23
Large-Scale Water Manipulations
Paul J. Hanson
Introduction
Researchers predict that increasing levels of greenhouse gases in the atmosphere will cause a 1 to
3SC increase in average global temperatures and
alter regional levels of precipitation (Kattenberg et
al. 1996; Rind et al. 1990). Changes in global temperature and altered precipitation patterns may lead
to dramatic changes in ecosystem productivity, biogeochemical cycling, and the availability of water
resources (Kirschbaum and Fischlin 1996; Melillo
et al. 1990). The responses of ecosystems to decreased water availability or increased occurrence
of drought is considered a key issue in climate
change scenarios (Wigley et al. 1984), because
changes in the carbon sequestration potential of
vegetation at the global scale may alter the accumulation of carbon dioxide (C02) in the atmosphere, potentially moderating or accelerating climate changes (Ojima et al. 1991; Wigley and Jones
1985). Because the actual direction and magnitude
of expected changes in precipitation are highly uncertain, and consensus scenarios for regional climate change do not exist, manipulation experiments can play a significant role in clarifying the
potential impacts of a range of climate change scenarios on highly valued ecosystems.
Controlled experiments in greenhouses and
growth chambers have provided us with a large
database of information concerning the impacts of
moisture manipulations on the physiology and
growth of forest tree seedlings and saplings (Ellsworth and Reich 1992; Hinckley et al. 1978; Kleiner et al. 1992; Kolb et al. 1990; Pezeshki and
Chambers 1986). However, concerns remain as to
the appropriateness of the extrapolation of smallscale and short-term data to mature tree responses
in forest stands where existing data are limited
largely to short-term responses to water stress
(Cregg et al. 1989; Hinckley et al. 1978) or to a
limited number of trees sampled or species observed (Dougherty and Hinckley 1981; Epron et al.
1992; Ginter-Whitehouse et al. 1983). Similar concerns and arguments can be made for the extrapolation of data for other ecosystems. In response to
concerns over the validity of small-scale experimental response data for environmental assessments, several recent reviews (Graham et al. 1990;
Mooney 1991; Mooney et al. 1991; Woodward
1992) have called for large-scale manipulation
experiments as the appropriate means by which
to study the impacts of changing climates on
ecosystems.
This chapter outlines important methodological
issues involved in the design and operation oflargescale precipitation manipulation experiments. Previous precipitation or water manipulation studies
historically focused on scenarios of increasing or
decreasing precipitation patterns and the resulting
impacts on small plants or saplings grown in pots
or small field plots (Hinckley et al. 1978; Kozlowski 1982). The primary objective of these studies was to determine the impact of irrigation or
drought on plant growth or the harvest index of a
particular crop. Irrigation methods for application
to agricultural crops and forest plantations have
been discussed previously (Hagan et al. 1967;
Stewart and Nielsen 1990) and they will not be extensively reviewed here. Instead, this chapter focuses on precipitation manipulation methods that
341
Large-Scale Water Manipulations
Paul J. Hanson
Introduction
Researchers predict that increasing levels of greenhouse gases in the atmosphere will cause a 1 to
3SC increase in average global temperatures and
alter regional levels of precipitation (Kattenberg et
al. 1996; Rind et al. 1990). Changes in global temperature and altered precipitation patterns may lead
to dramatic changes in ecosystem productivity, biogeochemical cycling, and the availability of water
resources (Kirschbaum and Fischlin 1996; Melillo
et al. 1990). The responses of ecosystems to decreased water availability or increased occurrence
of drought is considered a key issue in climate
change scenarios (Wigley et al. 1984), because
changes in the carbon sequestration potential of
vegetation at the global scale may alter the accumulation of carbon dioxide (C02) in the atmosphere, potentially moderating or accelerating climate changes (Ojima et al. 1991; Wigley and Jones
1985). Because the actual direction and magnitude
of expected changes in precipitation are highly uncertain, and consensus scenarios for regional climate change do not exist, manipulation experiments can play a significant role in clarifying the
potential impacts of a range of climate change scenarios on highly valued ecosystems.
Controlled experiments in greenhouses and
growth chambers have provided us with a large
database of information concerning the impacts of
moisture manipulations on the physiology and
growth of forest tree seedlings and saplings (Ellsworth and Reich 1992; Hinckley et al. 1978; Kleiner et al. 1992; Kolb et al. 1990; Pezeshki and
Chambers 1986). However, concerns remain as to
the appropriateness of the extrapolation of smallscale and short-term data to mature tree responses
in forest stands where existing data are limited
largely to short-term responses to water stress
(Cregg et al. 1989; Hinckley et al. 1978) or to a
limited number of trees sampled or species observed (Dougherty and Hinckley 1981; Epron et al.
1992; Ginter-Whitehouse et al. 1983). Similar concerns and arguments can be made for the extrapolation of data for other ecosystems. In response to
concerns over the validity of small-scale experimental response data for environmental assessments, several recent reviews (Graham et al. 1990;
Mooney 1991; Mooney et al. 1991; Woodward
1992) have called for large-scale manipulation
experiments as the appropriate means by which
to study the impacts of changing climates on
ecosystems.
This chapter outlines important methodological
issues involved in the design and operation oflargescale precipitation manipulation experiments. Previous precipitation or water manipulation studies
historically focused on scenarios of increasing or
decreasing precipitation patterns and the resulting
impacts on small plants or saplings grown in pots
or small field plots (Hinckley et al. 1978; Kozlowski 1982). The primary objective of these studies was to determine the impact of irrigation or
drought on plant growth or the harvest index of a
particular crop. Irrigation methods for application
to agricultural crops and forest plantations have
been discussed previously (Hagan et al. 1967;
Stewart and Nielsen 1990) and they will not be extensively reviewed here. Instead, this chapter focuses on precipitation manipulation methods that
341