23.3 Definition of Terrestrial Disturbance Regime
the assessment process. Scale has two components:
grain, the degree of resolution or smallest recognizable element, and extent, the total area or time
span over which observations are made (Turner et
aI., 1989; Wiens, 1989). Varying either the grain
or extent of a sample alters our observations. Increasing grain size tends to reduce observed variance among samples because each sample is larger
and encompasses more variation internally; increasing extent tends to increase observed variance
because, in larger areas, more kinds of environmental conditions and events are observed (Wiens,
1989).
When conducting ecological assessments, it is
preferable to avoid a single arbitrary choice of scale
based purely on expediency (Wiens, 1989). Different kinds of disturbance create patterns at different scales. Even for a single disturbance event
within one landscape, considerable variation occurs
in the scale of resulting pattern. For both reasons,
a multi scale approach is essential. The scale at
which patterns are identified influences the assessment of these patterns, and patterns that are discerned at different scales may not be comparable.
For this reason, the grain and extent of any observation should be stated explicitly (Turner et aI.,
1989; Reed et aI., 1993).
23.3.2 Spatial Scale
Disturbances differ in a variety of spatial characteristics, including size (e.g., patch area, total disturbance area, area per time period), distribution
(location with respect to various physical or biotic
gradients), and landscape pattern (e.g., patch
shape, complexity, area-to-perimeter ratio, orientation and dispersion, degree of connectivity or contiguity, and relationship to the surrounding matrix).
These parameters affect the rate, degree, and nature of colonization, succession, and recovery from
disturbance in both terrestrial and marine systems
(Runkle, 1985; Sousa, 1985; White et aI., 1999).
Smith and Urban (1988) explored the effect of
different scales of observation in developing an individual tree-based forest succession model. They
concluded that at a broad scale a forest stand might
appear to be in a near-equilibrium state, with succession progressing toward a specific overall mean
species composition. An individual gap within the
stand, however, constantly undergoes structural
changes driven by the life-history traits of a few
dominant individuals. Stand dynamics, therefore,
tend to be driven by the demographics of individuals at a small spatial scale and by overall stand
biomass and composition at a broader scale (Smith
341
and Urban, 1988). This work illustrates the potential to overlook relevant mechanisms if we limit
ourselves to one spatial scale.
23.3.3 Temporal Scale
Temporal characteristics of disturbances include
frequency (the number of disturbance events per
unit time), rotation period (the time needed to disturb an area equal in size to the study area), return
interval (the time between disturbances), regularity (or periodicity), and contagion (rate and probability of spread) (White et al., 1999). Many factors
influence temporal variation, especially the periodicity of a particular disturbance type. These may
include endogenous feedback mechanisms (e.g., increased fuel loading and subsequent flammability
with long fire return intervals) or exogenous factors (e.g., long-term climatic cycles, the recurrence
of El NifLO). Deducing a characteristic or representative temporal periodicity may prove elusive, however, because disturbance regimes may not be temporally stable, especially in light of recent and
ongoing human alterations to many systems.
Ecological assessments often seek to separate recent human influences from naturally occurring dynamics. However, human effects often may extend
backward for longer time periods than we have generally addressed. In many locations in the United
States, there is growing evidence of widespread
manipulation of the fire regime by native Americans and European settlers, who used fire to drive
game and prepare spring forage. The degree to
which these anthropogenic activities may have affected the development and structure of the most
fire-prone ecosystems is unknown, but it is speculated that their influence might have been profound
in areas that were otherwise naturally fire protected
(Baker, 1992b; Ware et aI., 1993).
Different assessment objectives will require different temporal scales. Assessing the historical
range of variability for a forested site will involve
a time scale on the order of centuries in order to
encompass several generations of the dominant
trees (Morgan et aI., 1994). An assessment of
shorter-term successional trends may be based on
data collected during a single decade.
As with spatial scale, the amount of observable
variation depends on the resolution of the observations, in this case the time span of individual observations. At small temporal scales, an assessment
of dynamic pattern may seem straightforward, but
may also be shortsighted. Consider the example of
assessing dynamic patterns in riparian environments. Stream meandering results in continuous
the assessment process. Scale has two components:
grain, the degree of resolution or smallest recognizable element, and extent, the total area or time
span over which observations are made (Turner et
aI., 1989; Wiens, 1989). Varying either the grain
or extent of a sample alters our observations. Increasing grain size tends to reduce observed variance among samples because each sample is larger
and encompasses more variation internally; increasing extent tends to increase observed variance
because, in larger areas, more kinds of environmental conditions and events are observed (Wiens,
1989).
When conducting ecological assessments, it is
preferable to avoid a single arbitrary choice of scale
based purely on expediency (Wiens, 1989). Different kinds of disturbance create patterns at different scales. Even for a single disturbance event
within one landscape, considerable variation occurs
in the scale of resulting pattern. For both reasons,
a multi scale approach is essential. The scale at
which patterns are identified influences the assessment of these patterns, and patterns that are discerned at different scales may not be comparable.
For this reason, the grain and extent of any observation should be stated explicitly (Turner et aI.,
1989; Reed et aI., 1993).
23.3.2 Spatial Scale
Disturbances differ in a variety of spatial characteristics, including size (e.g., patch area, total disturbance area, area per time period), distribution
(location with respect to various physical or biotic
gradients), and landscape pattern (e.g., patch
shape, complexity, area-to-perimeter ratio, orientation and dispersion, degree of connectivity or contiguity, and relationship to the surrounding matrix).
These parameters affect the rate, degree, and nature of colonization, succession, and recovery from
disturbance in both terrestrial and marine systems
(Runkle, 1985; Sousa, 1985; White et aI., 1999).
Smith and Urban (1988) explored the effect of
different scales of observation in developing an individual tree-based forest succession model. They
concluded that at a broad scale a forest stand might
appear to be in a near-equilibrium state, with succession progressing toward a specific overall mean
species composition. An individual gap within the
stand, however, constantly undergoes structural
changes driven by the life-history traits of a few
dominant individuals. Stand dynamics, therefore,
tend to be driven by the demographics of individuals at a small spatial scale and by overall stand
biomass and composition at a broader scale (Smith
341
and Urban, 1988). This work illustrates the potential to overlook relevant mechanisms if we limit
ourselves to one spatial scale.
23.3.3 Temporal Scale
Temporal characteristics of disturbances include
frequency (the number of disturbance events per
unit time), rotation period (the time needed to disturb an area equal in size to the study area), return
interval (the time between disturbances), regularity (or periodicity), and contagion (rate and probability of spread) (White et al., 1999). Many factors
influence temporal variation, especially the periodicity of a particular disturbance type. These may
include endogenous feedback mechanisms (e.g., increased fuel loading and subsequent flammability
with long fire return intervals) or exogenous factors (e.g., long-term climatic cycles, the recurrence
of El NifLO). Deducing a characteristic or representative temporal periodicity may prove elusive, however, because disturbance regimes may not be temporally stable, especially in light of recent and
ongoing human alterations to many systems.
Ecological assessments often seek to separate recent human influences from naturally occurring dynamics. However, human effects often may extend
backward for longer time periods than we have generally addressed. In many locations in the United
States, there is growing evidence of widespread
manipulation of the fire regime by native Americans and European settlers, who used fire to drive
game and prepare spring forage. The degree to
which these anthropogenic activities may have affected the development and structure of the most
fire-prone ecosystems is unknown, but it is speculated that their influence might have been profound
in areas that were otherwise naturally fire protected
(Baker, 1992b; Ware et aI., 1993).
Different assessment objectives will require different temporal scales. Assessing the historical
range of variability for a forested site will involve
a time scale on the order of centuries in order to
encompass several generations of the dominant
trees (Morgan et aI., 1994). An assessment of
shorter-term successional trends may be based on
data collected during a single decade.
As with spatial scale, the amount of observable
variation depends on the resolution of the observations, in this case the time span of individual observations. At small temporal scales, an assessment
of dynamic pattern may seem straightforward, but
may also be shortsighted. Consider the example of
assessing dynamic patterns in riparian environments. Stream meandering results in continuous