- Top of page
- ANALYSIS OF TWO PAPERS
- NORMATIVE SCIENCE
In the July 2006 issue of Integrated Environmental Assessment and Management, a paper by Harwell and Gentile was published assessing the present ecological significance of the impacts from the Exxon Valdez oil spill (EVOS). First, this paper compares the major conclusions of Harwell and Gentile and a paper reviewing the current impacts of EVOS by Peterson et al. as published by Science in 2003. Stark differences exist between the conclusions of the 2 papers regarding continuing impacts. Part of the difference appears to be the infusion of different social values or policy goals into each. Normative science is the use or interpretation of data in support of specific values or policies. Examples of values or policies intertwined with science are constructs such as ecosystem health, ecosystem integrity, ecological significance, and recovery. Examination of the environmental risk assessment and toxicology literature reveals that the symptoms of normative science are common and the implications widespread. Separation of science from policy or at a minimum a transparent acknowledgment of the science–policy interaction is clearly necessary in order to obtain a clear picture of the ecological system under investigation.
ANALYSIS OF TWO PAPERS
- Top of page
- ANALYSIS OF TWO PAPERS
- NORMATIVE SCIENCE
The recent publication by Harwell and Gentile (2006) revisits the persistence of impacts of the 1989 Exxon Valdez oil spill (EVOS). The impacts of the spill and their persistence are controversial. This is not just a scientific debate, as the controversy also occurs in the public domain as demonstrated by a recent article in Time magazine (Caplan 2006). This short commentary compares Harwell and Gentile (2006) to the widely cited Science article of Peterson et al. (2003) and other developments. The context of the comparison is the insertion of normative terminology and frameworks into environmental toxicology, risk assessment, and environmental management as illuminated by Lackey (2004).
Peterson et al. (2003) clearly state that persistent oil and chronic exposures continue to cause impacts to wildlife. The assessment model that they use is a derivation of a before– after control–impact (BACI) as described by Parker and Wiens (2005). Peterson et al. also state that indirect effects and the resulting cascades within the marine community have prevented recovery of the system. Table 1 of the article (page 2085) summarizes the comparison between the old and new paradigms for oil toxicology for marine ecological systems.
The new paradigm is comprised of an assumption of increased persistence of the oil, a role for chronic effects on organisms, and cascades of indirect effects for populations and communities. The lack of inclusion of indirect effects was stated as one of the weaknesses of ecological risk assessment in the estimation of resource loss.
In contrast, Harwell and Gentile (2006) dismiss the importance of indirect effects and cascades due to the EVOS event. They state,
We believe that the primary drivers for cascading effects in the PWS ecosystem relate to fundamental processes driven by climatic and oceanographic variability, especially involving nutrient upwelling and fish productivity (p 231).
The contrast is made more explicit by Harwell and Gentile later in the paper:
Some would argue that so-far-unrealized indirect or cascading effects from EVOS could in the future plague PWS (such as suggested by Peterson et al. 2003). We quite disagree: the very real cascading effects on the PWS ecosystem over the past several decades from fundamental natural stressors, especially climate and oceanographic variability, and the pervasive anthropogenic stressor of overexploitation of marine resources are significant and would overwhelm any proposed EVOS signal (p 238).
Peterson et al. (2003) identify instances of population-level effects as additional support for the oil toxicity hypothesis. Harwell and Gentile do not cite a source for the contention in the above quote that cascading effects are due only to larger-scale stressors. Conversely, Peterson et al. do not eliminate regional scale stressors as a causal factor for some of the observed effects. Clearly there is an important difference in how the 2 studies regard indirect effects and cascades.
Many other similar contrasts exist between Peterson et al. (2003) and Harwell and Gentile. However, many of the differences in the interpretation of the available information may be tied to a more fundamental issue. The defining of “effect” in each paper is quite different and is a more important contrast. Understanding this contrast is critical to the fields of environmental toxicology, risk assessment, and management.
Harwell and Gentile state that an effect with “ecological significance” includes 2 factors: 1) Whether a change in a valued ecosystem component (VEC) is of sufficient type, intensity, extent, and/or duration to be important to the structure, function, and/or health of the system and 2) whether a change in a VEC is sufficient to exceed natural variability or alters the innate variability of the system. The 1st criterion incorporates words such as “important” and “health,” terms that are inherently value driven. The 2nd criteria discusses natural variability, but over what time frame? Neither criterion discusses spatial or temporal scales, begging the point about whether we are talking impacts over all of Prince William Sound and variability over time frames of 1, 10, 100 y or since the last glaciation. Both criteria have underlying and innate value systems that are being applied by the authors.
“Recovery” is the analogous term used in Peterson et al. (2003). “Recovery” is not defined within the paper, but the use of the term appears to be synonymous with return to conditions prior to the EVOS or that no statistically significant difference exists between oiled and not-oiled sites. This is the BACI model as defined by Parker and Wiens (2005). Recovery assumes an equilibrium state, but that is now recognized as a poor model for describing the dynamics of ecological structures or environmental impacts (Wu and Loucks 1995; Kapustka and Landis 1998). How alike is alike? Are there criteria for specifying the statistical power of a field study to define the level of population or reproductive performance compared to pre-event levels? Determining sufficient statistical power or how close measurements have to be to be similar enough becomes a trade off between resources and value judgments on how close is close enough.
Note that in the definitions of “ecological significance” and “recovery,” there are societal values incorporated into each. In the Harwell and Gentile (2006) definition of ecological significance, it is straightforward to identify the value-laden terminology (“important,” “ecosystem health”). The term “recovery” has value-driven assumptions that are implied in the assumptions about equilibrium. It is not realistic to expect any person or group to completely disassociate themselves from the cultural value systems imposed by religion, political affiliation, ethnic background, or an employer's value system and mission. However, the onus is on the authors in reporting results to lay that context aside. As the fundamental definitions of each paper are driven by different sets of value systems, both papers have at least some of the definitive characteristics of normative science.
- Top of page
- ANALYSIS OF TWO PAPERS
- NORMATIVE SCIENCE
How can such divergent conclusions about the effects of the EVOS be reached by reviews using information from the same site? R. Lackey, in a series of papers (Lackey 2001, 2004, 2006), has investigated the interaction between values, policy, and the use of science in environmental management. He coined the phrase “normative science” (Lackey 2001). Normative science is described as science that is developed, presented, or analyzed on the basis of an assumed policy or outcome. One of the best examples presented by Lackey of policy woven into science is the term “ecosystem health” and the related term “integrity.” His analysis clearly demonstrates that an underlying policy or value statement of ecosystem health is that it is a specified system even as modified by human activity. Ecosystem health is a policy statement, not a scientific one. The term “ecosystem integrity” also has policy implications. Ecosystem integrity is defined as a system as it would exist without human interference, often labeled as pristine, reference, or baseline conditions (Lackey 2001). The assumption underlying integrity is that there is a specified state for an ecological system given its location, bioregion, and so on. Ecosystem health and integrity are Clementian ecology based on assumptions of stability and specificity that have been found inadequate in describing the properties of ecological structures (Wu and Loucks 1995; Kapustka and Landis 1998; Wu and David 2002). Since ecosystem health is a policy statement and ecosystem integrity is an outmoded idea, a definition of ecological significance or recovery that invokes these constructs is a policy statement.
The use of value-laden terms such as “ecological significance” or “ecologically significant” are widely used in the literature. In a search of Environmental Toxicology and Chemistry and IEAM from June 2001 to June 2006, 75 papers were found to use those terms at least once. The same search parameters for the journal Environmental Science and Technology resulted in 50 papers being identified. Clearly the terminology is widespread in the scientific literature. A review of the citations reveals that the vast majority of the citations are not policy statements but purport to be scientific investigations.
In environmental science, separating what is science from what is a policy is difficult. Ecological significance as defined by Harwell and Gentile (2006) is based on a policy statement, as long as the range of the impact is within the range of natural variation that the impact is acceptable. “Recovery” is the analogous term in Peterson et al. (2003). “Recovery” is not defined within the paper, but the use of the term appears to be synonymous with return to conditions prior to the EVOS or that no statistically significant difference exists between oiled and not-oiled sites. There are a number of definitions for “recovery” (reviewed in Parker and Wiens 2005); the decision of which to use is a policy decision.
Has our research as a community of scientists been corrupted by the advocacy of specific values by the use of terms such as “ecological significance” or “recovery”? Lackey (2006), in his “ecological policy axiom 5,” states that many advocates cloak their arguments as science to hide policy preferences. Has the use of terminology without a scientific root led practitioners to advocate consciously or unknowingly for a particular policy outcome? Has the legacy of Silent Spring (Carson 1962), an excellent example of persuasive writing and policy promotion, led to a scientific culture trying to emulate that style?
So what are the alternatives? Vaccination? I concur with Lackey (2001) that there are 2. The 1st is to cease the use of normative science and to relegate terms such as “ecological significance,” “integrity,” and “recovery” (in its Clementian context) to the dustbin; describe clearly what the purpose and findings of the study are; and communicate the state of the ecological system and the ongoing dynamics. Statements such as “We believe ” (see Hartwell and Gentile 2006) are the province of items or paradigms taken on faith rather than science. The 2nd alternative is to understand that ecological policy is a complex and multicomponent decision-making process that cannot be summarized in metaphorical clichés such as “ecosystem health,” “ecological significance,” “integrity,” or “recovery.” If a policy goal is being discussed or supported, it should be clearly defined and obvious to the reader.
One of my favorite men of science is T.H. Huxley, one of the instigators of connecting science to environmental management. In his time, religion, societal values, and science were particularly difficult to distinguish. One of my favorite quotes attributable to him is “God give me strength to face a fact though it slay me” or, as I would add, regardless of the policy of my employer, family, colleagues, or funding source.
In a journal such as IEAM, with a mixture of science, policy, and commentary, the differentiation is even more important. The outcome for me of this evaluation is an appreciation for just how important it is for this journal and others to ensure that the review process is adequate in differentiating policy from science and the caveat regarding normative science. Would it be ironic if controversies, such as that illustrated above for the impacts of the EVOS, are really policy debates, not a controversy of scientific merit?