364
(e.g., Adamse and Britz 1996). Ideally, these lamps
should automatically vary the output as the background sunlight level changes (e.g., Barnes et al.
1995; Fiscus et al. 1996; Olszyk et al. 1996), although some studies have kept a constant output
for a set amount of daylight hours (e.g., Olszyk et
al. 1996). Only a few field experiments of UV-B
radiation enhancement have considered the plant
community or ecosystem level, and none has been
long-term. Barnes et al. (1995) examined the effects
of increased UV-B radiation resulting from a 20%
decrease in stratospheric ozone on canopy photosynthesis and light competition in wheat and wild
oat in Utah, U.S.A., using a modulated fluorescent
lamp system, suspended 40 cm above the experimental plants. Fiscus et al. (1996) used fluorescent
lamps suspended in open-topped field chambers to
study response of soybean to three levels of UV-B
radiation corresponding to various changes in the
total column ozone thickness.
The use of UV lamps in the field is costly, and
plagued with problems. The lamps and filters do
not provide ideal simulations of solar UV-B radiation (for a detailed analysis, see Caldwell et al.
1986). There are problems of optical degradation,
aging, and temperature sensitivity, and also geometrical constraints which prevent an accurate
UV-B radiation enhancement simulation. An added
difficulty is that since relatively small amounts of
energy are involved in these manipulations, and the
UV-B radiation phenomena are so wavelengthspecific, small discrepancies can result in substantial inaccuracies (Caldwell and Flint 1994). Although there is a need for ecosystem-level field
studies of UV-B radiation impacts of stratospheric
ozone depletion, the use of UV-B enhancing lamps
is problematic for long-term field studies. We encourage the reader to consult the work of Caldwell
and coworkers for more information before undertaking field manipulations of this kind.
General Recommendations
As a prerequisite to undertaking a climate manipulation study, ecologists should familiarize themselves with the science of global warming and the
latest climate model predictions for changes in temperature, precipitation, and other variables. The
best place to begin is the most recent report by the
Karin P. Shen and John Harte
Intergovernmental Panel on Climate Change (IPCC
1995)
There is a need for large-scale, long-term field
studies of ecosystem response to global and regional change, including warming, moisture
changes, elevated CO 2 , and increased UV-B radiation. Large-scale, whole-ecosystem experiments
are the best means of examining the individual and
net effects of the different feedbacks within the system. In addition to their own worth, field manipulations provide context for the many smaller-scale
(time and space), single-species studies that exist.
Experiments may need to run for several years before trends become discernible (Coulson et al.
1996). Whenever possible, long-term field studies
should be year-round, in order to capture the entire
annual cycle of response.
There is no single ecosystem manipulation
method that will be appropriate in all situations. For
whole-ecosystem warming manipulations in relatively low canopy systems, the use of overhead IR
heaters to increase radiant energy input to ecosystems is probably the best option. As discussed
above, it is preferable for many reasons to maintain
a constant power output rather than a thermostated
temperature rise. Climate manipUlations should incorporate monitoring of microclimate throughout
the experiment, on an appropriately fine spatial and
temporal scale, and researchers should examine
variance and extremes in addition to mean values
so as to gain insight to diel and seasonal variations
in microclimate response to the treatment.
The task of extrapolating experimental results on
ecosystem responses to climate change on the scale
of plots to that of landscapes or continents is now
widely accepted as one of the central challenges
facing global change science (e.g., Anderson 1992;
Ehleringer and Field 1993; Lubchenco 1991; Rosswall et al. 1988). Predicting community and ecosystem responses to environmental changes from
study of individual physiological processes requires
consideration of the time scales over which physiological processes exert their effects (Chapin and
Shaver 1996). Problems of small scale size in field
studies may sometimes be circumvented by combining information from laboratory, field, and modeling studies (e.g., Harte et al. 1992). Comparing
results from ecosystem field manipulations and
from correlation analysis along elevational gradients can generate insights into how transient re-
(e.g., Adamse and Britz 1996). Ideally, these lamps
should automatically vary the output as the background sunlight level changes (e.g., Barnes et al.
1995; Fiscus et al. 1996; Olszyk et al. 1996), although some studies have kept a constant output
for a set amount of daylight hours (e.g., Olszyk et
al. 1996). Only a few field experiments of UV-B
radiation enhancement have considered the plant
community or ecosystem level, and none has been
long-term. Barnes et al. (1995) examined the effects
of increased UV-B radiation resulting from a 20%
decrease in stratospheric ozone on canopy photosynthesis and light competition in wheat and wild
oat in Utah, U.S.A., using a modulated fluorescent
lamp system, suspended 40 cm above the experimental plants. Fiscus et al. (1996) used fluorescent
lamps suspended in open-topped field chambers to
study response of soybean to three levels of UV-B
radiation corresponding to various changes in the
total column ozone thickness.
The use of UV lamps in the field is costly, and
plagued with problems. The lamps and filters do
not provide ideal simulations of solar UV-B radiation (for a detailed analysis, see Caldwell et al.
1986). There are problems of optical degradation,
aging, and temperature sensitivity, and also geometrical constraints which prevent an accurate
UV-B radiation enhancement simulation. An added
difficulty is that since relatively small amounts of
energy are involved in these manipulations, and the
UV-B radiation phenomena are so wavelengthspecific, small discrepancies can result in substantial inaccuracies (Caldwell and Flint 1994). Although there is a need for ecosystem-level field
studies of UV-B radiation impacts of stratospheric
ozone depletion, the use of UV-B enhancing lamps
is problematic for long-term field studies. We encourage the reader to consult the work of Caldwell
and coworkers for more information before undertaking field manipulations of this kind.
General Recommendations
As a prerequisite to undertaking a climate manipulation study, ecologists should familiarize themselves with the science of global warming and the
latest climate model predictions for changes in temperature, precipitation, and other variables. The
best place to begin is the most recent report by the
Karin P. Shen and John Harte
Intergovernmental Panel on Climate Change (IPCC
1995)
There is a need for large-scale, long-term field
studies of ecosystem response to global and regional change, including warming, moisture
changes, elevated CO 2 , and increased UV-B radiation. Large-scale, whole-ecosystem experiments
are the best means of examining the individual and
net effects of the different feedbacks within the system. In addition to their own worth, field manipulations provide context for the many smaller-scale
(time and space), single-species studies that exist.
Experiments may need to run for several years before trends become discernible (Coulson et al.
1996). Whenever possible, long-term field studies
should be year-round, in order to capture the entire
annual cycle of response.
There is no single ecosystem manipulation
method that will be appropriate in all situations. For
whole-ecosystem warming manipulations in relatively low canopy systems, the use of overhead IR
heaters to increase radiant energy input to ecosystems is probably the best option. As discussed
above, it is preferable for many reasons to maintain
a constant power output rather than a thermostated
temperature rise. Climate manipUlations should incorporate monitoring of microclimate throughout
the experiment, on an appropriately fine spatial and
temporal scale, and researchers should examine
variance and extremes in addition to mean values
so as to gain insight to diel and seasonal variations
in microclimate response to the treatment.
The task of extrapolating experimental results on
ecosystem responses to climate change on the scale
of plots to that of landscapes or continents is now
widely accepted as one of the central challenges
facing global change science (e.g., Anderson 1992;
Ehleringer and Field 1993; Lubchenco 1991; Rosswall et al. 1988). Predicting community and ecosystem responses to environmental changes from
study of individual physiological processes requires
consideration of the time scales over which physiological processes exert their effects (Chapin and
Shaver 1996). Problems of small scale size in field
studies may sometimes be circumvented by combining information from laboratory, field, and modeling studies (e.g., Harte et al. 1992). Comparing
results from ecosystem field manipulations and
from correlation analysis along elevational gradients can generate insights into how transient re-