Contemplations on Ecological Risk Assessment
THOMAS E. LOVEJOY
I approach this subject with a basic concern about the environment and from a scientist's perspective. In addition, I look at it through a biological diversity lens and with a background both in conservation biology and ecology.
What does someone concerned with the Amazon rainforest make of ecological risk assessment? It is interesting to apply the ecological risk assessment approach to the Amazon. First, the problem of Amazon deforestation is so blatant that hazard identification is obvious (loss of biological diversity, degraded landscapes, regional climate disruption, and greenhouse-gas production). If I think of my own research involvement with the effects of habitat fragmentation, the examination of the effect of scale (fragment size) on community structure and species richness could be considered a dose-response study. When the results are applied by Jim Tucker of NASA in analyzing the effects of Amazon deforestation on biological diversity, it is a form of exposure assessment.
It has been fascinating to look at the process of ecological risk assessment over a wide variety of case studies: from a single species (whether in terms of human health or oysters), through multispecies systems (the Georges Bank fishery) and ecosystems (what the spotted owl issue is really all about), to the scale of the entire biosphere includ-
ed in the purview of Bill Cooper's committee on priority-setting for EPA.
One important question that arises is, what counts as a problem? For humans narrowly, the critical end point for determining whether something is a problem is pretty clear: it is essentially a question of human health. For more complex systems there are multiple determining end points. On the one hand, there is a danger of wasting energy and time in debating what these end points might be, rather than acting on the problem. On the other hand, there is a danger of stifling constructive debate by concentrating exclusively on action. Curiously, at the planetary scale, matters seem to integrate into simplicity with clean-cut end points like stable atmospheric composition, maintenance of biological diversity, and normal levels of UVb radiation. One idea that sticks in my mind is that what happens is a scientific question and what we want is a value question. Someone else expressed the latter in another way: So what if a few robins bite the dust?
What counts as a problem—i.e., a risk—is especially complex when it comes to ecosystems. The tendency, of course, is to look primarily at ecosystem function. The rare species within an ecosystem usually play only a small role in an ecosystem's function. A focus on ecosystem function in risk assessments of ecosystems results in an assessment based on the few species that contribute the most to ecosystem function. A tropical forest, however, would require a different approach because a vast array of rare species constitute the bulk of the biomass and ecosystem function.
Focusing on ecosystem function tends to lead to a snapshot approach that overlooks how species have important functions at different times. I think, in this regard, of a yeast discovered by a graduate student at the Academy of Natural Sciences of Philadelphia. It is normally rare, apparently outcompeted by other organisms, because of an unusual metabolic pathway that skips over a few normal steps. When faced with rising mercury concentrations in its environment, the yeast is suddenly at an advantage, because the steps that its pathway skips are vulnerable to mercury compounds. Furthermore, the yeast is capable of reducing mercury compounds to their elemental form. The yeast population explodes, and its vacuoles fill with mercury, which is then deposited on rock surfaces. Mercury in the aquatic community is cleaned up, and the yeast becomes rare again. Examples of this sort might in fact be unusu-
al, but the larger point is that many rare species in ecosystems can represent infrequent but recurring conditions in which those species do play important roles.
Indeed, a focus on ecosystem function to the exclusion of biological diversity overlooks biological diversity itself as an end point—one that represents potential resources (including genes), intellectual resources (evidence of how living systems can work), and environmental indicators. The latter brings to mind the fascinating story of TBT. One wonders whether, if people did not eat oysters, anyone would have noticed or cared that TBT at only 2 parts per trillion would cause the female dog whelk to grow a penis (imposex).
Nonetheless, I cannot help noting the difference between the oyster population today, filtering a volume of water equal to the Chesapeake Bay once a year, as opposed to once a week before the major population decline. This presents the term keystone species in a new, expanded light.
The question of what is sufficient evidence for action is central in the business of ecological risk assessment. For cancer producing substances, there are, for the moment, reasonably precise working definitions that use laboratory studies of what constitutes limited and sufficient evidence. In the example of TBT or Georges Bank, the evidence is more circumstantial, although common sense suggests causality. Certainly, action should not always wait for an understanding of mechanisms of causality, especially when dose-response linkage is clear. Yet at the same time, there is real value, as pointed out by Dr. Maki, in pushing ahead with research to understand causal mechanisms. That would help, for example, in evaluating substitute compounds for TBT or CFCs. In the last analysis, assessment is an iterative process, and action and policy cannot wait forever.
Another tough question is what constitutes acceptable solutions. There are, in fact, two definitions of what is acceptable: what works scientifically and what society is willing to accept. It was clear from the fisheries discussion that fishing-fleet managers resist the notion of risk as a driver of decisions. The same must be involved in the spotted owl old-growth controversy. A point made about the latter was that, although a range of alternatives were presented, only two relative extremes entered into the debate. Whether the other alternatives might have been acceptable or not, there clearly will be times when there is
only one scientifically acceptable solution (habitat type being destroyed, etc.), and it is important to err on the safe side. As Robert May points out, some arguments do not have two sides—it is a figment of journalism and wishful thinking to believe that all arguments do. Some questions do not have two sides—at least not the two perceived on first examination. Sometimes the alternatives or solutions lie outside the envelope being considered. One wonders how the spotted owl/old-growth debate would have played out had the question of alternative means of making a living for the timber workers been a major part of the exercise early on.
It is in the area of solutions that the term uncertainty often raises its head. The term can easily be manipulated to put scientists on the defensive. Uncertainty is fundamentally part of the inherent honesty of the scientific process. Indeed, it is part of the normal way in which we discuss almost anything and is generally endemic to decision-making. Uncertainty is at least as applicable to the countervailing view of what recommended change science may put forth. We need to recognize that and not permit ourselves to be put on the defensive.
One initial talk divided ecological risk assessment into two kinds: one is essentially reactive (someone notices a problem) and involves initially retrospective and ecoepidemiological problems; the other is active and predictive. Clearly, there is a need to move toward greater emphasis on the active, but it is folly to assume that we will ever know enough to avoid surprises altogether. Clearly, too, there is a need to set priorities: it is impossible to do everything at once. There is a serious challenge, in that ecological risk assessment is vastly more complicated than society will ever understand, and at least in principle, society wants it all done. That puts a tremendous premium on sound risk assessment, on communication, and on priority-setting. EPA's effort to set priorities, giving great emphasis to problems of substantial spatial and time dimensions, is an extremely important exercise—essentially ecological risk assessment on a planetary scale.
Early in the workshop, I participated in a discussion of the role of science vs. the role of values in setting the agenda. A devil's advocate suggested that the role of science was to work on problems once values were set. There is, of course, some truth in that: if society wants something, science responds. But in setting those values, society must be informed by science. The new priorities set by EPA's Science Advi-