.
Dynamic Terrestrial Ecosystem Patterns and Processes
23.5.2 Interpreting Observational Data
Observational data typically consist of descriptive
stand data and an assessment of obvious ecosystem
trends, including overall stand structure and composition and an approximation of successional
stage. These data are usually augmented with objectively collected quantitative data, derived using
techniques such as the releve method, random
quadrats, or various plotless methods. Ecosystem
attributes that should be noted include any changes
in ecosystem components, such as the absence of
species or the presence of exotic species, as well
as changes in ecosystem processes, such as relict
fire-dependent species and the absence of species
reproduction (White and Walker, 1997). These
types of evidence are critical to understanding the
recent and long-term disturbance history of a site,
especially the interval since last disturbance and the
magnitude and frequency of regular and/or past disturbance.
An assessment of successional stages and trends
is usually an important objective in the collection
of observational data. Successional analysis may be
based on any number of factors, depending on the
ecosystem involved, but in forested stands it typically focuses on a comparison of the composition
and structure of the overs tory stratum to the composition and structure of the seedling and sapling
strata. Successional stage analysis provides a great
deal of information about the intensity of and interval since disturbance. The size, shape, and dispersal pattern of different successional patches
across the landscape convey a wealth of information about current disturbance dynamics. Although
succession is best quantified by repeated measures
over time in the same location or by analyzing historical records for the location in question, an extensive array of sampling points across the landscape can be used to both characterize typical
successional stages in areas of similar topographic
and edaphic conditions (especially when the date
of disturbance is known) and to identify aberrations
from these typical stages, which may indicate distinctive disturbance events.
Although observational data clearly offer the
most effective way to understand a particular system's present and recently past dynamic status, it
must be restated that single-time-frame observations can be misleading (White and Walker, 1997);
it is only when they are coupled with long-term
data that a realistic perception of the disturbance
regime can be constructed. Many ecosystem
processes function at nonstationary rates depend-
were used in conjunction with written records, a
detailed analysis of anthropogenic disturbance history could be constructed. Identifying the sequence
of historic disturbance events, including fire, logging, and conversion to farmland, makes it possible to separate these events from natural disturbances, such as windthrows, lightning-set fires, ice
damage, and floods.
A fmal limit to using historical data pertains to
the nature of the site and the disturbance regime itself. In areas subject to infrequent and catastrophic
disturbance, not only may the historical record be
too limited in duration to capture the disturbance
pattern, but more recent disturbances may erase all
trace of previous disturbances. In areas where human influence has been particularly high, evidence
of a natural disturbance regime may be lacking.
Ecosystems dominated by nonwoody vegetation,
lacking tree rings, fire scars, and structural evidence of release events, are also difficult to assess
from the historical record (Morgan et al., 1994).
Sorting out anthropogenic from natural disturbance
in some landscapes can be difficult, particularly
when past human effects alter the occurrence of
natural disturbances. A particular problem is assessing the impact of Native Americans on the natural fire regime, which, until recently, has been
generally underestimated (White et al., 1999). Only
the results of fires can be observed, and differentiating between fires ignited by lightning, Native
Americans, or Euro-American settlers can be challenging.
Limitations in availability or quality of on-site
historical data can be augmented by using more
complete data from intact similar systems. Areas
believed to have similar vegetation, topography,
climate characteristics, and unaltered disturbance
regimes can be used as models for systems for
which data are unavailable. Large wilderness areas
and public lands are often used in this manner. Intact sites are rare, however, because the effects of
fire suppression, climate alteration, air pollution,
and exotic invasion are widespread. In many parts
of the western United States, for example, the extirpations of native herbivores (e.g., bison) or the
introduction of nonnative herbivores (cattle and
sheep) have permanently altered vegetation structure on a regional scale (Billings, 1990). An additional impediment to substituting one site for another is that protected areas are often limited in their
range of ecological conditions (Morgan et al.,
1994). Most large wilderness areas in the United
States, for example, are limited to mountainous regions.
Dynamic Terrestrial Ecosystem Patterns and Processes
23.5.2 Interpreting Observational Data
Observational data typically consist of descriptive
stand data and an assessment of obvious ecosystem
trends, including overall stand structure and composition and an approximation of successional
stage. These data are usually augmented with objectively collected quantitative data, derived using
techniques such as the releve method, random
quadrats, or various plotless methods. Ecosystem
attributes that should be noted include any changes
in ecosystem components, such as the absence of
species or the presence of exotic species, as well
as changes in ecosystem processes, such as relict
fire-dependent species and the absence of species
reproduction (White and Walker, 1997). These
types of evidence are critical to understanding the
recent and long-term disturbance history of a site,
especially the interval since last disturbance and the
magnitude and frequency of regular and/or past disturbance.
An assessment of successional stages and trends
is usually an important objective in the collection
of observational data. Successional analysis may be
based on any number of factors, depending on the
ecosystem involved, but in forested stands it typically focuses on a comparison of the composition
and structure of the overs tory stratum to the composition and structure of the seedling and sapling
strata. Successional stage analysis provides a great
deal of information about the intensity of and interval since disturbance. The size, shape, and dispersal pattern of different successional patches
across the landscape convey a wealth of information about current disturbance dynamics. Although
succession is best quantified by repeated measures
over time in the same location or by analyzing historical records for the location in question, an extensive array of sampling points across the landscape can be used to both characterize typical
successional stages in areas of similar topographic
and edaphic conditions (especially when the date
of disturbance is known) and to identify aberrations
from these typical stages, which may indicate distinctive disturbance events.
Although observational data clearly offer the
most effective way to understand a particular system's present and recently past dynamic status, it
must be restated that single-time-frame observations can be misleading (White and Walker, 1997);
it is only when they are coupled with long-term
data that a realistic perception of the disturbance
regime can be constructed. Many ecosystem
processes function at nonstationary rates depend-