Background
Most global change experiments to date have focused on one or two factors that are changing globally. For example, the 1992-1997 Open-top Chamber Experiment at Jasper Ridge focused largely on the effects of higher CO2 concentrations and, to some degree, higher nutrient levels on grassland ecosystems. However, global change entails multiple factors changing concurrently. The Jasper Ridge Global Change Experiment (JRGCE) was designed to explore the impact on grassland ecosystems of four major components of global change that are caused, at least partially, by burning of fossil fuels: warming, elevated CO2, increased precipitation, and increased nitrogen deposition. This is the first study to examine these four factors in a natural ecosystem. The studyıs initial principal investigators were Christopher Field of the Carnegie Institution of Washington and Harold Mooney of Stanford University. The first year of field treatments was the 1998-99 growing season.
The principal advantages of studying grassland ecosystems are that 1) they respond quickly to manipulation of global change factors, and 2) they are fully functional ecosystems with a high diversity of plant functional types. Although the community is dominated by annual grasses, it also includes perennial grasses and many non-grasslike plants, which are collectively termed "forbs." Among the forbs, there are annuals and perennials, some nitrogen-fixing species, and a wide range of flowering times.
Experimental Design
The experiment is located at Stanford Universityıs Jasper Ridge Biological Preserve. This area has a typical Mediterranean climate with a cool, wet growing season (October to May) and a hot, dry summer (June to September). The experiment site is an annual grassland on sandstone-derived soil. The design involves 32 plots, each 2 meters in diameter. Each of these plots receives a CO2 treatment and a warming treatment. The CO2 treatment options are ambient CO2 or enriched CO2 (at approximately twice the ambient level, or 700 ppm). The heat treatments are ambient heat or warming of 1.0-1.5 degrees centigrade at the vegetation canopy level and about half this level at the soil surface. These temperature and CO2 increases are within the range of estimates for the end of this century. Each 2-meter diameter plot is divided into four quadrants of equal size. Each of these quadrants receives a water treatment and a nitrogen treatment. The water treatment options are ambient or a precipitation addition of 50% applied after each natural rainfall event, plus end of rainy season water additions to simulate extending the rainy season by 3 weeks. This water treatment was chosen to simulate the predicted wetter climate on the Pacific Coast. The nitrogen treatment options are ambient or a nitrogen addition that raises total nitrogen deposition to a level approximating some of the more heavily polluted areas in Europe (an addition of 7 grams N per square meter per year). The design results in 128 treatment units (32 plots times 4 quadrants) which provide 8 replicates of all 16 possible combinations of treatments. Four additional plots identified in the field do not have any equipment installed. These plots are used to determine possible effects of experimental apparatus.
Each of the 32 circular plots is separated from the surrounding grassland by a fiberglass sheet that extends 0.4 meters below the ground. The quadrants within the plot are separated by a similar fiberglass barrier below ground to minimize soil and root interactions. Aboveground vertical netting minimizes seed and litter dispersal between quadrants. Each quadrant contains 4000-8000 individuals of locally-dominant higher plants.
Elevated CO2 is achieved by a free-air CO2 enrichment (FACE) system that emits pure CO2 from orifices in plastic tubes that surround the plot at the canopy level. CO2 emission is adjusted continuously for wind speed and wind direction so that CO2 concentrations averaged over every few minutes are almost always close to the target of 700 ppm. Warming is achieved with an array of four overhead infrared heaters. Dummy heater hoods are installed over plots not receiving the heat treatment. Water additions are applied to quadrants by above ground spray emitters. Nitrogen addition is achieved by adding dissolved or encapsulated calcium nitrate to the soil surface.
Central Hypothesis
The project is integrated around the hypothesis that grassland responses to multiple global change factors can be predicted from the responses to single factors. Results supporting this hypothesis would strengthen the likelihood that we can accurately predict ecosystem responses to multifactorial global change based on the results of many previous experiments involving single factors. Results refuting this hypothesis would indicate that single factor experiments have less relevance in understanding the impacts of more complex, and more realistic, global changes. Because of the projectıs high degree of replication and the factorial combination of treatments, this hypothesis can be tested with statistical models that determine whether measured responses to combined treatments are different from the sum of the responses to separate, individual treatments.
The adjustment of plant water budgets played an important role in several responses observed in the experiment. Higher levels of atmospheric CO2 can cause plants to partially close leaf pores (stomata) through which CO2 enters leaves for photosynthesis and water vapor escapes to the atmosphere. This partial closing of the stomata under elevated CO2 can allow the plants to acquire more CO2 with less water loss. Lower plant water loss (higher water use efficiency) can result in increased soil moisture.
The intention is to operate the experiment for a total of at least 15 years in order to measure both short-term and long-term effects of global change on the grassland community.
Measuring and Analyzing Effects of Global Change
Central measurements being made fall into four major categories; net plant primary production (NPP) which is total tissue production less production consumed by respiration, disposition of belowground carbon, the status of water and nutrients available to the plants, and changes in species that comprise the grassland community.
Plant production is measured both above and below ground using a variety of techniques that look at processes contributing to growth, such as photosynthesis, as well as the amount of biomass at different times during the season. During the growing season, a non-invasive technique (NDVI) based on measuring reflected light wavelengths is used to estimate aboveground biomass. At the end of the growing season, plants from a small portion of each plot are harvested, separated by species, and weighed to measure aboveground biomass. Belowground biomass is assessed by measuring root turnover and by analyzing soil cores. Roots are monitored with a small video camera inserted into transparent observation tubes that extend under each quadrant. Root growth and turnover is determined from a time series of video images of roots growing outside the tube. Several times during the growing season a coring device is used to extract soil samples. Root mass and microbial biomass of each soil core is measured.
The disposition of underground carbon is also monitored by analyzing the amount and types of microbes in the soil and by measuring CO2 released through respiration in the soil. Because the added CO2 has a unique isotopic signature, the fate of the added carbon from the CO2 treatment can be tracked through analysis of the isotopic signature of CO2 respired from the soil.
The amounts of water and nutrients available to plants are measured periodically. Small plastic tubes are inserted below the soil surface in each quadrant. The tubes contain resins that bind nitrogen-containing compounds used by the plants. Every few months the resin is replaced and the old resin is analyzed for nitrogen content. A suction device is also used to extract water from the soil which is analyzed to determine nutrient concentrations. The amount of soil moisture available to the plants is determined indirectly by measuring the propagation and reflection of an electrical pulse between two parallel steel rods installed in each quadrant (time domain reflectometry or TDR).
Changes in the species that comprise the grassland community are expected to be an important impact of global change. One non-invasive technique being used to study species composition of the plots involves marking positions in a pre-determined pattern from a reference point above the plot and recording the identities of the plants closest to each position. The harvest of aboveground biomass from a small portion of each quadrant at the end of the season is analyzed by species to provide a second measure of diversity.
Some Findings
After three years of field treatments, tests of the hypothesis that the effects of the four treatments are additive yielded some interesting results which were reported in the December 6, 2002 issue of Science in a paper titled "Grassland Responses to Global Environmental Changes Suppressed by Elevated CO2". This paper reported that elevated CO2 reduced grassland growth when combined with increases in other global change factors. These results demonstrate that ecosystem responses to global change depend strongly on interactions of factors and responses can be nonadditive. Some of the findings reported in Science are outlined below.
Over the first two years of treatments, elevated CO2 had no significant effect on net primary production (NPP) across all treatment combinations. In the third year, elevated CO2 stimulated above ground biomass by an average of 32.6% when all other factors were at ambient levels. This increase was comparable to those seen in other single factor studies. Each treatment involving increased temperature, nitrogen, or precipitation, alone or in combination, tended to increase aboveground and belowground biomass in the third year. However, when CO2 was added there was a significant suppressive effect on NPP in the third year. For example, the combination treatment of temperature, nitrogen and precipitation produced the largest stimulation (84% increase), but when CO2 was added to this treatment the increase was reduced to 40%. The paper concludes: "The NPP response of this ecosystem to multiple global changes was not a simple combination of responses to individual global change factors."
The mechanism causing the suppressive effect of CO2 in the ecosystem is not known. Analysis of data indicates that the suppression may be due to a nutrient limitation that increases over time as the demand for the nutrient increases. For example, studies at Jasper Ridge have shown that several indicators of microbial activity increase under elevated CO2. It is possible that microbial growth is being stimulated by increased carbon loss from roots, and increased microbial consumption of a nutrient is causing a limitation for the plants. Another possible mechanism relates to the plantsı allocation of production between roots and above ground biomass. Allocation to root production was consistently decreased by elevated CO2. This root reduction could result in a nutrient limitation since roots forage for nutrients. Other possible mechanisms are also being investigated. Ultimately, a combination of factors may be found to be responsible for the suppressive effect of CO2 in this ecosystem.
Some other JRGCE findings not covered in the Science paper are discussed below.
The Jasper Ridge Open Chamber Experiments showed that elevated CO2 increased soil moisture due to greater water use efficiency. JRGCE studies during 1999 and 2000 found that both warming and elevated CO2 increased soil moisture during the onset of seasonal drought in California grassland. Independently, as single factor treatments, warming and elevated CO2 caused comparable increases in soil moisture in both years. When both warming and elevated CO2 were present, the increase was approximately additive, resulting in a 15% increase in soil moisture available during the drought onset period. Warming accelerated plant senescence (aging) resulting in less water use late in the growing period and higher soil moisture levels. This finding was unexpected since global warming has been predicted to increase evapotranspiration, causing declines in soil moisture. Higher moisture levels at the beginning of the annual drought period could increase invasion of California grasslands by woody plants and late-season annuals.
Global change factors have the potential to affect species diversity and the composition of natural communities. After three years of treatments, species diversity was increased by added precipitation and decreased by added nitrogen and elevated CO2. Combined nitrogen addition and elevated CO2 reduced diversity by over 15% after 3 years, but their effects were almost fully offset by increases in diversity caused by warming and increased precipitation. The dominant functional groups in the grassland were annual grasses and forbs (non-woody plants excluding grasses and grass-like plants). These functional groups showed different responses to global change factors. Added nitrogen suppressed forb production and increased grass production. Increased precipitation strongly enhanced forb production and had little effect on grasses. Warming enhanced forb production and reduced grass biomass. The combined effects of all four treatments on grasses and forbs tended to offset one another, resulting in a slight overall increase in forb production. Treatment responses by individual species of grasses and forbs tended to be weak, but they were sufficiently consistent to produce stronger responses at the functional group level.
A two year study investigated the response of Baccharis pilularis (coyote bush, a common grassland invader) to global change factors. Across all treatment groups during the first year, elevated CO2 accelerated germination, reduced herbivory damage, and increased survival during the annual onset of drought. During the second year, warming and increased precipitation strongly facilitied establishment of Baccharis. These results indicate that global change may accelerate woody shrub invasion of California grasslands.
The response to global change factors by the dominant nitrogen fixer in the ecosystem, Vicia sativa, was investigated in a two year study. Nitrogen additions strongly suppressed biomass and seed production in both years. Warming increased biomass and seed production in both years. Neither elevated CO2 nor increased precipitation showed significant effects. However, the combination of elevated CO2, warming, and increased precipitation caused the highest increase in production by far.
All of these findings reinforce the importance of ecosystem experiments that allow relatively easy manipulation of multiple global change factors and that produce rapid responses over multiple generations of plants. Also, since long time periods are often necessary for the processes being studied to reach equilibrium (particularly belowground processes), the findings emphasize the need for long term studies in order to predict ultimate ecosystem effects.
Directions of Current Studies
The core experimental framework operating today will continue the existing global change factor manipulations long enough to quantify the transition from the initial transient responses to the long-term responses involving in nitrogen stocks, changes in soil carbon, and species shifts. These responses should be clearly discernable by the end of this decade. Other studies will emphasize responses related to biodiversity, biological invasion, and plant molecular biology.
In addition, a National Science Foundation grant received in 2002 supports investigation of the responses of biogeochemical cycles to global change factors at a deeper level. Specifically, studies are testing the following four hypotheses:
1. Long-term responses of cycles of carbon, nitrogen, and water to global change factors are driven more strongly by changes in the composition of the plant community than by physiological changes in the plants.
2. Long-term changes in the composition and function of the plant community reflect controls by the level and timing of availability of key resources, especially nitrogen, water, and light.
3. Long-term changes in nitrogen availability are controlled by microbial processes.
4. Long-term changes in microbial processes reflect changes in the function and composition of the microbial community which responds to changes in the composition of the plant community providing its substrates.
These directions support the long-term objective of the JRGCE, which is to develop a model framework for generalizing the underlying principles of California grassland ecosystem responses to globally changing environmental factors, and other ecosystems as well.
References
Shaw, MB, Zavaleta, ES, Chiariello, NR, Cleland, EE, Mooney, HA, and Field, CB (2002) Grassland responses to global environmental changes suppressed by elevated CO2. In: Science 298, pp. 1987-1990.
Zavaleta, ES, Thomas, BD, Chiariello, NR, Asner, GP, Shaw, MR, and Field, CB (2002) Plants reverse warming effect on ecosystem water balance. Submitted.