May/June 2001

Precautionary Principle: Applications to Seismic Hazard Analysis

In the environmental field, the Precautionary Principle has recently been promoted as a guiding principle for decision making. Statements of the Precautionary Principle are varied, so my first task in this discussion is to provide a satisfactory definition. A starting point is the statement from the Wingspread conference in Racine, Wisconsin in 1998. The central statement of this version is: "When an activity raises threats of harm to human health or the environment, precautionary measures should be taken even if some cause and effect relationships are not fully established scientifically. In this context the proponent of an activity, rather than the public, should bear the burden of proof." I would restate the Precautionary Principle for seismic hazard analysis as: "Where earthquakes threaten harm to human safety and the built environment, precautionary measures should be taken even when the hazard is not fully established scientifically."

The Wingspread statement urges, "Corporations, government entities, organizations, communities, scientists and other individuals must adopt a precautionary approach to all human endeavors." The statement emphasizes that its application should be an open process: "The process of applying the Precautionary Principle must be open, informed and democratic and must include potentially affected parties." Organizations such as the Nevada Seismological Laboratory or the U.S. Geological Survey would have a particular responsibility, if this were accepted to guide our philosophical approach toward seismic hazard analysis, to assure openness in the process. Openness is especially important because we, as seismologists, have our own biases, which can at times be seen as self-serving. The informed discussion among affected parties is needed for balance.

The Precautionary Principle, as an element of environmental protection, has been used in several international treaties and declarations. One of the broadest is the Rio Declaration on Environment and Development, resulting from the United Nations Conference on Environment and Development in Rio de Janeiro, Brazil, 3-14 June 1992. Principle 15 states: "In order to protect the environment, the precautionary approach shall be widely applied by States according to their capability. Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation." The Rio Declaration describes consensus principles adopted during the United Nations Conference on Environment and Development, but it lacks the status of a formal treaty. A Communication from the Commission of the European Communities in 2000 states that "[a]pplying the precautionary principle is a key tenet of its policy", particularly for risk management. The Commission of the European Communities considers that the Precautionary Principle applies quite broadly, as indicated from the following: "Although the precautionary principle is not explicitly mentioned in the [European Community] Treaty except in the environmental field, its scope is far wider and covers those specific circumstances where scientific evidence is insufficient, inconclusive or uncertain and there are indications through preliminary objective scientific evaluation that there are reasonable grounds for concern that the potentially dangerous effects on the environment, human, animal or plant health may be inconsistent with the chosen level of protection."

The Precautionary Principle has been used, at least in some cases, as a philosophical guide for environmental policy in the United States. In a speech at the National Academy of Sciences on 13 October 2000, reflecting on state actionsfor coastal protection, Christine Todd Whitman, Governor of New Jersey at the time and now Director of the Environmental Protection Agency under President George W. Bush, said: "As states like New Jersey enter a new era of public policy, with science playing such an important role, I believe policymakers need to take a precautionary approach to environmental protection. By this I mean we must 1) acknowledge that uncertainty is inherent in managing natural resources, 2) recognize it is usually easier to prevent environmental damage than to repair it later, and 3) shift the burden of proof away from those advocating protection toward those proposing an action that may be harmful."

How would the Precautionary Principle apply to seismic hazard analysis? The emphasis in probabilistic seismic hazard analysis for "ordinary" structures is on finding the best estimate of the hazard. To account for epistemic uncertainty, multiple models may be considered for the seismicity and the ground motions, although in some cases a single "best estimate" input model may still be used. The multiple models are weighted according to informed judgments of the probabilities that they are correct, in order to achieve the best estimate of the hazard. The precautionary approach could result in an open decision to favor those credible models that give the higher hazard estimates. This approach would recognize that it is easier to build a structure that is stronger than required by the current best estimates of the hazard (but matching high estimates) than it is to retrofit in the future if more studies demonstrate that the high hazard model is correct.

It should be stated that I believe that for the most critical facilities, where designs are likely to be based on very low probabilities of occurrence estimated in special studies, the design and regulatory approval process already assures that a precautionary approach is taken. My concern is for "ordinary" structures regulated by building codes. Multiple models will certainly be considered in any state-of-the-art probabilistic seismic hazard analysis for a critical structure, and at low probabilities the models that give a higher hazard have a sufficient effect on the mean.

I will briefly describe three examples of how the Precautionary Principle might be applied in a probabilistic seismic hazard analysis. First, in western Nevada, results from global positioning system measurements of crustal deformation indicate that the region is deforming by about 5 ± 2 mm/yr within a zone about 100 km wide on the east side of the Sierra Nevada crest, but active faults that express that motion have not yet been found and historical seismicity is too low to achieve such a rate. The precautionary approach would be to develop an input seismicity model that is consistent with the full 5 mm/yr through the region, and faults in the zone with unknown slip rates may be assumed to have slip rates consistent with the geodesy. Second, the existence of precariously balanced rocks at many locations in southern California and Nevada suggests that the PSHA at those sites, and consequently for the region, is overestimated at extremely low probabilities. By the Precautionary Principle, this information would not be incorporated into the PSHA (as an alternative hypothesis to the present approach) until strong confirmatory evidence has been accepted. Third, measured ground motions at high frequencies from the 1999 Chi-Chi and Izmit earthquakes are lower than extrapolations from smaller earthquakes, and lower than a very limited number of observations from earthquakes that were almost as large. There are differing explanations. At the 2000 Fall American Geophysical Union meeting, Paul Somerville suggested that low accelerations may be associated with faults causing surface rupture, while I suggested that faults with large amounts of total geological slip, high geological slip rates, or rupturing through softer rocks may favor lower accelerations. By the Precautionary Principle, the 1999 observations, based on the statistics of only one thrust and one strike-slip earthquake, would not be used to lower the mean hazards at high frequencies until the conditions associated with the low accelerations are understood and the predominance of those conditions is confirmed.

If the precautionary approach is taken, the seismological community is forced to deal with some very difficult questions associated with the threshold of plausibility for a hypothesized seismic hazard. Compounding our occasional tendencies to sound alarmist, its application could reopen arguments that probabilistic seismic hazard analysis hoped to solve, such as judging how much evidence is needed to consider a particular hypothesis viable. In addressing this, we can apply the guidelines for application of the Precautionary Principle that were adopted by the Commission of the European Communities, described in a Communication issued in 2000. This Communication ties the decision process to a careful evaluation of available scientific evidence. According to these guidelines, "[r]ecourse to the precautionary principle presupposes that potentially dangerous effects deriving from a phenomenon, product or process have been identified, and that scientific evaluation does not allow the risk to be determined with sufficient certainty." The potentially dangerous effects of earthquakes are of course beyond debate, and circumstances where the hazard has not been determined with sufficient certainty are abundant (e.g., those mentioned above), so these criteria for invoking the Precautionary Principle are satisfied. When the Precautionary Principle is invoked, the Communication from the Commission of the European Communities states that "implementation ... should start with a scientific evaluation, as complete as possible, and where possible, identifying at each stage the degree of scientific uncertainty" and must satisfy the following criteria: proportionality, nondiscrimination, consistency, examination of costs and benefits, subject to review, and capability of assigning responsibility for gathering scientific evidence for a more comprehensive hazard assessment:

Proportionality, which "means tailoring measures to the chosen level of protection", and consistency, which "means that measures should be of comparable scope and nature to those already taken in equivalent areas in which all scientific data are available", are standard features of present probabilistic seismic hazard analyses. Assignment of 5 mm/yr to the deformation near Reno will automatically assure that Reno is assigned a hazard similar to other regions with similar fault activity rates. Nondiscrimination means "comparable situations should not be treated differently." This can be assured by establishing general rules for applying the Precautionary Principle in seismic hazard analyses; a simple one (which would need to be debated) could be to always choose the larger of the deformation rates determined from geodesy or geological observations. The examination of costs and benefits is potentially the least straightforward of these criteria. The output of a probabilistic seismic hazard analysis allows the decision maker to develop a cost-benefit analysis of different performance levels considering the probabilities of the seismic hazards. However, as described by the Communication from the Commission of the European Communities, the examination of costs and benefits would involve a higher-level comparison of the costs (more expensive construction) and benefits (better earthquake performance) of invoking the Precautionary Principle in the first place. The Communication envisions an examination that includes noneconomic considerations and explicitly states that "protection of health takes precedence over economic considerations." Thus, in the case of PSHA, this criterion does not need to be a stumbling block. Any decision would be subject to review in the periodic revisions of the national hazard maps planned by the U.S. Geological Survey. The capability of assigning responsibility for gathering scientific evidence for a more comprehensive hazard assessment is an important aspect, as any hypothesized seismic source which cannot be tested by gathering more scientific evidence is likely to lack the status to be considered a credible hypothesis. The decision to invoke the Precautionary Principle, when associated with specific questions about the input to the PSHA, would simultaneously define the problem causing the specific uncertainty and could thus focus attention on a problem of interest for the research and public-service scientific communities.

Why should we adopt the Precautionary Principle? The nature of the Building Code as it is applied in the seismically active western states is changing fundamentally. The Uniform Building Code for the later half of the 20th century used broad zones-Zone 1, Zone 2a or 2b, Zone 3, and Zone 4-for code-based earthquake-resistant design criteria. For instance, California near the San Andreas system is entirely in Zone 4. In contrast, the International Building Code 2000 directly adopts the probabilistic seismic hazard models that were developed by Art Frankel and his colleagues at the U.S. Geological Survey, with some modifications only near the most active faults, as the basis of seismic-resistant design. The engineer reads ground motions from the contours on the map as the starting point for the hazard description. This change in philosophy means that the engineers are using maps with far more detail than before (and in some areas, perhaps, more detail than is justified on the basis of our current knowledge). Consequently, design criteria are much more sensitive to the details of the latest model for seismic hazards: the mapping and slip rates on active faults, the smoothed historical seismicity, and the ground motion prediction models that depend on what is still a very limited set of strong-motion records from sites close to large earthquakes. In this context, an open and intentional decision to adopt the more hazardous of the credible models may provide a seismic-resistant design policy with greater stability. This is appropriate since discoveries and improved understanding of the seismic hazard are certain to be coming at a steady or accelerating rate, especially with the first results of GPS surveys now being reported, the initiation of the Advanced National Seismic System, and the likely deployment of the Plate Boundary Observatory.

In summary, at least until our understanding of the seismic hazard stabilizes, a precautionary approach is the best approach that the seismological community can adopt for dealing with the uncertainties in the input to the seismic hazard analysis and for providing the architecture and engineering communities with a relatively stable, but still defensible, estimate of the seismic hazard.

Some References on the Precautionary Principle

Because some of the readers of Seismological Research Letters are probably not well acquainted with the literature on the Precautionary Principle, the following is a short, annotated list of references. For convenience, those without copyright are reproduced at John Anderson's Web pages: http://www.seismo.unr.edu.

Appell, D. R. (2001). The new uncertainty principle, Scientific American, January 2001, 18-19.
    This was the first article on the Precautionary Principle that caught my attention and led me into the literature on the topic. Comments in this article are the reason for my selection of the Wingspread statement as a starting point for the presentation of the Precautionary Principle.

Commission of the European Communities (2000). Communication from the Commission on the Precautionary Principle, COM (2000) - 1, Brussels, Belgium.
    This Communication is important because of its official source. Its emphasis on the scientific process that needs to accompany any discussion of whether to apply the Precautionary Principle is, in my viewpoint, highly appropriate for the earthquake community.

Foster, K. R., P. Vecchia, and M. H. Recacholi (2000). Science and the Precautionary Principle, Science 288, 979-981.
    This article discusses the Precautionary Principle in the context of its application to regulation of human exposure to radio-frequency radiation. It has a good, compact discussion of the range of definitions of the Precautionary Principle.

Montague, P. (1998). The Precautionary Principle, Rachel's Environment and Health Weekly #586, 19 February 1998. See http://www.rachel.org/home_eng.htm. (Go to "rachels" and then to "back issues")
    This Web page, as far as I know, is the primary source for the Wingspread Statement of the Precautionary Principle. This statement of the Precautionary Principle has several useful concepts but leans away from a process driven by science toward one where citizen concerns are emphasized.

United Nations (1992). Rio Declaration on Environment and Development, 13 June 1992 (UN Doc./CONF.151/5/ Rev.1). See http://sedac.ciesin.org/pidb/texts/ rio.declaration.1992.html.
    This accompanied a relatively important global conference. The United Nations Conference on Environment and Development (UNCED), held from 3-14 June 1992 in Rio de Janeiro, Brazil, brought together more than 150 nations, 1,400 nongovernmental organizations (NGO's), and 8,000 journalists. A discussion of the conference itself can be read in Haas, P., M. Levy, and T. Parson (1992), Appraising the Earth Summit: How should we judge UNCED's success?, Environment 34, 6-11, 26-33. The declaration is not a treaty, but is still important for its statement of principles.

Whitman, C. T. (2000). Remarks of Governor Christine Todd Whitman, "Effective Policy Making: The Role of Good Science" Symposium on Nutrient Over-Enrichment of Coastal Waters, Washington, DC, Friday, 13 October 2000. This speech is available on the Web at http://gos.sbc.edu/w/whitman2.html.
    This is interesting as a statement by a leading politician. She appears to endorse the importance of good science and the use of the Precautionary Principle when the scientists don't yet know the answer. The above Web page still worked in March 2001, even though Ms. Whitman was no longer Governor of New Jersey.

John G. Anderson
Seismological Laboratory
University of Nevada
Reno, Nevada 89557

To send a letter to the editor regarding this opinion or to write your own opinion, contact Editor John Ebel by e-mail or telephone him at (617) 552-8300.

Posted: 16 June 2001
Link Updated: 23 January 2002