Assessing the Response of Terrestrial Ecosystems to Potential Changes in Precipitation

Abstract

Changes in Earth's surface temperatures caused by anthropogenic emissions of greenhouse gases are expected to affect global
and regional precipitation regimes. Interactions between changing precipitation regimes and other aspects of global change
are likely to affect natural and managed terrestrial ecosystems as well as human society. Although much recent research has
focused on assessing the responses of terrestrial ecosystems to rising carbon dioxide or temperature, relatively little research
has focused on understanding how ecosystems respond to changes in precipitation regimes. Here we review predicted changes
in global and regional precipitation regimes, outline the consequences of precipitation change for natural ecosystems and
human activities, and discuss approaches to improving understanding of ecosystem responses to changing precipitation. Further,
we introduce the Precipitation and Ecosystem Change Research Network (PrecipNet), a new interdisciplinary research network
assembled to encourage and foster communication and collaboration across research groups with common interests in the impacts
of global change on precipitation regimes, ecosystem structure and function, and the human enterprise.

Full text

October 2003 / Vol. 53 No. 10 • BioScience 941
Articles
The responses of terrestrial ecosystems to global
environmental change, and the resulting impacts on the
natural resources on which humans depend, are topics of great
societal concern and current scientific interest (Vitousek
1994). Anthropogenic emissions of greenhouse gases are ex-
pected to raise the mean temperatures of Earth’s surface by
1.4°C to 5.8°C during this century (Houghton et al. 2001).
Such warming is likely to alter patterns of global air circula-
tion and hydrologic cycling that will change global and
regional precipitation regimes (Houghton et al. 2001).
Corresponding changes in air and soil temperatures, soil
water and nutrient contents, and concentrations of atmos-
pheric carbon dioxide ([CO2]) are likely to alter the func-
tioning of natural and managed ecosystems in terrestrial en-
vironments. Because these changes will co-occur with ongoing
changes in global land use and land cover that have already
affected biodiversity and natural resources,impacts on human
societies are expected (Vitousek 1994).
Considerable research has been directed at understanding
the effects of increased temperature and [CO2] on the
Jake F. Weltzin (e-mail: jweltzin@utk.edu) is an assistant professor in the Department of Ecology and Evolutionary Biology, University of Tennessee, Knoxville, TN 37919.
Michael E. Loik is an assistant professor and Brent M. Haddad is an associate professor in the Department of Environmental Studies,University of California, Santa
Cruz, CA 95064. Susanne Schwinning is a postdoctoral associate and Guanghui Lin is an associate research scientist at the Biosphere 2 Center, Columbia University,
Oracle, AZ 85623. David G. Williams is an associate professor in the Departments of Renewable Resources and Botany, University of Wyoming, Laramie, WY 82071.
Philip A. Fay is a plant ecologist at the Natural Resources Research Institute, Duluth, MN 55811. John Harte is a professor in the Energy and Resources Group, Uni-
versity of California, Berkeley, CA 94720. Travis E. Huxman is an assistant professor in the Department of Ecology and Evolutionary Biology, University of Arizona,
Tucson, AZ 85721.Alan K. Knapp is a professor in the Division of Biology, Kansas State University, Manhattan, KS 66506. William T. Pockman is an assistant pro-
fessor in the Department of Biology, University of New Mexico,Albuquerque, NM 87131. M. Rebecca Shaw is a researcher in the Department of Global Ecology, Carnegie
Institution of Washington, 260 Panama Street, Stanford, CA 94305. Eric E. Small is an assistant professor in the Department of Geological Sciences, University of
Colorado, Boulder, CO 80309. Melinda D. Smith is a postdoctoral fellow at the National Center for Ecological Analysis and Synthesis,Santa Barbara, CA 93101.
Stanley D. Smith is a professor in the Department of Biological Sciences, University of Nevada, Las Vegas, NV 89154. David T. Tissue is an associate professor
and John C. Zak is a professor in the Department of Biological Sciences, Texas Tech University, Lubbock, TX 79409. © 2003 American Institute of
Biological Sciences.
Assessing the Response
of Terrestrial Ecosystems
to Potential Changes in
Precipitation
JAKE F. WELTZIN, MICHAEL E. LOIK, SUSANNE SCHWINNING, DAVID G. WILLIAMS, PHILIP A. FAY, BRENT M.
HADDAD, JOHN HARTE, TRAVIS E. HUXMAN, ALAN K. KNAPP,GUANGHUI LIN, WILLIAM T. POCKMAN, M. REBECCA
SHAW, ERIC E. SMALL, MELINDA D. SMITH, STANLEY D. SMITH, DAVID T. TISSUE, AND JOHN C. ZAK
Changes in Earth’s surface temperatures caused by anthropogenic emissions of greenhouse gases are expected to affect global and regional precipi-
tation regimes. Interactions between changing precipitation regimes and other aspects of global change are likely to affect natural and managed
terrestrial ecosystems as well as human society. Although much recent research has focused on assessing the responses of terrestrial ecosystems to
rising carbon dioxide or temperature, relatively little research has focused on understanding how ecosystems respond to changes in precipitation
regimes. Here we review predicted changes in global and regional precipitation regimes, outline the consequences of precipitation change for
natural ecosystems and human activities, and discuss approaches to improving understanding of ecosystem responses to changing precipitation.
Further, we introduce the Precipitation and Ecosystem Change Research Network (PrecipNet), a new interdisciplinary research network assembled
to encourage and foster communication and collaboration across research groups with common interests in the impacts of global change on precipi-
tation regimes, ecosystem structure and function, and the human enterprise.
Keywords: global change, community, ecosystem, precipitation, soil moisture

structural and physiological dynamics of terrestrial ecosystems
(e.g., Koch and Mooney 1996, Shaver et al. 2000).Although
there is a long history of investigation of linkages between pre-
cipitation and terrestrial ecosystems (Noy-Meir 1973, Leith
1975), little research has focused on how anticipated changes
in precipitation might affect terrestrial ecosystems. We
suggest that shifts in precipitation regimes may have an even
greater impact on ecosystem dynamics than the singular or
combined effects of rising [CO2] and temperature, especially
in arid and semiarid environments. For example, precipita-
tion substantially influenced plant and ecosystem response
to elevated [CO2] in an arid ecosystem (Smith et al.
2000). Moreover, environmental degradation in drought-
susceptible regions negatively affects nearly one billion
people occupying about 30% of the world’s land surface
(FAO 1993). Thus, a research focus on these regions and on
the effects of changing precipitation patterns would yield
information necessary to mitigate potentially negative impacts
of climate change on human well-being.
This article addresses the following basic questions
concerning precipitation change research:
•How will global and regional precipitation patterns
change in the near future?
•How does precipitation influence the dynamics of
natural ecosystems?
•Which ecosystems and ecosystem processes are sensitive
to changes in precipitation?
•How will changes in precipitation alter human–
ecosystem interactions?
•What approaches are available to study the effects of
precipitation change?
We will highlight the need for multidisciplinary approaches
and the challenges in interpreting limited data sets within the
context of global change. These factors have motivated the for-
mation of an interdisciplinary research network, PrecipNet
(Precipitation and Ecosystem Change Research Network),
to promote additional precipitation studies, strengthen col-
laborative research, and facilitate exchange of information
about the impacts of precipitation change on terrestrial
ecosystems and on the natural resources that support human
activities.
How will global and regional precipitation
patterns change in the near future?
General circulation models (GCMs) are used to describe the
complex dynamics of mass and energy exchange, momentum,
and hydrologic cycling within Earth’s surface–atmosphere
system. The most widely accepted models predict increases
in mean global precipitation of up to 7% during this century,
depending on the model used and on how the exchange of
greenhouse trace gases from terrestrial and oceanic sources
is defined (Houghton et al. 2001). One common prediction
from these models, regardless of the model used or the
scenario of trace gas emissions employed, is that the amount
of precipitation in the tropics and at midlatitudes and high
latitudes will increase over this century,while precipitation at
subtropical latitudes will decrease. Moreover, the intensity of
precipitation events and the frequency of extreme events,
which have already increased across the globe, are predicted
to increase further (Easterling et al. 2000).
However, scenarios for many specific geographic regions
remain ambiguous, with unresolved discrepancies between the
outputs of different models. For example, a Canadian Cen-
tre for Climate Modelling and Analysis model (CGCM1)
predicts reductions in summer and winter precipitation in the
Southeast and Great Plains regions of the United States by
2095, whereas a model developed by the Hadley Centre for
Climate Prediction and Research (HadCM2) predicts in-
creased precipitation throughout most of the United States,
and particularly the Southwest, over the same time period (fig-
ure 1; NAST 2000). Both models predict that tropospheric
warming will increase evaporation rates and thus increase the
severity of drought despite potential increases in precipitation
in some regions (NAST 2000).
One of the major challenges in predicting precipitation pat-
terns at scales that are meaningful for ecosystem function and
land management is the representation of effects imposed by
surface topography and other landscape features. Most recent
GCMs operate with a spatial resolution of about 2.5° (lati-
tude/longitude) square or coarser.At this scale, varied topog-
raphy and other landscape features (e.g., coastline,lake, and
orographic effects) can modify local precipitation patterns.
Thus, the uncertainty associated with predictions for topo-
graphically complex regions such as the western United States
is relatively high. Regional climate models (e.g., Giorgi et al.
1998) can bring resolution to about 45 kilometers square
(Snyder et al. 2002).However, the improved resolution must
be weighed against uncertainty in long-term predictions.
Moreover, interactions between El Niño–Southern Oscillation
(ENSO) and the Pacific Decadal Oscillation, which operate
at different spatial and temporal scales, may affect regional pre-
cipitation in complex and as yet unpredictable ways (Collier
and Webb 2002). Analogous to the need for enhanced spatial
resolution in climate models, there is also a need for greater
temporal resolution. Most models produce output on seasonal
or monthly time steps, but the organisms that dominate
ecosystem responses to climate change can be sensitive to pre-
cipitation patterns on shorter scales, such as the number of
storms per rainy season, the relationship between precipita-
tion timing and magnitude (e.g., fewer large storms versus
more frequent small storms), or variation in the duration of
the rainy or dry season.
While current climate models seem unable to make reliable
predictions about the magnitude or even the direction of
precipitation change on smaller, biologically meaningful
scales, they do indicate that many regions of the world will ex-
perience alterations in precipitation regimes over the next 100
years. The scientific community should consider the conse-
quences of a range of possible climate scenarios, and land and
942 BioScience • October 2003 / Vol. 53 No. 10
Articles

water managers should develop strategies to mitigate the
most negative impacts of likely climate scenarios on natural
ecosystems and human society.
How does precipitation influence the
dynamics of natural ecosystems?
Soil moisture is the direct link between precipitation and
ecological systems. Therefore,understanding the effects of pre-
cipitation on soil moisture has been a central goal for hy-
drologists and soil physicists for many years (Noy-Meir 1973)
and remains an active field of research (e.g., Eagleson 2002).
The basic phenomena associated with precipitation events—
interception, infiltration,and runoff—are relatively well un-
derstood; the main difficulty lies in describing rates of soil
moisture change between precipitation events (McAuliffe
2003). These rates are driven chiefly by evaporation from
soils, transpiration by plants, horizontal and vertical soil
water transport, and hydraulic redistribution of soil water,all
of which depend in complex ways on vegetation and soil
characteristics and on the timing and size of precipitation in-
puts.
In arid and semiarid ecosystems, there is a good correla-
tion between event size and infiltration depth: Water from
larger rainfall events infiltrates more deeply (Sala et al. 1981),
but infiltration, storage, and use depend on the season and on
patterns of organismal activity. In summer, evaporation and
transpiration remove nearly all water from shallow soil
layers within days of rainfall, so that in the absence of rapid
drainage through macropores, water does not infiltrate deeply
into the soil profile. In winter, evaporation and transpiration
are limited, so water can accumulate and infiltrate deeper into
the soil profile. This spatial and temporal partitioning of
water has been shown to have ecological and evolutionary im-
plications for plant water use strategies (e.g., physiology and
morphology; Cohen 1970, Walter 1979, Schwinning and
Ehleringer 2001). Thus, changes in the seasonality or variability
of precipitation—both predictions of most GCMs (Houghton
et al. 2001)—are likely to affect the distribution of soil mois-
ture in space and time, with ramifications for the perfor-
mance of species and their interactions with other organisms.
Hydraulic redistribution (Burgess et al. 1998) and phenotypic
plasticity may buffer the effects of changes in soil moisture
regimes and thereby increase the resilience of ecosystems to
changes in patterns of precipitation, but the potential for
this buffering effect is not known.
Heterogeneity in environmental conditions and resource
supply rates plays a central role in producing and maintain-
ing patterns of species diversity (Tilman and Pacala 1993,
October 2003 / Vol. 53 No. 10 • BioScience 943
Articles
Figure 1. Predictions of seasonal precipitation regimes for the continental United States from the HadCM2 model, Hadley
Centre for Climate Prediction and Research, for (a) summer (June, July, August) and (b) winter (December, January, Febru-
ary), and from the CGCM1 model, Canadian Centre for Climate Modelling and Analysis, for (c) summer and (d) winter.
Colors indicate trend in precipitation for 2090 as percentage changes relative to the period 1960–1990 (NAST 2000). Source:
US Global Change Research Program public archives (19 August 2003; www.usgcrp.gov/usgcrp/nacc/background/scenarios/
found/figs.html).
80
Percentage
> 100
60
40
20
–60
–40
–20
0
–80
b
a c
d
–100

Chesson 2000). Changes in the stochastic patterns of a
variable environmental factor, such as precipitation, may
have potentially stronger effects on ecological systems than
changes in average conditions or changes in other factors
that are relatively stable over time and space (e.g., [CO2])
(Knapp et al. 2002). Therefore, it is important to focus research
on spatial and temporal variation in precipitation rather
than on yearly or seasonal averages.
Studying the consequences of precipitation variability is far
more difficult than studying the consequences of averages or
gradual changes in climate factors. Patterns and processes of
precipitation regimes occur across a broad spectrum of
spatial and temporal scales (figure 2). Moreover,there may be
lag effects in the responses of ecosystems to changes in pre-
cipitation regimes; for example, if changes in patterns of
precipitation that occur at decadal scales are expected, the
effects of these changes may become apparent only after
perhaps a century under the new regime, when actual climate
patterns may have already shifted again. Faced with this
logistical and conceptual challenge, researchers across all
disciplines must pay special attention to developing experi-
ments at appropriate spatial and temporal scales, practicing
restraint in data interpretation, and developing models and
analyses that prudently extrapolate long-term effects from
short-term data. This will require understanding (or at least
considering) the relative importance of the various biotic
and abiotic factors that drive the ecosystem; the sensitivities
and lag times of the component species and processes; and the
recent climatic, evolutionary, and societal history of the
ecosystem.
Which ecosystems and ecosystem processes
are sensitive to changes in precipitation?
Clearly,arid and semiarid regions of the world are highly de-
pendent on the availability of water, which more than any
other factor dominates recruitment, growth and reproduction,
nutrient cycling, and net ecosystem productivity (figure 3;
Noy-Meir 1973, Leith 1975,Smith et al. 1997).For example,
predicted increases in summer precipitation might contribute
to a substantial “greening” across wide areas of the arid South-
west, primarily by increasing the density and relative pro-
duction of C4grasses (Neilson and Drapek 1998). In addition,
precipitation is often a limiting factor in more mesic terres-
trial ecosystems. For example,native tallgrass prairies in the
US Central Plains experience substantial interannual varia-
tions in production that are tightly coupled to annual pre-
cipitation (Sala et al. 1998). Similarly, prairie irrigated to re-
place evapotranspiration losses during the growing season
produced on average 26% more biomass than control plots
that received only ambient precipitation (Knapp et al.2001).
Knapp and Smith (2001) concluded that herbaceous-
dominated systems, such as grasslands and old fields of the
central United States, exhibit greater in-
terannual variability than other systems in
aboveground net primary production
(ANPP) under current precipitation
regimes and thus may be more responsive
to future shifts in precipitation. In tem-
perate forests, net primary production
and stand water use are correlated with in-
terannual variation in precipitation and
the frequency and periodicity of drought,
and differential growth and survivorship
of juvenile trees may ultimately shift
species composition (Hanson and Weltzin
2000). Thus, it appears that most ecosys-
tems are sensitive to precipitation change;
however, at this point the potential con-
sequences of these sensitivities are largely
unknown.
Changes in global and regional precip-
itation regimes are expected to have im-
portant ramifications for the distribution,
structure, composition, and diversity of
plant, animal, and microbe populations
and communities and their attendant
ecosystems (Easterling et al. 2000,
Houghton et al. 2001, Weltzin and
McPherson 2003).Long-term monitoring
studies suggest that recent climatic and
atmospheric trends, which are anomalous
relative to past climate variation, are
944 BioScience • October 2003 / Vol. 53 No. 10
Articles
Figure 2. Variations in the spatial and temporal distribution of various factors
that comprise or dictate precipitation regimes, from individual convective storms
with local distributions and short duration to hemisphere and global-scale
oscillations in atmospheric conditions that occur on a decadal scale. Abbrevia-
tions: AO, Arctic Oscillation; ENSO, El Niño–Southern Oscillation; IPO, Inter-
decadal Pacific Oscillation; NAO, North Atlantic Oscillation; PDO, Pacific
Decadal Oscillation.

already affecting species physiology, distribution, and pheno-
logy (Hughes 2000). Moreover, the secondary effects of
changes in species composition on ecosystem processes are
likely to be as important as the direct effects of climate change.
Changes in species composition could affect primary and
secondary production, rates of decomposition and biogeo-
chemical cycling, frequency and intensity of wildfire, avail-
ability of water resources, and fluxes of energy and materials
between the biosphere and the atmosphere (see figure 3;
Pastor and Post 1988, Hungate et al.1996, Grime et al.2000,
Bachelet et al. 2001). In time, changes in community and
ecosystem structure are likely to cause feedback effects: A
change in soil organic matter content and concordant changes
in water-holding capacity, for example, might engender fur-
ther changes in plant composition, litter chemistry, and rates
of decomposition. Changes in precipitation may also in-
crease the susceptibility of ecosystems to invasion by nonnative
plant species (Weltzin et al. 2003) and affect the spatial and
temporal dynamics of consumers at other trophic levels
(Ernest et al. 2000, Staddon et al. 2003).
October 2003 / Vol. 53 No. 10 • BioScience 945
Articles
Figure 3. Conceptual model of interactions between global-scale climate processes; land use and cover; soil moisture; and
ecosystem-, community-, population-, and individual-level processes. Soil moisture is also controlled by local- to landscape-
scale characteristics of soil and hydrologic characteristics (e.g., texture, slope, vegetation cover, antecedent moisture condi-
tions). Climate change and soil moisture affect each level of the hierarchy across a range of spatial and temporal scales (solid
lines). Responses of individuals and populations indirectly control soil moisture and community-, ecosystem-, and global-
scale processes. Factors within each level of the hierarchy are capable of interacting. Abbreviations: [CO2], concentration of
carbon dioxide; NPP, net primary production.

Although perhaps of secondary importance in arid and
semiarid ecosystems, other factors of global change are
expected to modify the effects of precipitation change (see
figure 3). Increases in temperature will affect rates of evap-
oration, with ramifications for ecosystem water budgets,
and may indirectly affect processes of soil respiration, net ni-
trogen mineralization, and plant productivity (Shaver et
al. 2000). Moreover, soil moisture regimes may be affected
if warming causes the primary composition of winter pre-
cipitation to shift from snow to rain or if snow melts ear-
lier in the spring. These effects on snow may be most im-
portant in ecosystems with relatively dry summers. Increases
in [CO2] may alter rates of plant transpiration or water use
efficiency or accentuate or attenuate the effects of increased
temperature or water stress on rates of assimilation and
production (Owensby et al. 1999, Shaw et al. 2002). Con-
versely, changes in precipitation may control plant and
ecosystem responses to changes in [CO2] and temperature
(Smith et al. 2000).
How will changes in precipitation alter
human–ecosystem interactions?
Changes in precipitation regimes are likely to alter the types
and quantities of goods and services that ecosystems provide
to humans. Models that incorporate predicted changes in
climate and [CO2] suggest that enhanced accumulation of bio-
mass in natural ecosystems during wet periods will lead to
greater fuel accumulation, with potential ramifications for
wildland fire regimes (Smith et al. 1997, 2000). Increases in
the variability of precipitation—but not necessarily in the to-
tal amount of precipitation—may reduce grassland produc-
tivity (Knapp et al. 2002) and livestock carrying capacity,ex-
acerbate overgrazing, increase rangeland susceptibility to
invasions by weed species, and lower agricultural income by
increasing input costs and reducing productivity. Changes in
precipitation timing and magnitude may also affect human
health. Heavy rainfall associated with the ENSO events of the
1990s increased seed and rodent populations, which favored
the virus that causes hantavirus pulmonary syndrome in hu-
mans (Yates et al. 2002). Studies that combine epidemiolog-
ical and climate-change modeling point to northward ex-
pansion of the North American range of mosquito-borne
diseases, including malaria, dengue, and West Nile virus
(Rogers and Randolph 2000).
What approaches are available to study
the effects of precipitation change?
Given the diversity of terrestrial ecosystems and the breadth
of potential response variables of interest, accurate forecasts
of the most likely response of ecosystems to changes in pre-
cipitation regimes will require considerable research. Past
studies of precipitation effects fall into four broad categories:
(1) long-term observations of population and community
change in conjunction with records of precipitation history,
(2) short-term experimental manipulations of soil moisture,
(3) hydroecologic modeling, and (4) cross-site comparisons.
In combination, these approaches provide the insight neces-
sary to form a more complete framework for research and
management.
Long-term observations. A number of long-term observations
of community change, particularly in arid environments
(e.g., Goldberg and Turner 1986), and reconstructions of
prehistoric precipitation and vegetation changes (McAuliffe
and van Devender 1998) have been critical to formulating the
basic ideas concerning the role of precipitation variability and
change in terrestrial communities and ecosystems.However,
because observational approaches rely on inferences drawn
from correlational analyses, conclusions from these studies are
necessarily uncertain. One difficulty lies in distinguishing
the effects of precipitation from the effects of other factors that
vary independently during the observation interval (e.g.,
changes in temperature or in the activity of organisms at
other trophic levels). In addition,extrapolating results from
these studies to predict potential consequences of future
climate scenarios is problematic because of the uncertainties
associated with selecting conditions that adequately represent
future climates.
Short-term experimental manipulations of soil moisture.
Experimental alteration of soil moisture is a logical way to de-
termine how precipitation change affects communities and
ecosystems on relatively short time scales. Several techniques
can be used to manipulate soil moisture, including plot-scale
irrigation and the establishment of rainout shelters to with-
hold rainfall over periods of time (figure 4). However, these
techniques face logistical and conceptual challenges (Weltzin
and McPherson 2003). Logistical constraints on the experi-
mental manipulation of precipitation include difficulties in
simulating the characteristics of actual precipitation (e.g.,
drop size, intensity, nutrient content, rainfall versus snowfall,
rates of infiltration and runoff); overwhelming effects of the
environment external to relatively small experimental units;
transport of irrigation water to often-remote field sites; and
undesired experimental artifacts (e.g., increased herbivory,al-
teration of microclimate). Conceptual limitations include
difficulties in determining the timing and magnitude of wa-
ter applications or withholdings vis-à-vis natural variation in
rainfall patterns; in the choice of adequate response vari-
ables and observational periods; and in scaling across space
and time. Although such constraints can be overcome by
careful design of experiments, funding often remains a lim-
iting factor.
To facilitate comparison across ecosystems and regions,
rainfall manipulation experiments should employ a com-
mon methodology and measure a common set of response
variables over a fixed period of time. That said, it is clear that
different ecological systems will require different manipula-
tive techniques: Simulation of rainfall in grassland, for ex-
ample, is certainly easier than in forest or woodland. More-
over, the particular precipitation regime chosen will vary
depending on the research question, which may focus on
946 BioScience • October 2003 / Vol. 53 No. 10
Articles

the role of means versus extremes of amount, summer ver-
sus winter precipitation, or high frequency versus high in-
tensity of rainfall events.
Modeling. Models used to investigate the role of precipitation
and water in ecosystems are numerous,ranging from highly
mechanistic models that explore the consequences of com-
plex hydrologic–ecologic process interactions to rule-based
models that seek to predict large-scale patterns. While mech-
anistic models employ exact mathematical relationships de-
rived from simplified physical models, rule-based models
employ “if–then” rules that summarize and synthesize system
behaviors that may have complex root causes. For example,
where a mechanistic model of plant water uptake during a
growing season could involve equations describing the rates
of water movement from the soil through the plant, as well
as rates of leaf gas exchange and the transformation of car-
bon gain into biomass, a rule-based model could simply
state,“If precipitation is below a threshold value, then some
fraction of it is taken up by plants, ELSE water uptake is at a
specified maximum.” In reality all complex models have
quantitative–mechanistic representations and rules and
differ only in the degree to which rules, rather than physical
processes, dominate the model results.
Mechanistic process models that link hydrology and veg-
etation can expose fundamental relationships between pat-
terns of precipitation, characteristics of soil, and properties
of vegetation. For example, recent advances in understand-
ing hydrologic transport in the soil–plant–atmosphere con-
tinuum have improved predictions of transpiration regula-
tion by plants and of limits to drought tolerance (Sperry et
al. 2002). In contrast, global-scale and rule-based models
can examine regional to continental relationships between pre-
cipitation and vegetation patterns or ecosystem processes
(e.g., VEMAP 1995).
Rule-based models have a variety of ecohydrologic
assumptions related to parameterization of precipitation in-
puts and modeling of soil water budgets (table 1). In general,
equilibrium models (e.g., MAPSS [Mapped Atmosphere–
Plant–Soil System], BIOME2, and DOLY [Dynamic Global
Phytogeography Model]; table 1) adjust leaf area index or
related variables to maximize annual ecosystem water uptake
or use, provided that other resources (e.g., light or nutri-
ents) are not limiting. For water-limited regions, this
assumption often leads to the conclusion that almost all the
water that enters the soil is removed by evapotranspiration
in the course of a year. Since the location of soil moisture
storage and the ratio of transpiration to evaporation depend
strongly on temperature, the seasonal distribution of pre-
cipitation plays a major role in selecting vegetation charac-
teristics, such as rooting depths and dought tolerance.
However, because equilibrium models are static, they are
unable to generate predictions for changes in precipitation
variability.
A new generation of models called dynamic global vege-
tation models, or DGVMs (e.g., IBIS [Integrated Biosphere
Simulator] and HYBRID; table 1), integrate the objectives of
vegetation and ecosystem modeling. Because rates of growth
and senescence can be calculated explicitly for all plant types,
these models can be executed dynamically, and vegetation pat-
terns emerge directly from the representation of resource
competition or recruitment and mortality events. These
models are also capable of simulating transient ecohydrologic
conditions, such as those caused by fire or by interannual vari-
ation in precipitation (e.g., El Niño or La Niña events).Dif-
ferences in the ecohydrologic assumptions of DGVMs may
have a greater impact on model solutions than they do in equi-
librium models, because DGVMs lack a common objective
function. Tests and intermodel comparisons of these highly
complex DGVMs should illustrate model sensitivities and im-
prove their convergence (Cramer et al. 2001).
October 2003 / Vol. 53 No. 10 • BioScience 947
Articles
Figure 4. Techniques for experimental manipulation of
precipitation. Top: A precipitation shelter in mesquite
(Prosopis) grassland south of Tucson, Arizona. Twelve
experimental plots (1.5 meters [m] by 1.8 m) under each
of a total of six such shelters (spread across two soil types)
are hand-watered 42 times each year to mimic shifts in
seasonal precipitation regimes. Photograph: Nathan
English. Bottom: One of 12 precipitation shelters in tall-
grass prairie at the Konza Prairie Research Natural Area
in northeastern Kansas. Experimental plots (7.6 m by
7.6 m) under each shelter are watered to simulate shifts
in seasonal timing of precipitation and changes in the
frequency of rainfall events within the growing season.
Photograph: Philip A. Fay.

One of the greatest uncertainties in global models is the rep-
resentation of root structure and function (Feddes et al.
2001), which in current models is oversimplified, with little
consideration of known hydraulic transport laws. For ex-
ample, hydraulic redistribution by plant roots (Burgess et
al. 1998) is not considered in any global model, though it may
be important for drought resilience, nutrient uptake,or com-
petitive interactions. Furthermore,physiological integration
of plant water uptake from layered soil is handled poorly
across models. Another limitation of current global models
is the representation of the root zone depth, which varies
between models but usually does not vary between biomes
within a model (but see Kleidon and Heimann 1998). Root
zone depth assignments can have large-scale effects on global
change predictions (Hallgren and Pitman 2000).
Cross-site comparisons. Manipulative experiments will always
be limited by the length of time over which a treatment can
be applied, the spatial scale over which moisture can be added
or withheld, and the response variables measured. Thus,
there is a need for alternative approaches that can be used to
extrapolate the results of isolated experiments. These include
cross-site comparisons or focused gradient studies, in which
the same research question and methodology are applied
along environmental gradients related to precipitation (e.g.,
amount, seasonality, or variations).
Most cross-site comparisons are observational, but these
space-for-time substitutions can contribute substantially to
scientific understanding of relationships between biotic and
abiotic variables (Leith 1975, Webb et al. 1978, Le Houerou
et al. 1988). For example, Knapp and Smith (2001) used data
from 11 sites in the US Long Term Ecological Research
Network to demonstrate that at continental scales, ANPP
was strongly correlated with mean annual precipitation
(MAP; figure 5). However, their research indicated that in-
terannual variability in ANPP was not related to variability
in precipitation; instead, maximum variability in ANPP
occurred in biomes where high potential growth rates of
948 BioScience • October 2003 / Vol. 53 No. 10
Articles
Table 1. Representation of precipitation and precipitation effects in selected global biogeography and biogeochemistry
models.
Precipitation or
Precipitation Number of Hydrological soil moisture effects
Model input soil layers processes modeled on biotic components Reference
MAPSS: Equilibrium Monthly mean 3 Interception, infiltration, LAIi~ monthly water added Neilson 1995
biogeography model as snow or rain runoff, snowmelt, downward to layers 1 and 2 (via
percolation between layers, equilibrium assumption)
base flow
BIOME2: Equilibrium Monthly mean 2 Runoff, downward percolation Daily Ti~ soil moisture in Haxeltine et al. 1996
biogeography model as rain between layers, bare soil root zone for plant type i;
evaporation monthly Ai~ monthly fixTi/
monthly water added; ∑fi~
monthly water added to layers
1 and 2 (via equilibrium
assumption)
DOLY: Equilibrium NPP Monthly mean 1 Interception, unspecified Gs~ soil moisture content; Woodward et al. 1995
model as rain outflow of prt in excess of ET LAI ~ monthly prt (via
equilibrium assumption)
CENTURY: Dynamic Monthly mean Variable Interception, surface evaporation, Pool decomposition rates ~ Parton et al. 1993
biogeochemistry model as snow or rain saturated downward flow between pool soil moisture; mineral N
layers, deep drainage leaching ~ saturated flow
between layers; production ~
monthly (prt + residual soil
moisture at 0 to 60 cm)/PET;
senescence ~ soil moisture at
0 to 60 cm; root/shoot ~ annual
prt; monthly T ~ soil moisture
content by layer
IBIS: Dynamic Hourly 6 Interception, runoff, surface Hourly Ai~ soil moisture in Foley et al. 1996
biosphere model evaporation, bidirectional water root zone for plant type i;
transport between layers, deep hourly Ti~ Ai
drainage
HYBRID: Dynamic Daily mean as 1 Interception, snowmelt, unspeci- Gs~ soil water potential; pool Friend et al. 1997
biosphere model snow or rain fied outflow of soil moisture decomposition rates ~ percent-
above 1.5 xfield capacity age of water-filled pore space
~, is a function of; A, assimilation rate; DOLY, Dynamic Global Phytogeography Model; ET, evapotranspiration; f, ground-cover fraction; Gs,stomatal
conductance for water vapor; i, plant type; IBIS; Integrated Biosphere Simulator; LAI, leaf area index; MAPSS, Mapped Atmosphere–Plant–Soil System;
N, nitrogen; NPP, net primary production; PET, potential evapotranspiration; prt, precipitation; T, transpiration rate.
Note: Interactions that do not directly involve precipitation or soil moisture are omitted.

herbaceous vegetation were combined with moderate vari-
ability in precipitation. A recent analysis of the same data sets
used by Knapp and Smith (2001) indicated that interannual
variability in ANPP was strongly influenced by MAP at the
most arid sites but only weakly related to MAP at more mesic
sites, particularly those within forest biomes.
Cross-site comparisons of manipulative experiments
have the potential to contribute even more information
about ecosystem sensitivities, critical thresholds, and
local- to broad-scale mechanisms that control the response
of a variety of ecosystems to changes in precipitation regimes.
Currently, opportunities for cross-site comparisons of
experimental manipulations are limited because of the
variety of methods employed for the application or re-
moval of precipitation, and because response variables and
assessment techniques differ from site to site (Weltzin and
McPherson 2003). As scientists conducting CO2enrich-
ment experiments determined more than a decade ago,
cross-site comparisons would be facilitated if researchers
agreed on common protocols for precipitation manipula-
tion and sampling that are applicable to all sites and com-
patible with a variety of research questions. This reasoning
motivated the formation of PrecipNet, described below.
PrecipNet: An interdisciplinary research
network focused on changing precipitation
regimes
Ecologists and hydrologists from various terrestrial
ecosystem study sites, along with climate model-
ers and social scientists, have formed PrecipNet, an
international and interdisciplinary network for
precipitation and ecosystem change research
(http://zzyx.ucsc.edu/ES/PrecipNet.htm). The
purpose of this network is to promote communi-
cation, intellectual exchange, and integration
of methods and results among research groups
interested in how potential future precipitation
regimes may affect physical and biological processes
across ecological, geographic, and disciplinary
boundaries (box 1). Most of the current PrecipNet
participants and their study sites are located in
arid or semiarid regions, where water availability
imposes the strongest control over community
and ecosystem dynamics and processes.However,
as awareness of the network has grown, a number
of national and international sites from more
mesic regions have been added. To date, research
at most sites focuses on mechanisms likely to
govern the response of the structure and func-
tion of communities and ecosystems to changes
in precipitation regimes. Study sites include aca-
demic research stations, private biological research
stations, and other sites at national and interna-
tional institutions dedicated to research, conser-
vation, or management.
Research needs and directions
Predictions of future precipitation regimes depend on out-
put from GCMs, which are constantly being improved. Most
GCMs are parameterized at the global scale, with grid cells that
can encompass entire biogeographic regions, although in-
creasing numbers are executed at regional scales (e.g., Giorgi
et al. 1998, Snyder et al. 2002). Prediction of the effects of pre-
cipitation change on vegetation will require output from
local or regional models at monthly or even daily temporal
resolutions. Such scenarios could form the basis for new
field experiments in ecosystems (e.g., grasslands) predicted
to be highly sensitive to precipitation change (Knapp and
Smith 2001).
The relationship between climate models and experiments
should be reciprocal: Model predictions can serve as most-
likely scenarios of climate change that delimit field experi-
ments, while the results from field experiments can facilitate
model parameterization, particularly if they incorporate gra-
dients of driving variables. In addition, climate models should
be linked with DGVMs to model feedbacks between terres-
trial vegetation and climate. Constructive interactions between
modelers and empiricists will strengthen linkages between
models and experiments, to the benefit of ecology,manage-
ment, planning, and policymaking. Critical observations of
October 2003 / Vol. 53 No. 10 • BioScience 949
Articles
Figure 5. Relationship between annual aboveground net primary pro-
duction (ANPP) and mean annual precipitation for 11 US Long Term
Ecological Research sites representing five biomes: (1) arctic and alpine
(ARC, Arctic Tundra; NWT, Niwot Ridge), (2) desert (JRN, Jornada; SEV,
Sevilleta), (3) grassland (CDR, Cedar Creek; KNZ, Konza Prairie; SGS,
Shortgrass Steppe), (4) old-field (KBS, Kellogg Biological Station), and
(5) forest (BNZ, Bonanza Creek; HBR, Hubbard Brook; HFR,
Harvard Forest).Reprinted with permission from Knapp and Smith
(2001). ©2001 American Association for the Advancement of Science.
Precipitation (millimeters)
ANPP (g per m2)
ANPP = 75.34 + 0.37 (Precipitation)

alterations in species composition after environmental per-
turbations (e.g., Allen and Breshears 1998) will complement
improved models of vegetation dynamics and enhance con-
fidence in predictions about the fate of communities and
ecosystems over decadal temporal periods. Moreover, addi-
tional research should focus on the representation of below-
ground processes, such as root structure and function, phe-
notypic plasticity, hydraulic redistribution, and water uptake
vis-à-vis root zone depth, which may vary within and between
biomes.
Although the research cited in this review provides a broad
cross-section of ecological research and study systems, many
important terrestrial systems remain relatively unstudied.
Research is notably sparse in deciduous forests, coniferous
woodlands and forests, shrublands, and tropical wet and
seasonal forests. The paucity of data from these and other
systems limits our ability to generalize about the response of
species, growth forms, life forms, community-level proper-
ties (e.g., productivity, diversity), or ecosystem attributes
(e.g., nutrient cycling, energy flows) to changing precipitation
regimes.
Even in systems that are being studied, background infor-
mation on the broader ecological, climatological, and socio-
logical circumstances of the study area is usually limited.
Moreover, it is unclear whether the particular site choices for
experiments are highly representative of the most common
background conditions of a region (e.g., characteristics of soil,
frequency of disturbance, and patterns of land use). To over-
come such limitations and uncertainties, experiments should
focus on interactions between various precipitation regimes
and other important factors. Where feasible, new precipita-
tion experiments should include elevated [CO2], increased
temperature, or both, to reflect the multiple interacting en-
vironmental changes that will coincide with global change
(e.g., Shaw et al. 2002). To this end, understanding “the re-
sponses of ecosystems to multiple stresses” is one of four
current research imperatives selected by the US Global Change
Research Program for the coming decade (CGCR 1999).
Most of the research described above focuses on the re-
sponse of only one trophic level—primary producers—to
changes in precipitation regimes. However, changes in pre-
cipitation will also affect consumers (Ernest et al. 2000) and
decomposers (Staddon et al. 2003), which will have feed-
back effects on vegetation though changes in rates of polli-
nation, seed dispersal, granivory, herbivory, nutrient
cycling, and substrate alteration. Clearly, more studies are
needed to address potential responses of other trophic levels
and especially how interactions between trophic levels con-
strain ecosystem responses. Finally, the transfer of technology
and ecological understanding to policymakers at the landscape,
regional, and national levels will be critical. Effective com-
munication will require a synthesis of information relevant
to the variety of different spatial and temporal scales
considered by ecologists, land managers, stakeholders, and
policymakers.
950 BioScience • October 2003 / Vol. 53 No. 10
Articles
The Precipitation and Ecosystem Change Research
Network (PrecipNet) was formed to address several
elements missing in the study of precipitation and eco-
system change and the assessment of resultant impacts
on humans.
•Research coordination, communication, and integration.
PrecipNet will establish a database for exhibiting and
organizing precipitation manipulation experiments
performed in various ecosystems on several spatial and
temporal scales, using a variety of tools. This will form
the basis for developing standard approaches for future
experiments designed to improve opportunities for mean-
ingful cross-experimental comparisons. It will also help
identify knowledge gaps and suggest opportunities for
research. PrecipNet will interact with other research net-
works, such as BASIN (Biosphere–Atmosphere Stable
Isotope Network), C.DELSI (Center for the Dynamics
and Evolution of the Land–Sea Interface), SAHRA (Sus-
tainability of Semi-Arid Hydrology and Riparian Areas),
and CIRES (Cooperative Institute for Research in Envi-
ronmental Sciences) Western Water Assessment.
•Regional comparisons of precipitation change and its
effects. The database will provide opportunities to analyze
intra- and interregional patterns and processes, such as
relationships between current precipitation regimes and
ecosystem structure and function, and potential impacts
of changes in precipitation regimes on different ecological
systems.
•Fostering multidisciplinary activities. PrecipNet will
sponsor activities that foster communication between
biologists, hydrologists, climate modelers, and social
scientists. These activities will include workshops to assess
the impacts, vulnerability, and mitigation of precipitation
change effects and to encourage the formation of multi-
disciplinary research groups.
•Promoting skill development and technology transfer.
PrecipNet will coordinate the exchange of graduate
students and postdoctoral researchers between research
groups to promote communication, facilitate cross-site
comparisons and proposal development, and increase
skills for working in multidisciplinary groups.
•Participants. PrecipNet will also sponsor interactions
between scientists, stakeholders, and the public.These
interactions will serve both to disseminate knowledge gen-
erated by PrecipNet members and to help members
develop and refine useful research questions.
Box 1. PrecipNet objectives

Acknowledgments
This work was conducted as part of the PrecipNet: Analysis
and Synthesis of Precipitation and Ecosystem Change Work-
ing Group (principal investigator M. E. L.), supported by
the National Center for Ecological Analysis and Synthesis, a
center funded by the National Science Foundation (NSF
grant no. DEB-0072909) and the University of California
and its Santa Barbara campus. The authors acknowledge the
support of funding agencies including the US Department of
Energy, the National Park Service, NSF, and the US
Department of Agriculture. Guy McPherson and George
Koch supported and encouraged the production of this
review, helped define some of the conceptual framework,
and provided suggestions that improved early drafts of the
manuscript.
References cited
Allen CD, Breshears DD. 1998. Drought-induced shift of a forest–woodland
ecotone: Rapid landscape response to climate variation. Proceedings of
the National Academy of Sciences 95: 14,839–14,842.
Bachelet D, Neilson RP, Lenihan JM, Drapek RJ. 2001.Climate change effects
on vegetation distribution and carbon budget in the United States.
Ecosystems 4: 164–185.
Burgess SSO,Adams MA, Turner NC, Ong CK.1998. The redistribution of
soil water by tree root systems. Oecologia 115: 306–311.
[CGCR] Committee on Global Change Research. 1999.Global Environmental
Change: Research Pathways for the Next Decade. Washington (DC):
National Academy Press.
Chesson P. 2000. Mechanisms of maintenance of species diversity. Annual
Review of Ecology and Systematics 31: 343–366.
Cohen D.1970.The expected efficiency of water utilization in plants under
different competition and selection regimes. Israel Journal of Botany 19:
50–54.
Collier M, Webb RH. 2002. Floods, Droughts, and Climate Change. Tucson:
University of Arizona Press.
Cramer W, et al. 2001. Global resonse of terrestrial ecosystem structure and
function to CO2and climate change: Results from six dynamic global
vegetation models. Global Change Biology 7: 357–373.
Eagleson PS. 2002. Ecohydrology: Darw inian Expression of Vegetation Form
and Function. Cambridge (United Kingdom): Cambridge University
Press.
Easterling DR, Meehl GA,Parmesan C, Changnon SA, Karl TR, Mearns LO.
2000. Climate extremes: Observations, modeling, and impacts.Science
289: 2068–2074.
Ernest SKM, Brown JH, Parmenter RR. 2000. Rodents, plants, and precip-
itation: Spatial and temporal dynamics of consumers and resources.
Oikos 88: 470–482.
[FAO] Food and Agriculture Organization. 1993. Sustainable Development
of Drylands and Combating Desertification. Rome: FAO of the United
Nations.
Feddes RA, et al. 2001. Modeling root water uptake in hydrological and
climate models. Bulletin of the American Meteorological Society 82:
2797–2809.
Foley JA, Prentice IC, Ramankutty N, Levis S, Pollard D, Sitch S, Haxeltine
A. 1996. An integrated biosphere model of land surface processes, ter-
restrial carbon balance, and vegetation dynamics. Global Biogeochem-
ical Cycles 10: 603–628.
Friend AD, Stevens AK, Knox RG, Cannell MGR. 1997. A process-based,
terrestrial biosphere model of ecosystem dynamics (HYBRID v3.0).
Ecological Modelling 95: 249–287.
Giorgi F,Mearns LO, Shields C, McDaniel L.1998. Regional nested model
simulations of present day and 2 xCO2climate over the central plains
of the United States. Climatic Change 40: 457–493.
Goldberg DE, Turner RM. 1986. Vegetation change and plant demography
in permanent plots in the Sonoran Desert, USA. Ecology 67: 695–712.
Grime JP, Brown VK, Thompson K,Masters GJ, Hiller SH, Clarke IP,Askew
AP, Corker D,Kielty JP.2000. The response of two contrasting limestone
grasslands to simulated climate change. Science 289:762–765.
Hallgren WS, Pitman AJ. 2000. The uncertainty in simulations by a global
biome model (BIOMES) to alternative parameter values. Global Change
Biology 6: 483–495.
Hanson PJ, Weltzin JF. 2000. Drought disturbance from climate change:
Response of United States forests. Science of the Total Environment
262: 205–220.
Haxeltine A, Prentice IC, Creswell DI. 1996. A coupled carbon and water flux
model to predict vegetation structure. Journal of Vegetation Science 7:
651–666.
Houghton JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, Dai X,
Maskell K, Johnson CA. 2001. Climate Change 2001: The Scientific
Basis. Contribution of Working Group 1 to the Third Assessment Report
of the Intergovernmental Panel on Climate Change. Cambridge (United
Kingdom): Cambridge University Press.
Hughes L. 2000. Biological consequences of global warming: Is the signal
already apparent? Trends in Ecology and Evolution 15: 56–61.
Hungate BA,Canadell J, Chapin FS III. 1996. Plant species mediate changes
in soil microbial N under elevated CO2.Ecology 77: 2505–2515.
Kleidon A, Heimann M. 1998. A method of determining rooting depth
from a terrestrial biosphere model and its impacts on the global water
and carbon cycle. Global Change Biology 4: 275–286.
Knapp AK, Smith MD. 2001. Variation among biomes in temporal dynam-
ics of aboveground primary production. Science 291: 481–484.
Knapp AK, Briggs JM, Koelliker JK. 2001. Frequency and extent of water
limitation to primary production in a mesic temperate grassland.
Ecosystems 4: 19–28.
Knapp AK, Fay PA, Blair JM,Collins SL, Smith MD, Carlisle JD, Harper CW,
Danner BT, Lett MS, McCarron JK. 2002. Rainfall variability, carbon
cycling and plant species diversity in a mesic grassland. Science 298:
2202–2205.
Koch GW, Mooney HA.1996.Carbon Dioxide and Terrestrial Ecosystems.
San Diego: Academic Press.
Le Houerou HN, Bingham RL, Skerbek W. 1998.Relationship between the
variabiliy of primary production and the variability of annual precipi-
tation in world arid lands. Journal of Arid Environments 15: 1–18.
Leith H. 1975. Modeling the primary productivity of the world. Pages
237–263 in Leith H, Whittaker RH, eds. Primary Productivity of the
Biosphere. Berlin: Springer-Verlag.
McAuliffe JR. 2003. The atmosphere–biosphere interface: The importance
of soils in arid and semi-arid environments. Pages 9–27 in Weltzin JF,
McPherson GR, eds. Changing Precipitation Regimes and Terrestrial
Ecosystems: A North American Perspective. Tucson: University of Ari-
zona Press.
McAuliffe JR, van Devender TR. 1998. A 22,000-year record of vegetation
change in the north-central Sonoran Desert. Paleogeography, Paleo-
climatology and Paleoecology 141: 253–275.
[NAST] National Assessment Synthesis Team,US Global Change Research
Program. 2000. Climate Change Impacts on the United States: The
Potential Consequences of Climate Variability and Change. New York:
Cambridge University Press.
Neilson RP. 1995. A model for predicting continental-scale vegetation dis-
tribution and water balance. Ecological Applications 5:362–385.
Neilson RP, Drapek RJ. 1998. Potentially complex biosphere responses to tran-
sient global warming. Global Change Biology 4: 505–521.
Noy-Meir I. 1973. Desert ecosystems: Environment and producers. Annual
Review of Ecology and Systematics 4: 23–51.
Owensby CE, Ham JM, Knapp AK, Auen LM.1999. Biomass production and
species composition change in a tallgrass prairie ecosystem after long-
term exposure to elevated atmospheric CO2.Global Change Biology 5:
497–506.
October 2003 / Vol. 53 No. 10 • BioScience 951
Articles

Parton WJ, et al. 1993. Observations and modeling of biomass and soil
organic-matter dynamics for the grassland biome worldwide. Global
Biogeochemical Cycles 7: 785–809.
Pastor J, Post WM. 1988. Response of northern forests to CO2-induced
climate change. Nature 334:55–58.
Rogers DJ, Randolph SE. 2000. The global spread of malaria in a future,
warmer world. Science 289: 1763–1766.
Sala OE, Lauenroth WK,Parton WJ, Trlica MJ. 1981. Water status of soil and
vegetation in a shortgrass steppe. Oecologia 48: 327–331.
Sala OE, Parton WJ, Joyce LA, Lauenroth WK. 1998. Primary production of
the central grassland region of the United States. Ecology 69: 40–45.
Schwinning S, Ehleringer JR. 2001.Water use trade-offs and optimal adap-
tations to pulse-driven arid ecosystems. Journal of Ecology 89: 464–480.
Shaver GR, et al. 2000. Global warming and terrestrial ecosystems: A
conceptual framework for analysis. BioScience 50: 871–882.
Shaw MR, Zavaleta ES, Chiariello NR, Cleland EE, Mooney HA, Field CB.
2002. Grassland responses to global environmental changes suppressed
by elevated CO2.Science 298: 1987–1990.
Smith SD,Monson RK,Anderson JE. 1997. Physiological Ecology of North
American Desert Plants. New York: Springer-Verlag.
Smith SD,Huxman TE,Zitzer SF, Charlet TN, Housman DC,Coleman JS,
Fenstermaker LK, Seemann JR, Nowak RS. 2000. Elevated CO2increases
productivity and invasive species success in an arid ecosystem.Nature 208:
79–82.
Snyder ML, Bell JL, Sloan LC, Duffy PB, Govindasamy B. 2002. Climate
responses to a doubling of atmospheric carbon dioxide for a climatically
vulnerable region. Geophysical Research Letters 29:U383–U386.
Sperry J, Hacke U, Oren R, Comstock J. 2002. Water deficits and hydraulic
limits to leaf water supply. Plant Cell and Environment 25: 251–263.
Staddon PL, Thompson K, Jakobsen I, Grime JP, Askew AP, Fitter AH. 2003.
Mycorrhizal fungal abundance is affected by long-term climatic
manipulations in the field. Global Change Biology 9: 186–194.
Tilman D, Pacala S. 1993. The maintenance of species richness in plant
communities. Pages 13–25 in Ricklefs RE, Schluter D, eds. Species Diversity
in Ecological Communities: Historical and Geographical Perspectives.
Chicago: University of Chicago Press.
[VEMAP] Vegetation/Ecosystem Modeling and Analysis Project. 1995.
Vegetation/ecosystem modeling and analysis project—comparing
biogeography and biogeochemistry models in a continental-scale study
of terrestrial ecosystem responses to climate-change and CO2doubling.
Global Biogeochemical Cycles 9: 407–437.
Vitousek PM. 1994. Beyond global warming: Ecology and global change.
Ecology 75: 1861–1876.
Walter H. 1979. Vegetation of the Earth and Ecological Systems of the
Geo-Biosphere. 2nd ed. New York: Springer-Verlag.
Webb W, Szarek S, Lauenroth W, Kinerson R, Smith M. 1978. Primary
productivity and water use in native forest, grassland, and desert eco-
systems. Ecology 59: 1239–1247.
Weltzin JF, McPherson GR, eds. 2003. Changing Precipitation Regimes
and Terrestrial Ecosystems: A North American Perspective. Tucson:
University of Arizona Press.
Weltzin JF, Belote RT, Sanders NJ. 2003. Biological invaders in a greenhouse
world: Will elevated [CO2] enhance the spread and impact of plant
invaders? Frontiers in Ecology and the Environment 1: 146–153.
Woodward FI, Smith TM, Emanuel WR.1995. A global land primary pro-
ductivity and phytogeography model. Global Biogeochemical Cycles 9:
471–490.
Yates TL, et al. 2002. The ecology and evolutionary history of an emergent
disease: Hantavirus pulmonary syndrome. BioScience 52: 989–998.
952 BioScience • October 2003 / Vol. 53 No. 10
Articles