15.1 INTRODUCTION

Technology is available, or envisioned, to accomplish experimental manipulation of atmospheric factors in ecosystems representative of global biomes. In some cases, however, it will be necessary to modify and scale up existing equipment originally designed and fabricated for use on smaller samples. Ecosystem environmental control facilities must be of adequate size to contain the ecosystem sample desired. The primary problem is control of atmospheric gases, particularly CO₂ and water vapor. It is, of course, also necessary to control air temperature and the energy and hydrological balances in studies of global change.

   It must be recognized that it is not possible to obtain ideal simulation of the atmospheric environments of open ecosystems. Rather, the goal is to apply the best technology and treatment scenarios available to obtain a degree of control sufficient to interpret the potential effects of global change.

   Ecosystem site selection will be dependent upon the availability of adequate technology to reach desired levels of control for each system. Only a few ecosystems have been studied and only three have been experimentally manipulated for the CO₂ variable. The only systems analyzed to date are all graminoid in structure. Oechel and colleagues (Oechel and Strain, 1985) enriched tundra plots for three years (1983-85), Drake et al. (1989) completed four years of treatment of representative plots of estuarian salt marsh (1987-89), and Owensby controlled tall-grass prairie at his site in Kansas in 1989 and 1990. Some autecological studies have been accomplished in the other ecosystems shown in cross-hatching in Figure 15.1, but system-level data exist only for graminoid communities. It will be necessary to obtain information on more representative systems from the matrix shown in Figure 15.1 if we are to develop a predictive capability for the effects of long-term global change on terrestrial ecosystems.

   When ecosystems are chosen for large-scale study of global change, they should represent the range of diversity indicated in Figure 15.1. It is possible to classify broadly the world's ecosystems into six biomes. Ranked roughly in order of predicted sensitivity to the Manabe Scenario (Manabe and Wetherald,1980) of global change (CO₂ increase, global warming with increasing temperature from equator to high latitudes, mid-latitude drying) they are: (1) boreal forest, (2) tundra, (3) tropical forest, (4) temperate forest, (5) grass- land, (6) desert.

   Ecosystems will be affected by as much as 11°C mean annual warming; (2) the middle latitudes will have moderate warming but significant drying in the zones of descending atmospheric circulation; and (3) in spite of the prediction that the tropical zone will be least sensitive to weather change, differential species responses to carbon dioxide fertilization will generate important ecosystem modification in the systems of highest biological diversity.

   In the past, field researchers examined effects of environmental change using animal enclosures or exclosures, irrigation and fertilization of field plots, automatic or manual rain covers, and simple plastic tents with various degrees of closure. Although these systems have allowed manipulation of some environmental factors (e.g. water, nutrients, herbivores), atmospheric gases could not be controlled. Here we examine various technologies that are available to control atmospheric gases. Current technologies being used for soil and air temperature, water and nutrient modification in ecosystem studies could be applied in the atmospheric control facilities.

   Facilities that have been used to control atmospheric gases (e.g. toxic air pollutants, CO₂, water vapor), include leaf chambers, sunlit controlled environment facilities, mobile greenhouses, larger more permanent green-houses, and open-top chambers. A system of gas lines that deliver CO₂ for release over open fields was described by Allen (1975), and McLeod et al. (1983) described a circular plot of stand-pipes for the release of toxic air pollutants. Hendrey and colleagues from the Brookhaven National Laboratory, USA, are refining McLeod's system for the continuous release of CO₂ over an agricultural field. These systems will be described in more detail in the following pages.