Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
11 For biologic markers to be accepted as indicators of the effects of specific air pollutants, a relationship must be shown to exist between the pollutants (causes) and the responses (effects) measured by the markers. In approaching its task of evaluating the potential usefulness of various markers of air-pollution effects, the committee found it necessary to consider several fundamental criteria for the establishment of cause and effect in complex environmental relationships. This section reviews those criteria and suggests guidelines for their application to biologic markers of air- pollutant stress and damage in forests. Air pollution affects forests and trees in many ways. Smith ( 1981 ) classified interactions between air pollutants and forests into three categories. When pollutant concentrations are low (Class I), forest vegetation and soil serve both as sinks and sources of pollutants. At intermediate concentrations (Class II), vegetation is subtly and adversely affected. Acute effects on forests, including morbidity and mortality, occur at high pollutant concentrations (Class III). Gradations between classes result in a continuum of effects on forests, occurring from the subcellular to the ecosystem levels. Chronic stresses can induce a series of changes (including species impoverish- ment) that are as systematic as plant succession, although less well recognized. Of the many examples of the effects of chronic stress, some of the best defined were shown in experimental studies of the ecologic effects of ionizing radiation (McCormick, 1963; Woodwell, 1970; Fraley, 1971; Woodwell and Houghton, in press), in which the primary effect was well defined and measurable, whereas secondary effects, such as insect damage, were clearly recognized as secondary. In forest stands affected by regional pollution, where effects of several stresses might be integrated, cause-and-effect relationships are not as clear as in the case of ionizing radiation. The challenge of relating specific causes to specific effects is complicated further by the similarity of patterns of forest changes caused by a wide variety of toxicants to changes caused by other stresses. Experimental studies have been designed to investigate the effects of specific pollutants or combinations of pollutants on forests or trees and to test whether a particular pollutant might be a factor in damage observed in an existing stand or single tree. However, in a forested region subjected to a variety of air pollutants and natural stresses, the effects of each stress are not necessarily unique and identifiable. Furthermore, simple field studies of affected forest stands cannot yield clear information about the specific causes of damage--only the general conclusion that the observed pollutants appear to be associated with the observed damage. The damage might eventually be noticed as changes in the distribution of species, rather than immediately as specific physiologic symptoms in individual plants. If an affected species were an ephemeral or an occasional participant in the community, such changes would be particularly difficult to recognize and attribute to a cause. Knowledge of the structure and physiology of forests and trees is now sufficient to develop a basis for detecting disruption or disturbance from a variety of causes. Although the task is complicated and efforts are incomplete, the committee's review suggested some answers to the questions posed during the workshop. The workshop papers confirmed that many biologic markers of stress are known for trees and forests, but that few (if any) such markers are unequivocal indicators of single causal agents. Because many natural and anthropogenic stress factors (summarized in Table 1 ) can combine to cause changes in the trends of forest metabolism, growth, and mortality (summarized in Table 2), it is difficult to link causes and effects with certainty. To help to clarify the links, the committee suggests a set of guidelines for inferring cause-and-effect relationships in cases of local and regional changes in the health and growth of trees
12 and forests. As in much of science, the relationships will be expressed as probabilities based on current information and experience, not as certainties. A related problem was addressed by bacteriologist Robert Koch in the nineteenth century, when the germ theory of disease was gaining wide acceptance (Koch, 1876~. Koch pointed to the need for systematic and rigorous proof that a disease was caused by a particular organism. To establish the cause-and-effect relationship, Koch advanced a set of postulates that now seem obvious. His postulates, summarized by Yerushalmy and Palmer (1959), were as follows: 1. The [causative] organism must be found in every case of the disease. 2. [The causative organism] must be isolated from patients and grown in pure culture. 3. When the pure culture is introduced into a susceptible subject, it must produce the disease. These postulates were formulated to deal with bacterial diseases and are widely accepted in the biomedical sciences. They clearly are inadequate to deal with diseases that have several etiologic components or to assess cause-and-effect relationships between tree damage and environmental pollutants. But they can be used to focus thinking about important questions and can help to rule out particular pathogens as causes. Hill (1965), Mosteller and Tukey (1977), Cochran (1983), and Holland (1986) have described various criteria for establishing cause-and-effect relationships. As Holland pointed out, it is necessary that, "for causal inference, each unit [e.g., each tree, each stand, etc.] be potentially exposable to any one of the causes. This committee sought criteria that might help to define the most probable causes of stress symptoms seen in nature. On the basis of a few common principles, the criteria for establishing causality that appeared most useful for assessing pollutant effects on forests are the following: · Strong correlation. Is there a consistent relationship between the measured effect and the suspected causers)? This criterion includes the concept of generality of association or consistency (Koch first postulate) and the concept of strength of correlation. Mosteller and Tukey (1977) wrote about "a clear and consistent associa- tion." · Plausibility or mechanism. Is there a biologic explanation of the mechanism of the observed association that is reasonable, coherent, or analogous to another case that is understood? Does it appear to contradict other, known mechanisms? Hill (1965) cautioned that "what is biologically plausible depends on the biological knowledge of the day."
13 Table 1. Environmental Stress Factors Known to Affect Forests and Trees Competition for resources: Growing space Nutrients Sunlight Water Weather and climate: Temperature extremes Drought High winds Low humidity High altitude with high ultraviolet radiation Heavy loads of ice or snow Wild fire Biologic agents: . - . rungs Insects Nematodes Bacteria Viruses and related forms Parasitic plants Predators Lack of essential symbionts Chemical factors: Deficiencies or imbalances in essential nutrients Toxic elements Allelopathic chemicals Herbicides Air pollutants: Toxic gases Toxic aerosols Growth-altering chemicals Acid deposition leading to direct injury Atmospheric deposition of toxic particulates Human disturbances: Mechanical injury to individual trees Clearing of forests Burning of forests Physical disturbance of soil induced by: Excessive drainage or flooding Compaction Erosion by wind or water Removal or destruction of organic matter Mechanical disturbance of soil
14 Table 2. Effects or Symptoms of Stress and Damage in Forests and Trees Effects on Individual Trees Visible symptoms of injury: Change in shape, size, color, and number of leaves Change in normal patterns of growth and development Change in normal patterns of foliage flushing or senescence Decreased annual height growth or radial increments Alteration of physiologic processes: Photosynthesis Respiration and metabolism Transpiration Mineral nutrition Transport and allocation of photosynthate Hormonal control of growth Symbiotic relationships with other organisms Changes in susceptibility to other stress factors Life-history changes: Decreased longevity; early onset of dieback Changes in reproductive behavior Effects on Forests Decreased productivity of stands Changes in age-class distribution Changes in normal patterns of competition and mortality Changes in normal patterns of community succession Changes in species composition Changes in nutrient cycling Changes in hydrologic behavior, watershed functions, or wildlife habitat Changes in genetic structure of populations Responsiveness or experimental replication. Can the effect be duplicated by Can the effect be stopped or prevented by removing the putative causal experiment? agent? Such a result is "strong causal evidence when you can find it" (Holland, 1986~. This criterion represents Koch's third postulate and includes Hill's (1965) idea of biologic gradient--i.e., dose-response relationship. · Temporality. A cause must precede its effect or at least be present at an appropriate time. Testing this criterion requires adequate histories of exposure and trends in a forest's growth and composition. It might be difficult to isolate subtle shifts in the genetic makeup of a tree population that increase the susceptibility of the population to other stresses. · Weight of evidence. The individual components of establishing cause-and-effect relationships--correlation, plausibility, responsiveness, and temporality--do not by themselves provide sufficient evidence. However, studies that provide information on
