Thirty years ago Charles F. Baes, Jr., a chemist at the U.S. Department of Energy’s Oak Ridge National Laboratory, wrote that the earth was undergoing a great “uncontrolled experiment,” one that would soon reveal the global consequences of rising greenhouse gas concentrations. Today scientists know that deforestation, land use and the burning of fossil fuels are warming our planet. We are less certain, however, about how climate change will alter forests and grasslands, as well as the goods and services these ecosystems provide society.

Much of the climate change news in the mass media comes not from experiments but observations. Scientists monitor Arctic sea ice, glaciers and natural events such as the timing of leaf appearance and inform the public when changes fall outside normal expectations. Recording this kind of information over time is important. But rather than waiting to see how an evolving climate slowly alters the biosphere, climate change biologists are conducting field experiments, often at large scales, to see how ecosystems will respond to more or less precipitation, rising concentrations of carbon dioxide (CO2) and warming temperatures. Experimental data are key to determining if and to what extent ecosystems will be affected by climate change in 10, 50 or 100 years and how those changes might feed back to further advance change. The results can help separate fact from fiction in the climate debate, which is charged with emotion.

For years researchers investigated how single plants—typically grown for several months inside climate-controlled chambers—responded to varied conditions. Understanding mechanisms at this scale is necessary. But we must also study plants in their proper context: actual ecosystems. Largely unknown to the public, several sizable outdoor experiments involving altered precipitation and CO2 concentrations have been under way for more than a decade, including those that are described in boxes on the following pages. Temperature experiments have begun as well. Enough data have now been generated to improve models that predict climate and vegetation changes, providing a more accurate picture of how woodlands, prairies and agricultural crops may change in an increasingly warmer world that is subject to different precipitation patterns and blanketed in more CO2.

Findings around the Globe
Experiments conducted worldwide show that plants and ecosystems possess a remarkable capacity to adjust to new conditions. But scientists expect that thresholds exist beyond which significant and potentially catastrophic responses will occur. As we explore these boundaries, we will find surprises, yet some conclusions informed by field experiment data can already be drawn:

  • Higher CO2 concentrations can enhance yields for commodities such as wheat, rice, barley, soybeans and cotton, but simultaneous warming, and in some locations ozone pollution, may well reduce or negate the “CO2 fertilization” effect. Climate changes will also alter interactions among crops, weeds, pathogens and insects, with the pests winning out as often as not.
  • Deciduous forests in the eastern U.S.—the kind that lose their leaves seasonally—are relatively insensitive to drought. Deep soils hold plenty of water to support the growth of large trees throughout much of the year. But surface soils hold little water and dry out quickly, causing high rates of mortality in young seedlings and small saplings—the forests of the future.
  • In a CO2-enriched atmosphere, greater root growth could provide more nutrients, enhancing the productivity of developing forests. Greater rooting with depth might also benefit plants in arid and dryland ecosystems by increasing access to soil water.
  • Global warming and rising CO2 concentrations could promote the invasiveness of many agricultural weeds, including Canadian thistle, lowering crop yields or demanding more herbicides. Exotic species may also pose problems. For example, recent experiments in the Mojave Desert by Stan Smith of the University of Nevada, Las Vegas, showed that in a year with unusually high rainfall, elevated CO2 concentration stimulated the spread of Bromus tectorum, or cheatgrass, which re­­duced plant species diversity, modified the food chain and raised the potential for fire.
  • Although the invasion of woody plants in world grasslands over the past 200 years has resulted mainly from overgrazing and from fire suppression, rising atmospheric CO2 concentrations may be contributing to the en­­croachment of trees and shrubs across the western U.S.
  • Future CO2 concentrations will affect plants in ways that could impact public health, including greater production of pollens that trigger allergies and greater growth and toxicity of poison ivy and other invasive species.

Complex Questions
The results of large outdoor experiments are telling, but most investigations have been conducted at middle latitudes and mostly in the U.S. and Europe. New experiments at a wider range of latitudes are needed to clearly predict the re­­sponse of boreal, tundra and tropical plants and ecosystems. Several years will be needed to prepare such experiments because they are likely to be scientifically complicated and located in remote regions. They will require significant engineering to ensure that altered conditions are imposed uniformly and that the infrastructure is robust enough to last for years.

Biologists must also build installations that not only alter CO2 concentration, temperature or precipitation patterns, but all three factors in combination. We have so far only scratched the surface. A new experiment near Cheyenne, Wyo., is evaluating how plants in a northern mixed-grass prairie will fare given simultaneous changes in CO2 concentration and temperature. In the first year of the Prairie Heating and CO2 Enrichment experiment, Jack Morgan of the U.S. Department of Agriculture’s Agricultural Research Service has found indications that warming in combination with higher CO2 concentration may enhance the abundance of warm-season grasses in the Great Plains, at the expense of cool-season grasses.

How best to manipulate multiple factors and how to account for such combinations, as well as possible feedbacks, in models are complex questions. We will need experimentally derived data very soon if we are to help society anticipate, plan and adapt to a climate that is already changing rapidly.