286
records derived from most other methods are also
subject to loss or degradation with time (Landres
et aI., 1999; Swetnam et al., 1999). Reconstructions
at local scales are obtained using plant macrofossil analysis and repeat photography. Results from
the other methods discussed can range from local
to regional in spatial scale. When historical data are
not available for a particular site, extrapolation of
data from another site may be unavoidable, but the
validity of such an extrapolation may depend on
the distance between the sites, the "distance decay"
problem (White and Walker, 1997; Swetnam et aI.,
1999).
All the methods described are limited in the patterns and processes they reconstruct, their spatial
and temporal resolutions, and their availability in
particular ecosystems. Limitations to many methods include the degree to which they describe vegetation structural characteristics, understory vegetation, spatial heterogeneity, and interactions
among disturbances. Other patterns and processes
potentially of interest in an assessment, such as animal populations or terrestrial biogeochemical
processes, are not characterized by any of the HRV
methods described. For some ecosystems, such as
the grasslands of the Great Plains and Columbia
Plateau, with sparse or absent historical records and
extensive human alteration, reconstructing HRV
may be difficult or impossible (Frissell and Bayles,
1996).
The limitations of HRV methods do not lessen
the value of historical information for ecological
assessments, but emphasize the need to consider as
many sources of information for determining HRV
as possible (Swetnam et aI., 1999). As a result of
methodological limitations, combinations of methods provide the most complete information about
HRV and can broaden the suite of entities and spatial and temporal extents considered in an assessment (Table 19.1). Multiple HRV sources provide
increased confidence in reconstructions when different lines of evidence converge (Swetnam et aI.,
1999). Many of the studies described in this chapter contain more than one HRV data source, such
as the combinations of pollen with plant macrofossil analysis and charcoal with dendroecological
analysis. Examples of other combinations of methods include one study in which information was derived from repeat photography, vegetation plots,
and analyses of tree rings and pollen (Veblen and
Markgraf, 1988) and another combining repeat
photography, fire-scar analysis, and simulation
modeling (Covington and Moore, 1994a). Shinneman and Baker (1997) evaluated a number of
sources of information, such as dendroecological
Methods for Determining Historical Range of Variability
studies, historical photographs, maps, and written
accounts, to assess HRV in ponderosa pine forests
in the Black Hills of South Dakota and Wyoming.
By combining multiple sources, they were able to
suggest that presettlement fire disturbance regimes
differed by location in the Black Hills. Kaufmann
et aI. (1998) reconstructed presettlement biotic and
disturbance variability as part of an ecological
assessment of the Sacramento Mountains, New
Mexico, using General Land Office survey data,
packrat and porcupine midden records, tree-ring
analysis, and repeat photography, along with old
survey records, historic accounts, and archeological records. Cole and Taylor's (1995) examination
of change in dune prairie-oak savanna contained
extensive use of multiple quantitative sources of information, including vegetation plots, aerial photography, fire-scar and multi stemmed tree analysis,
land survey data, and pollen and charcoal records.
In this case, the use of a range of methods covering a number of temporal scales enabled current
rapid rates of vegetation change to be placed in the
context of 4000 years of vegetation and fire history. Information that is integrated across several
scales is most likely to produce a comprehensive
description of HRV (Cole and Taylor, 1995).
Determination of HRV in critical ecosystem
components can provide important information for
the planning process following an assessment. Cissel et aI. (1998) explicitly incorporated HRV into
a landscape plan developed to guide management
activities in a 7600-ha planning area in western
Oregon, in which landscape and watershed management objectives were based on the inferred
HRV in landscape conditions and disturbance
processes. The determination of historical disturbance regimes included (1) use of dendroecological methods and plot-level data to compile a fire
history and construct fire-event maps, (2) interpretation of landslide and debris-flow occurrences
from aerial photographs, maps of existing condition, and field surveys, and (3) assessment of riparian vegetation dynamics and disturbance history
by means of a 40-year time series of aerial photographs. This case study illustrates how information on landscape conditions and disturbance
regimes can be integrated for developing objectives
and management prescriptions based on HRV to
guide management activities.
The objectives of an ecological assessment
should be the primary guide in selecting methods
for obtaining HRV, including determining the appropriate spatial and temporal scales for conducting analyses. However, the characteristics of a particular area included in an assessment will
records derived from most other methods are also
subject to loss or degradation with time (Landres
et aI., 1999; Swetnam et al., 1999). Reconstructions
at local scales are obtained using plant macrofossil analysis and repeat photography. Results from
the other methods discussed can range from local
to regional in spatial scale. When historical data are
not available for a particular site, extrapolation of
data from another site may be unavoidable, but the
validity of such an extrapolation may depend on
the distance between the sites, the "distance decay"
problem (White and Walker, 1997; Swetnam et aI.,
1999).
All the methods described are limited in the patterns and processes they reconstruct, their spatial
and temporal resolutions, and their availability in
particular ecosystems. Limitations to many methods include the degree to which they describe vegetation structural characteristics, understory vegetation, spatial heterogeneity, and interactions
among disturbances. Other patterns and processes
potentially of interest in an assessment, such as animal populations or terrestrial biogeochemical
processes, are not characterized by any of the HRV
methods described. For some ecosystems, such as
the grasslands of the Great Plains and Columbia
Plateau, with sparse or absent historical records and
extensive human alteration, reconstructing HRV
may be difficult or impossible (Frissell and Bayles,
1996).
The limitations of HRV methods do not lessen
the value of historical information for ecological
assessments, but emphasize the need to consider as
many sources of information for determining HRV
as possible (Swetnam et aI., 1999). As a result of
methodological limitations, combinations of methods provide the most complete information about
HRV and can broaden the suite of entities and spatial and temporal extents considered in an assessment (Table 19.1). Multiple HRV sources provide
increased confidence in reconstructions when different lines of evidence converge (Swetnam et aI.,
1999). Many of the studies described in this chapter contain more than one HRV data source, such
as the combinations of pollen with plant macrofossil analysis and charcoal with dendroecological
analysis. Examples of other combinations of methods include one study in which information was derived from repeat photography, vegetation plots,
and analyses of tree rings and pollen (Veblen and
Markgraf, 1988) and another combining repeat
photography, fire-scar analysis, and simulation
modeling (Covington and Moore, 1994a). Shinneman and Baker (1997) evaluated a number of
sources of information, such as dendroecological
Methods for Determining Historical Range of Variability
studies, historical photographs, maps, and written
accounts, to assess HRV in ponderosa pine forests
in the Black Hills of South Dakota and Wyoming.
By combining multiple sources, they were able to
suggest that presettlement fire disturbance regimes
differed by location in the Black Hills. Kaufmann
et aI. (1998) reconstructed presettlement biotic and
disturbance variability as part of an ecological
assessment of the Sacramento Mountains, New
Mexico, using General Land Office survey data,
packrat and porcupine midden records, tree-ring
analysis, and repeat photography, along with old
survey records, historic accounts, and archeological records. Cole and Taylor's (1995) examination
of change in dune prairie-oak savanna contained
extensive use of multiple quantitative sources of information, including vegetation plots, aerial photography, fire-scar and multi stemmed tree analysis,
land survey data, and pollen and charcoal records.
In this case, the use of a range of methods covering a number of temporal scales enabled current
rapid rates of vegetation change to be placed in the
context of 4000 years of vegetation and fire history. Information that is integrated across several
scales is most likely to produce a comprehensive
description of HRV (Cole and Taylor, 1995).
Determination of HRV in critical ecosystem
components can provide important information for
the planning process following an assessment. Cissel et aI. (1998) explicitly incorporated HRV into
a landscape plan developed to guide management
activities in a 7600-ha planning area in western
Oregon, in which landscape and watershed management objectives were based on the inferred
HRV in landscape conditions and disturbance
processes. The determination of historical disturbance regimes included (1) use of dendroecological methods and plot-level data to compile a fire
history and construct fire-event maps, (2) interpretation of landslide and debris-flow occurrences
from aerial photographs, maps of existing condition, and field surveys, and (3) assessment of riparian vegetation dynamics and disturbance history
by means of a 40-year time series of aerial photographs. This case study illustrates how information on landscape conditions and disturbance
regimes can be integrated for developing objectives
and management prescriptions based on HRV to
guide management activities.
The objectives of an ecological assessment
should be the primary guide in selecting methods
for obtaining HRV, including determining the appropriate spatial and temporal scales for conducting analyses. However, the characteristics of a particular area included in an assessment will