362
in Switzerland, choosing to achieve a fixed 2SC
temperature differential between adjacent heated
and control plots by continually adjusting the IR
heater output. In their experiment, called Free Air
Temperature Increase (FATI-in analogy to the
FACE experiments of enhanced CO 2 -although
this a misnomer since IR heaters do not directly
warm the intervening air), they used large heaters,
one per plot, mounted vertically at the plot edge
and angled to irradiate downward over the plot.
Their control plots had dummy heaters (no output)
to cancel out shading effects. They screened out
radiation below 800 nm to avoid unwanted photomorphogenic effects. An alternative to filtering is
to select heaters designed to emit radiation essentially only in the IR portion of the spectrum. Nijs
and coworkers were interested in the effect of the
experimental design on canopy and leaf temperatures and thus only ran their study for a period of
weeks. In another warming experiment, Bridgham
et al. (1995) used overhead heaters to investigate
effects of global warming on trace gas fluxes in bog
ecosystems in Minnesota, in the United States.
Rather than apply the manipulation in situ, they established large mesocosms (2.2 m 2 and 0.6 m in
depth) with intact vegetation communities, and
placed them in containers for manipulation by the
heaters. Though not technically an in situ field
study, this comes close to an actual ecosystem manipulation, and has the advantages of convenience
and easier access.
Overhead heaters can be applied in many ecosystems, and are relatively easy to install in tundra,
desert, meadow, and agricultural systems. Since IR
heaters do not directly heat the air, this technique
cannot simulate the convective heating effect of
global warming. It would be impractical to use suspended heaters for whole-ecosystem warming in
forests, because of the need to suspend them over
the canopy. Moreover, a dense canopy would virtually eliminate the possibility of soil warming. Indeed, no method of controlled heating of whole
dense forests has been demonstrated, but forest soil
heating can be achieved either by heating wires or
pipes at the soil surface, set to provide a constant
power output, or by IR lamps suspended under the
canopy.
Field Methods: Other Climate Variables
Global warming will alter many climatic variables
besides temperature. Changes in amount and duration of winter snowpack, amount and distribution
Karin P. Shen and John Harte
of rainfall, cloud cover, and soil moisture are some
effects relevant to ecosystem processes.
Changes in snow depth and the duration of the
snowmelt period can limit the distribution of plant
species and may induce a variety of growth responses, including changes in morphology, flowering, and seed-set, as well as taxonomic shifts toward more opportunistic species. Snow fences
provide a simple and inexpensive means of locally
altering snowpack. Snow shoveling has also been
employed to alter snow depth (Dunne 1996). Snow
fences have been used in long-term experiments at
the LTER site at Niwot Ridge, Colorado, U.S.A.
(Brooks et al. 1996; Walker 1991). Brooks et al.
found that the installation of a 2.6 X 60 m snow
fence at this site resulted in a snowpack significantly deeper and longer lasting than before construction of the fence. Sturges (1989) examined the
effect of induced snow accumulation on mountain
big sagebrush by installing a snow fence 3.8 m tall.
Chapin and Shaver (1996) used snow fences to determine the effects of shortened growing season in
a tussock tundra ecosystem. They erected a 40-m
long, 1.5-m tall snow fence perpendicular to the
prevailing winter wind and parallel to the slope of
their field site, creating late-lying snowdrifts and
shortening the growing season by 2 weeks. Galen
and Stanton (1995) also used snow fences to alter
growing season length. Tabler (1980) discusses geometry and density of drifts formed by snow
fences. Although snow fences exert a dramatic local impact on snow accumulation, their effects cannot be precisely controlled. As we have discussed,
greenhouses and buried resistance wires also have
major drawbacks for winter application. Infrared
lamps provide a useful way to simulate effects of
global warming on winter snow cover, but the melt
pits they produce may accumulate wind-blown
snow and thus unrealistically increase total snowmelt water available to the heated plots.
Global warming will also affect rainfall patterns.
Although soil moisture is affected by direct warming, the combined effect on soil moisture of both
an altered thermal regime and altered precipitation
rates has not been investigated at field scale. Tarpaulins have been used to intercept rainfall on experimental plots in a savanna ecosystem in Arizona,
U.S.A., to investigate the impact of global change
on seedling recruitment (MacPherson 1995) and on
montane meadow plots along an elevational gra-
in Switzerland, choosing to achieve a fixed 2SC
temperature differential between adjacent heated
and control plots by continually adjusting the IR
heater output. In their experiment, called Free Air
Temperature Increase (FATI-in analogy to the
FACE experiments of enhanced CO 2 -although
this a misnomer since IR heaters do not directly
warm the intervening air), they used large heaters,
one per plot, mounted vertically at the plot edge
and angled to irradiate downward over the plot.
Their control plots had dummy heaters (no output)
to cancel out shading effects. They screened out
radiation below 800 nm to avoid unwanted photomorphogenic effects. An alternative to filtering is
to select heaters designed to emit radiation essentially only in the IR portion of the spectrum. Nijs
and coworkers were interested in the effect of the
experimental design on canopy and leaf temperatures and thus only ran their study for a period of
weeks. In another warming experiment, Bridgham
et al. (1995) used overhead heaters to investigate
effects of global warming on trace gas fluxes in bog
ecosystems in Minnesota, in the United States.
Rather than apply the manipulation in situ, they established large mesocosms (2.2 m 2 and 0.6 m in
depth) with intact vegetation communities, and
placed them in containers for manipulation by the
heaters. Though not technically an in situ field
study, this comes close to an actual ecosystem manipulation, and has the advantages of convenience
and easier access.
Overhead heaters can be applied in many ecosystems, and are relatively easy to install in tundra,
desert, meadow, and agricultural systems. Since IR
heaters do not directly heat the air, this technique
cannot simulate the convective heating effect of
global warming. It would be impractical to use suspended heaters for whole-ecosystem warming in
forests, because of the need to suspend them over
the canopy. Moreover, a dense canopy would virtually eliminate the possibility of soil warming. Indeed, no method of controlled heating of whole
dense forests has been demonstrated, but forest soil
heating can be achieved either by heating wires or
pipes at the soil surface, set to provide a constant
power output, or by IR lamps suspended under the
canopy.
Field Methods: Other Climate Variables
Global warming will alter many climatic variables
besides temperature. Changes in amount and duration of winter snowpack, amount and distribution
Karin P. Shen and John Harte
of rainfall, cloud cover, and soil moisture are some
effects relevant to ecosystem processes.
Changes in snow depth and the duration of the
snowmelt period can limit the distribution of plant
species and may induce a variety of growth responses, including changes in morphology, flowering, and seed-set, as well as taxonomic shifts toward more opportunistic species. Snow fences
provide a simple and inexpensive means of locally
altering snowpack. Snow shoveling has also been
employed to alter snow depth (Dunne 1996). Snow
fences have been used in long-term experiments at
the LTER site at Niwot Ridge, Colorado, U.S.A.
(Brooks et al. 1996; Walker 1991). Brooks et al.
found that the installation of a 2.6 X 60 m snow
fence at this site resulted in a snowpack significantly deeper and longer lasting than before construction of the fence. Sturges (1989) examined the
effect of induced snow accumulation on mountain
big sagebrush by installing a snow fence 3.8 m tall.
Chapin and Shaver (1996) used snow fences to determine the effects of shortened growing season in
a tussock tundra ecosystem. They erected a 40-m
long, 1.5-m tall snow fence perpendicular to the
prevailing winter wind and parallel to the slope of
their field site, creating late-lying snowdrifts and
shortening the growing season by 2 weeks. Galen
and Stanton (1995) also used snow fences to alter
growing season length. Tabler (1980) discusses geometry and density of drifts formed by snow
fences. Although snow fences exert a dramatic local impact on snow accumulation, their effects cannot be precisely controlled. As we have discussed,
greenhouses and buried resistance wires also have
major drawbacks for winter application. Infrared
lamps provide a useful way to simulate effects of
global warming on winter snow cover, but the melt
pits they produce may accumulate wind-blown
snow and thus unrealistically increase total snowmelt water available to the heated plots.
Global warming will also affect rainfall patterns.
Although soil moisture is affected by direct warming, the combined effect on soil moisture of both
an altered thermal regime and altered precipitation
rates has not been investigated at field scale. Tarpaulins have been used to intercept rainfall on experimental plots in a savanna ecosystem in Arizona,
U.S.A., to investigate the impact of global change
on seedling recruitment (MacPherson 1995) and on
montane meadow plots along an elevational gra-