24. Ecosystem Climate Manipulations
Developing biosphere models to couple with
GCMs is one way to learn about ecosystem consequences of global warming. Another is using observed statistical associations of organisms with
specific climate conditions, combined with GCM
predictions, to determine how current species'
ranges may be altered by global warming. Because
vegetation is a dominant factor in structuring ecosystems, most studies have focused on how climate
influences the distribution of vegetation (Davis and
Zabinski 1992; Woodward 1992). Present geographical ranges for species of interest are compared with climate zone maps to determine the
bounding values of temperature and moisture that
correspond to each species' range limit. Pollen and
fossil records reveal distributions under past climates. Global circulation model predictions of
global warming are then used to predict future
locations of suitable climates for each species,
yielding a quantitative estimate of the range displacement. Such studies indicate large potential displacements in geographical ranges (Davis and Zabinski 1992). Because many species will not be
able to adjust rapidly enough, and because migration may be impeded by human settlements, these
studies imply reductions in geographical range and
thus in popUlation size.
Although studies involving mechanistic models
or statistical associations are valuable, they have
many limitations. Model evaluation and improvement are hampered by a lack of appropriate observed data (Dickson 1995; Gates et al. 1995). Developing better models and thus better predictions
requires the grounding of laboratory and especially
field studies of ecosystem response to climate
change. It is to these that we now turn.
Methods of Ecosystem
Climate Manipulation
Response of ecosystems to future climate change
can be assessed with experiments that manipulate
environmental variables (e.g., temperature, CO 2 , or
light levels) in a manner that mimics climate
change. This can be done in a laboratory growth
chamber, or by a variety of field methods. To date,
the majority of manipUlation studies have involved
enhanced CO 2 concentration, have taken place in a
355
growth chamber, and have focused on a single species. Recently, long-term field manipulations of
whole-ecosystem response (or representative samples of ecosystems) to climate change have become
more common.
Laboratory Methods: Growth Chambers
Laboratory growth chambers manipulate environmental variables, including temperature, moisture,
and light (intensity, spectrum, and day length).
These studies have largely been limited to soil processes or to response of a single species (e.g., Hunt
et al. 1996). Ecosystem studies in small laboratory
chambers, referred to as "microcosms," can comprise intact field samples, but more typically are
communities "assembled" in the laboratory. Verhoef (1996) lists some considerations for the design
and implementation of microcosm experiments.
Growth chambers have many advantages. They enable precise control of environmental conditions,
and are easier to understand than complex, realworld systems. Chamber studies are easily replicable and repeatable and generate quick results,
making statistically significant findings more likely,
and the costs can be modest. Chamber studies give
mechanistic insights and so can be helpful in developing simulation models.
Unfortunately, as a result of their inherently
small spatial scale and their assemblage history,
over time laboratory chamber ecosystems tend to
diverge considerably from real-world ecosystems
(Dudzik et al. 1979). Some processes and organisms are too large, wide-ranging, or slow to include
in a growth chamber study, while others change so
quickly that they may reach unrealistic rates or
population densities in the course of a microcosm
experiment (Carpenter 1996).
Mesocosm experiments, one step up from microcosms, are conducted in larger, more versatile
growth chambers. Komer and coworkers have designed mesocosm experiments with tropical (Korner and Arnone 1992) and spruce ecosystems
(Hattenschwiler and Komer 1996). A notable
mesocosm facility is the Ecotron (Lawton 1996;
Lawton et al. 1993), a collection of integrated
growth chambers that can be used to study community and ecosystem response to climate change.
The Ecotron attempts to construct, maintain, and
manipulate entire model ecosystems and simulta-
Developing biosphere models to couple with
GCMs is one way to learn about ecosystem consequences of global warming. Another is using observed statistical associations of organisms with
specific climate conditions, combined with GCM
predictions, to determine how current species'
ranges may be altered by global warming. Because
vegetation is a dominant factor in structuring ecosystems, most studies have focused on how climate
influences the distribution of vegetation (Davis and
Zabinski 1992; Woodward 1992). Present geographical ranges for species of interest are compared with climate zone maps to determine the
bounding values of temperature and moisture that
correspond to each species' range limit. Pollen and
fossil records reveal distributions under past climates. Global circulation model predictions of
global warming are then used to predict future
locations of suitable climates for each species,
yielding a quantitative estimate of the range displacement. Such studies indicate large potential displacements in geographical ranges (Davis and Zabinski 1992). Because many species will not be
able to adjust rapidly enough, and because migration may be impeded by human settlements, these
studies imply reductions in geographical range and
thus in popUlation size.
Although studies involving mechanistic models
or statistical associations are valuable, they have
many limitations. Model evaluation and improvement are hampered by a lack of appropriate observed data (Dickson 1995; Gates et al. 1995). Developing better models and thus better predictions
requires the grounding of laboratory and especially
field studies of ecosystem response to climate
change. It is to these that we now turn.
Methods of Ecosystem
Climate Manipulation
Response of ecosystems to future climate change
can be assessed with experiments that manipulate
environmental variables (e.g., temperature, CO 2 , or
light levels) in a manner that mimics climate
change. This can be done in a laboratory growth
chamber, or by a variety of field methods. To date,
the majority of manipUlation studies have involved
enhanced CO 2 concentration, have taken place in a
355
growth chamber, and have focused on a single species. Recently, long-term field manipulations of
whole-ecosystem response (or representative samples of ecosystems) to climate change have become
more common.
Laboratory Methods: Growth Chambers
Laboratory growth chambers manipulate environmental variables, including temperature, moisture,
and light (intensity, spectrum, and day length).
These studies have largely been limited to soil processes or to response of a single species (e.g., Hunt
et al. 1996). Ecosystem studies in small laboratory
chambers, referred to as "microcosms," can comprise intact field samples, but more typically are
communities "assembled" in the laboratory. Verhoef (1996) lists some considerations for the design
and implementation of microcosm experiments.
Growth chambers have many advantages. They enable precise control of environmental conditions,
and are easier to understand than complex, realworld systems. Chamber studies are easily replicable and repeatable and generate quick results,
making statistically significant findings more likely,
and the costs can be modest. Chamber studies give
mechanistic insights and so can be helpful in developing simulation models.
Unfortunately, as a result of their inherently
small spatial scale and their assemblage history,
over time laboratory chamber ecosystems tend to
diverge considerably from real-world ecosystems
(Dudzik et al. 1979). Some processes and organisms are too large, wide-ranging, or slow to include
in a growth chamber study, while others change so
quickly that they may reach unrealistic rates or
population densities in the course of a microcosm
experiment (Carpenter 1996).
Mesocosm experiments, one step up from microcosms, are conducted in larger, more versatile
growth chambers. Komer and coworkers have designed mesocosm experiments with tropical (Korner and Arnone 1992) and spruce ecosystems
(Hattenschwiler and Komer 1996). A notable
mesocosm facility is the Ecotron (Lawton 1996;
Lawton et al. 1993), a collection of integrated
growth chambers that can be used to study community and ecosystem response to climate change.
The Ecotron attempts to construct, maintain, and
manipulate entire model ecosystems and simulta-