Click here for the PDF version

TRANSATLANTIC UNCERTAINTY COLLOQUIUM (TAUC):
MAKING ENVIRONMENTAL DECISIONS IN SPITE OF UNCERTAINTY

M B Beck, K H Reckhow, and K A Oye

1          OBJECTIVES 

Problem Context

Making a decision under uncertainty is ubiquitous. Its formal study has been extensive and intensive, for over several decades (Berger, 1985; Morgan and Henrion, 1990; Funtowicz and Ravetz, 1990; Freeman, 1993; Pratt et al, 1995; Clemen, 1996; Cullen and Frey, 1999; Reichert and Borsuk, 2005; Krayer von Krauss et al, 2005). While labels may differ, three broadly different categories of uncertainty are generally recognized for the archetypal outcome of an uncertain (future) event: (i) Statistical Uncertainty, in which the possible outcomes of the event are known, as well as the numerical probabilities of occurrence of each; (ii) Scenario Uncertainty, where the outcomes are known, but not all of their respective probabilities; and (iii) Ignorance, where neither all of the outcomes, nor (self-evidently) their respective probabilities, are known. We might loosely associate aleatory uncertainty with the first of these, while epistemic uncertainty would tend to be better aligned with the last. Studies on how environmental regulatory decisions are made in the face of epistemic uncertainty are sparse indeed.

The social (and cultural) context of making environmental decisions entails rather different kinds of uncertainty. Members of the different social solidarities of Cultural Theory (Thompson et al, 1990), for example, will generally hold quite heterogeneous views on the consequences of action. Broad categories of stakeholder outlooks on the man-environment relationship — the solidarities of Cultural Theory (Thompson et al, 1990; Thompson, 1997) — map across to Holling’s “Myths of Nature”, as drawn from Systems Ecology (see, for example, Holling et al, 2002). Thus, the entrepreneurial “market-oriented”, or “individualist”, stakeholder understands nature as being benign, able to recover stability in its functions following all manner of largely unfettered insults and injury; the “government-oriented”, or “hierarchist”, stakeholder sees nature as generally tolerant, but perverse on occasion and, therefore, in need of curbs on man’s activities; while a member of the “civil-society, non-governmental activist”, or “egalitarian”, social solidarity views the behavior of nature as ephemeral, likely to be plunged into disaster and radical, undesirable change at the least interference from man and his technological innovations. In a healthy democracy, such a diversity of aspirations for the future is to be celebrated. Stakeholders will embrace attitudes towards risk that are more variegated than those formally accounted for in utility theory. Consensus, as the mechanism for negotiating the outcome of the rule-making process, is unlikely to succeed, or be maximally effective.

Arguing that “a commitment to democracy does not demand the attainment of consensus over policy decisions”, Coglianese (2001) has pointed to imprecision and ambiguity as one of several pitfalls of the device. Evidence indicates it does not diminish the rate of legal challenges to rulings. Other modes of engaging the public’s participation in negotiating regulatory decisions may be just as effective as the device of consensus-seeking, if not more so. But they leave uncertain the level of dissatisfaction with the outcome at which stakeholders might challenge the rule once expressed (Coglianese, 2001). Even when policy formulation does not rely literally on the attainment of consensus — as is true under most US regulations written under the Administrative Procedures Act — policy-makers are well advised to achieve broad credibility for making sound decisions. An examination of generally credible organizations, e.g., the National Research Council and the Health Effects Institute in the US, shows the crucial importance of demonstrating that the issues have not been pre-determined before the analysis is performed, that key parts of the assessment process are open to all interested groups, and that the selection of analysts is made with an eye toward scientific standing, not political placement (McCray, 2004).

We live and work in times of abundant global communication. Within the space of one month in 2003 a Workshop was held in Rockville, Maryland, on Uncertainty, Sensitivity, and Parameter Estimation for Multimedia Environmental Modeling (Nicholson et al, 2004), while another Workshop — of the European Union (EU) Harmoni-CA project (Modeling in Support of the EU Water Framework Directive; www.Harmoni-CA.info) — was held at the Royal Netherlands Academy of Arts and Sciences, Amsterdam. Scientists and engineers on the two sides of the Atlantic were proceeding on parallel fronts, notably in respect of the computational analysis of uncertainty and the assurance of quality in models used in a regulatory decision-making situation, yet seemingly unaware of each others’ work on these subjects. In 2004 a project on Adaptive Implementation of the Total Maximum Dail Load (TMDL) program under the US Clean Water Act was begun (Reckhow et al, 2004); in 2005 a project on Adaptive Approaches to Integrated Water Resources Management, under the EU Water Framework Directive, was also begun (NeWater; Pahl-Wostl and Kabat, 2004). Both seek to address issues of uncertainty in a management context; yet there are no formal links whatsoever between the two projects.

There is much to be gained from exploiting the resulting variety, and therefore richness, of traditions in handling uncertainty on the two sides of the Atlantic and then channeling this previous experience onto enquiry addressing both epistemic uncertainty (or ignorance), negotiating procedures, and the plurality of social solidarities (Thompson et al, 1990) — beyond the conventional duality of “markets” and “hierarchies” — in the making of regulatory environmental decisions.

Objectives

The purposes of the proposed project are to:

1          Identify the key issues preventing a more appropriate, effective application of incomplete, insecure science in guiding the promulgation and assessment of regulatory actions in an increasingly open, disputatious, participatory form of environmental stewardship.

2          Establish (and launch) an ongoing, multi-disciplinary, transatlantic network of experts working at the intersection of environmental science, policy-making, and law, within the context of dealing with uncertainty, i.e., the TransAtlantic Uncertainty Colloquium (TAUC).

“Effective application” under objective (1) will cover the use of models for generating foresight, as much as procedures for handling epistemic uncertainty in a legal setting. Defining will be enquiry into the nature of uncertainty from the greatest possible variety of disciplinary backgrounds: from anthropology, law, and economics, to engineering and the natural and mathematical sciences. In contexts where there is “disputation” — precisely because of the diversity of perspectives on the man-environment relationship — TAUC and its related activities will have the longer-term objective (beyond the span of this proposal) of avoiding polarization where possible, moving instead towards debate on the basis of a shared formalism for handling uncertainty.

2          BACKGROUND

In December, 1999, the journal Nature published a specially commissioned supplement entitled Impacts of Foreseeable Science. One of its articles bears the title “Science’s New Social Contract with Society” (Gibbons, 1999). It argues we must now ensure that scientific knowledge is “socially robust”, not merely “reliable”, as in the past; and that its production be seen by society to be both transparent and participatory: 

‘[S]ocially robust’ knowledge has three aspects. First, it is valid not only inside but also outside the laboratory. Second, this validity is achieved through involving an extended group of experts, including lay ‘experts’. And third, because ‘society’ has participated in its genesis, such knowledge is less likely to be contested than that which is merely ‘reliable’.

The complete case is argued in the book Re-Thinking Science (Nowotny et al, 2001), with the subtitle “Knowledge and the Public in an Age of Uncertainty”.

This change should hardly surprise us. It is the culmination of a steady drift over the decades: away, specifically in the environmental sciences, from an accompanying management stance of command-and-control, to one of participation and democracy (Darier et al, 1999). The intellectual history of working on the science and policy of environmental management reflects the same evolution: from its genesis within Operations Research (OR), i.e., application of the scientific principles of laboratory experimentation to military and subsequently civil operations; through Applied Systems Analysis (ASA), with its focus on the singular, largely technocratic environmental decision-maker (Beck, 1997); thus to Integrated Assessment (IA), with its (re-)introduction of the human dimension and its accompanying plurality of perspectives on the man-environment relationship (van Asselt and Rotmans, 1996; Thompson, 1997). In the early 1990s, we saw the rise of Post-Normal Science (Funtowicz and Ravetz, 1993; Ravetz and Funtowicz, 1999); born of an earlier enquiry into the nature of Uncertainty and Quality in Science for Policy (Funtowicz and Ravetz, 1990); and a form of science which, when implemented in the conduct of integrated environmental assessments and decision-making, would entail the democratization of knowledge by an extension of the peer-community for quality assurance. Now we have the newly minted notion of Sustainability Science (Kates et al, 2001), which calls for a way of doing science marked by social learning and participation.

It was not the case, of course, that uncertainty was not a consideration in the earlier OR and ASA stages in the evolution of environmental stewardship, but rather that the technocrats at the helm may not have been held accountable for addressing the uncertainty suffusing the context of their proposed rule-makings and management actions.


Policy and Administration

At the most strategic of levels we see two trends: one towards the growing recognition of “uncertainty”; the other towards a greatly widening circle of persons with an acknowledged stake in scrutinizing the knowledge base and the process of making decisions on the prosperity of the man-environment relationship, including regulatory decision-making. For the time-being the participatory character of the latter is less litigious in the countries of the EU than in the US. It could well be argued, nevertheless, that the EU ought to benefit from something of what one now sees in the US of the implications of the Government Performance and Results Act (GPRA), the Data Quality Act (DQA), and their oversight from the Office of Management and Budget (OMB) of the White House. Specifically, the EPA’s Council on Regulatory Environmental Modeling (CREM) has been established and this, in turn, has set in motion the current reviews of the Council’s draft Guidance Document on the Selection and Use of Models (Pascual et al, 2004) and associated considerations of uncertainty. One such review is being conducted by the EPA’s Science Advisory Board (SAB), another by the National Research Council (NRC). Driven by much the same trends (again in the US) — that the demand for economic analysis of policy is growing, including assessment of the accuracy/uncertainty of associated economic forecasts (Harrington et al, 1999) — evaluation of models under uncertainty was a core theme in the February, 2005, Workshop of the US EPA’s Office of Water on “Water Quality Modeling for National-scale Economic Benefit Assessment”. Significantly, the driver of such economic benefit assessment is the general public’s perception of what constitutes improvements in the “swimmable”, “fishable”, or “aquatic health” status of surface water bodies, with all the uncertainty (and diversity) thereto attaching (Viscusi et al, 2004).

In the European context, the government and civil service of the Netherlands has long been a leader in developing the formal means of handling uncertainty in science for environmental policy (since the 1980s). The then “philosophical” ideas of Funtowicz and Ravetz, likewise emerging from the 1980s (Funtowicz and Ravetz, 1990), have been brought to very practical fruition in the now formally announced (January, 2005) Netherlands Environmental Assessment Agency (MNP) Guidance for Uncertainty Assessment and Communication (see van der Sluijs et al (2003), for example). Development of this (Dutch) Guidance, though international in authorship, has been fundamentally European in its origins. Little of its heritage, other than the notion of “model pedigree” (Funtowicz and Ravetz, 1990; van der Sluijs et al, 2005), which imposes a succinct matrix of judgements on the summary outcomes of peer review of the knowledge base of a domain of science, has made it across the Atlantic (pedigree surfaces in the US EPA’s CREM draft Guidance Document). The extended peer review envisaged in the (European) work on Post Normal Science is championed in the US by Sheila Jasanoff (Jasanoff, 1994). She does so essentially single-handedly, most recently in exchanges over US OMB Guidelines on peer review, and, in particular, in respect of the distinction between peer review of conventional laboratory science and peer review of science for policy and regulatory decision-making.

Two governments, those of the US and the Netherlands, have for two decades led the way in engaging formally with the issue of uncertainty in science for environmental policy, with but remarkably little engagement — on this matter — with each other. 

Important aspects of making environmental decisions in spite of uncertainty have to do with new technologies and novel substances and materials. Anyone responsible for stewardship of the environment will wish to pursue the fine line between the risks of promulgating rules and regulations that, on the one hand, leave the environment inadequately protected, or, on the other hand, impose undue burdens on society, the economy, and industry, in meeting the said regulations. Benchmark publications on the subject of risk assessment appeared in the 1980s, most notably the NRC’s so-called “Red Book” (NRC, 1983), with its UK counterparts (Royal Society, 1983, 1985). Taking stock of developments in risk assessment and management in the US over the twenty years since publication of the Red Book, Finkel (2003) argued that practice was still falling far short of the 1983 book’s standards — for acknowledging and managing model uncertainty in a transparent and consistent manner — all the valiant efforts of both the US EPA and US Occupational Safety and Health Administration (OSHA) notwithstanding. Indeed, Finkel’s paper enters into an extraordinarily detailed analysis of the logic for when and how default models (with conservative assumptions, erring on the side of being overly protective) should be replaced by other models arising out of scientific progress.

There are cultural distinctions of some complexity and subtlety to be noted here. At a recent (2004) symposium in Copenhagen, Denmark, on “Uncertainty and Precaution in Environmental Management” (UPEM I), a former Danish Minister of the Environment was heard to complain about the difficulties the “Anglo Saxons” have in embracing the precautionary principle. This was an international symposium, of central relevance to this proposal (as we shall see), yet one where participation was dominated by the Europeans, with but a small representation from North America, and almost no participants from other parts of the world. There can be worthy cultural reasons for treating UK and US attitudes as identical, but in other significant respects, they differ. In particular, there is a penchant for quantification and explicit procedural terms and conditions in the latter; in the UK, however, there has been a history of a more implicit, more verbal, less quantitative approach to agreements and definitions, including perspectives on risk assessment.

Juxtaposed thus with the Red Book of the early 1980s, Brown’s book on Environmental Threats: Perception, Analysis and Management (Brown, 1989) emerged from research supported by the UK’s Economic and Social Research Council (ESRC). One key contribution asked whether there was any difference between Engineering and Anthropology (Thompson, 1989). It argued the case against the “mathematization” of risk assessments, since this sought to re-cast Engineering as an experimental, laboratory-type enterprise. This, it asserted, was to deny the basis and origins of Engineering as a discipline distinct from Science, wherein balancing attitudes towards risk in engineering design and analysis had become “internalized” over the centuries. It is what engineers have been taught as the quintessentially distinctive, non-quantitative, process of “engineering judgement”: the matter of balancing perspectives roughly between the two social solidarities gathered around “markets” (reflecting unfettered risk-taking entrepreneurship) and “hierarchies” (from which derive procedural controls by governments for protecting society and the environment). The contemporary challenge for Engineering, Thompson wrote, was to recognize a third social solidarity — the “egalitarian” solidarity (closely associated, as already observed, with civil society actors, such as Greenpeace and Friends of the Earth) — which had risen to prominence in the closing decades of the twentieth century, altering thus radically the process of balancing differing social and individual attitudes towards risk. Understanding how to move forward, in the presence of this plurality of perspectives on the man-environment relationship, could not solely be reduced to insights gained from a quantitative fault-tree, or computational, predictive exposure analysis, according to Thompson (1989).

There is another dimension to matters of risk and technology, most readily apparent in the context of foresight and decision-making under uncertainty in respect of the prospective consequences of climate change (a cluster of Centers was established in 2004 by the NSF to address this problem). A significant response there will require large-scale changes in the system by which we produce and consume energy, in order to reduce carbon emissions to the atmosphere. Substantial uncertainties in both the costs and benefits of such technology investments (from both the public and private sectors) derive from underlying uncertainties in the likelihood of research discoveries, the ultimate cost of producing and using innovations, future interest-rate fluctuations, and so on. Existing research for guiding decision-making in this setting (for example, Ringuest et al, 1999), to inform preferred strategies for managing technology R&D portfolios, is surprisingly limited.

In this evolving setting, wherein it is now much more readily apparent that science-based models often fail to embrace state variables (or outcomes) of vital importance to the public’s perception of the environment, the potential for employing Belief Networks (Varis, 2002), or Bayesian Networks (Borsuk et al, 2003), as models has begun to be exploited. The tentative conclusion is that the field has been transferring its attention away from Statistical Uncertainty, and accounts of uncertainty as randomness, towards Scenario Uncertainty (with probabilities of outcomes rising and falling in a Bayesian updating sense) and Ignorance, (components of which may be anything but random), even to the notion of designing models specifically for the discovery of ignorance (Beck, 2002, 2005). Indeed, there has been a trend towards exploring the consequences of a plurality of cultural/social perspectives, not only (as one might expect) on the normative goals for desirable states of the environment, but also on the scientific composition of a model. Provocative though this may seem — to suggest that a member of Cultural Theory’s egalitarian social solidarity (within the group of public stakeholders) would incorporate into a model of the climate system bits of the science/knowledge base different from those of his/her counterparts in the hierarchist or individualist social solidarities — this has been done (van Asselt and Rotmans, 1996).

In the US, then, we can see a comprehensive government-wide concern for handling uncertainty in environmental simulation models, including in the extremely challenging computational context of the inexorable drift towards virtual realities (Nicholson et al, 2004). In the EU we see something of the same, but more distinctively a growing interest in addressing uncertainties deriving from stakeholder perceptions and aspirations (Pahl-Wostl and Kabat, 2004).
 

Adaptive Management

The National Research Council’s (NRC) review of the US TMDL program made much of the issue of uncertainty in the underlying science base (NRC, 2001) and the current Adaptive Implementation project (Reckhow et al, 2004) is a direct outgrowth of the challenges in dealing with the issue. By implication, such “adaptivity” must also be an approach to be deployed in making environmental decisions in spite of uncertainty.

We know what adaptive management is (Holling, 1978). And we know its time must have come when the NRC issues a report entitled “Adaptive Management for Water Resources Planning”, as its response to the charge of assessing the methods of the US Army Corps of Engineers in this domain (NRC, 2004). In essence, policy therein fulfils two functions: to probe the behavior of the environmental system in a manner designed to reduce uncertainty about that behavior, i.e., to enhance learning about the nature of the physical system; and to bring about some form of desired behavior in that system. It is possible, however, to conceive of going beyond adaptive management thus defined, in what has been called adaptive community learning (Beck et al, 2002; Beck, 2002). If, in the more open, participatory setting of environmental stewardship, one were to start from eliciting the aspirations of stakeholders for the future of their cherished piece of the environment — both their hopes and fears — what meaningful questions can formal computational analyses with models address, under such conditions of gross uncertainty about the longer-term future and dissonance amongst the plurality of stakeholder aspirations? The procedure of adaptive community learning is one response to such a question. It enfolds the prospect of generating foresight by tapping into {stakeholder imagination & scientific models}, this being the complement of the more familiar use of {scientific empirical observations & scientific opinion}. It holds out the prospect too of identifying priorities for purchasing further science, as necessary, that are robust against a lack of consensus about the community’s future aspirations. In particular, adaptive community learning ought both to subsume the principles of adaptive management (as defined above) and include actions, or a process of decision-making, whereby the community of stakeholders experiences learning about itself, its relationship with the valued piece of the environment, i.e., the community-environment relationship, and the functioning of the physical environment (Beck et al, 2002). Just as adaptive management celebrates a prudent measure of experimentation, so does adaptive community learning (Norton and Steinemann, 2001). The process will be iterative and continuous, one of “always learning, never getting it right” (Price and Thompson, 1997).

The public record on the recent use of adaptive strategies among the US regulatory agencies shows, furthermore, how hard it is to introduce them in practice. Despite strong and steady encouragements from a string of both Democratic and Republican Administrations since the late 1970s, it is extremely rare that public agencies have been able to adapt existing rules to new knowledge as it emerges (Finkel, 2003; McCray, forthcoming). In fact, the single prominent example of a program that introduces routine adjustments in the light of new science is in setting the particulates standards in the US EPA’s air pollution program. We still need to understand those factors — among the interest groups and the policy programs themselves — that seem to thwart the simple idea of policy being routinely adapted to accommodate new science.

Whether in North America, where adaptive community learning and its embedded notion of the computational analysis of reachable futures echoes some of the ideas of Robinson (2003), or in the EU, where it is reflected in the work of Kieken (2002), can environmental regulations really entail the dual functions of adaptive management: to manage and to learn from one and the same policy action? It is not hard to imagine a legal challenge, to the effect that government would be playing “fast and loose” with the environment, if it were to promulgate “experimental” actions.
 

Legal Context

When decision stakes are high, there is discord amongst the social groupings, and incomplete, insecure, or immature science has to be mobilized in a context of policy proximity and data poverty — all conditions very different from the setting in which science is conventionally undertaken — the making of environmental decisions will be subject to challenge, and frequently so in a legal context. The legal process, more so than coming to a view in the scientific community on the pedigree of a theory, must come to a summary judgement on whether the inferences drawn — to make the decision — are justified given the evidence. Regulatory agencies can be asked to explain their science in court; courts can determine whether regulatory decisions were supported on a rational basis (as opposed to being arbitrary or capricious); and they can rule on whether scientific evidence is to be admitted into a trial. Where such distillation has to be achieved, down to the very essence of a yes/no judgement on the legitimacy of a complex assembly of uncertain, disparate bits of the science bases, often under the prospect of looking into the future (as opposed to past misdemeanors), the legal process will demand the sharpest of thinking about all of the foregoing. Moreover, it will require some assessment of what might lie in the realm of Ignorance, or epistemic uncertainty (Pascual, 2005).

In their review of thirty years of judicial challenges to US EPA rule-makings, McGarity and Wagner (2003) concluded that “with respect to the judicial review of modeling exercises, the DQA [Data Quality Act] is likely to have a limited effect, at most”. The length of their paper and the number of cases cited suggest an already high rate of challenges, as also apparent a decade earlier, in respect of groundwater models (Bair, 1994). While models are by no means the same as all of science, rule-makings about environmental protection must so often address issues of a forward look into the future that they cannot be made without the use of models. Indeed, there is a paradox here (Beck et al, 1997). The greater the degree of extrapolation from past conditions, so the greater must be the reliance on a model as the instrument of prediction; hence, the greater is the desirability of being able to quantify the validity (or reliability) of the model, yet the greater the degree of difficulty in doing just this.

Finkel (2003) pays thus much attention to the issue of how agencies, such as the US OSHA and EPA, should evaluate new models, as candidates for employment in the assessment of risk, against criteria such as “scientific support”, “goodness of fit” (to history), and “replicability”, all of whose application — he notes — will be subjective. Three things are important here. First, the legal approach to standards of evidence, or burdens of proof, are brought to bear on thinking about model evaluation; second, as we have seen above, the key is acceptance or rejection, in effect, of competing entire assemblies of theories about the behavior of complex systems; and third, Finkel’s discussion of what happens as time passes, where default models are retained or replaced by better alternatives, is redolent of the Bayesian updating framework in statistics. And this last, of course, recalls the role of Bayesian nets in the work of Borsuk et al (2003) and their employment in facilitating court discussions as to when to accept competing theories for explaining whether crimes have, or have not, been committed (Pascual, 2005; see also similar insights from McNally, 2003).

3          APPROACH

If the trend has been away from a primary concern hitherto with Statistical Uncertainty (aleatory uncertainty) towards epistemic uncertainty (Ignorance); if the monolith of society (stakeholders) has now been differentiated and fractionated into the more subtle pluralities of understandings and perspectives of Cultural Theory (Thompson et al, 1990); if this plurality will have an increasing bearing on the negotiations of rule-making and judgements about scientific reliability; and if there has been but little cross-fertilization of ideas between Europe and the US, especially between governmental agencies, then approaches to the making of environmental decisions in spite of uncertainty may be on the cusp of substantial change. It is the purpose of this proposal to debate all of these macroscopic assertions, in particular, to draw this debate together from across the many disciplinary fields in which it is taking place, thus to frame the issues for the future of environmental decision-making in spite of uncertainty, and to establish a collaborative network (TAUC) for their further elaboration and resolution.

Our work program is divided between the two Objectives set out in Section 1, according to the schedule of Figure 1. 

Program of Work # 1: To Identify Key Issues for Future Research Agenda

A majority of the requested funding will be associated with the mid-project Workshop (task (f) in the following sequence), to be hosted in the US in month 9.

 (a)       Establish Workshop Organizing Committee. Composition of the Committee must strike balances between membership from the US and EU, between members representing government agencies and academia, and amongst the many contributing disciplines. Membership will include senior US EPA policy officers and their counterparts in the EU, for instance, from the European Commission’s Directorates General for the Environment and for Research and Development, the European Environment Agency, the RIVM/MNP of the Netherlands, the UK Environment Agency, support staff from the EU Parliament, and the Organization for Economic Cooperation and Development (OECD). Such commitments — in effect, from these key agencies — will be the basis on which to compose the subsequent Steering Committee of TAUC (see below under Program of Work #2).

 (b)       Assemble Inventory of Understandings of Uncertainty. Each PI will assume responsibility for surveying two out of the six subject domains listed in Section 2, as follows: PI Beck, Technology and Risk and Models and Methods of Analysis; PI Reckhow, Policy and Administration and Adaptive Management; PI Oye, Science and Society and Legal Context. The purpose here will be to identify key contributing individuals, identify similarities and differences of approach on the two sides of the Atlantic, and ensure that the framework of the inventory meshes disciplinary contributions together in a seamless fashion. The three PIs will together be responsible for preparing a first draft of a synthesis/overview paper (tasks (e), (f) and (g) below).

 (c)       Commission Background Papers. Given the rounded context created by task (b), specific individuals will be assigned responsibilities to prepare domain review papers for presentation at the Workshop and subsequent publication in the open, peer-reviewed literature. These papers will be key to defining the content of the Workshop and to establishing the initial work program of TAUC. The Workshop Organizing Committee and, especially, the PIs will oversee preparation of these papers, through teleconferencing.

 (d)       Contact/assemble Workshop Participants. The Workshop is to be by invitation only. Authors of the literature cited herein are prima facie candidates (amongst others) for participation. A number of policy-related persons, drawn from the agencies listed under task (a) above, have likewise already been identified for participation.

 (e)       Design Workshop. The Organizing Committee is to design the Workshop, using a mix of formal presentations and discussions and breakout sessions, with very special attention paid to achieving synthesis across disciplinary contributions. Four criteria will be uppermost in the design process, that: (i) all participants should emerge with a sound understanding of the whole of the subject; (ii) issues for future research should be readily identifiable; (iii) attention can be focused on shaping the TAUC network (see Program of Work #2); and (iv) a path towards promoting TAUC within the setting of the second IWA Symposium on “Uncertainty and Precaution in Environmental Management” (UPEM II) is apparent (see also Program of Work #2). To realize (i) and (ii), the six domain review papers will be presented in plenary sessions, reflecting thus the six themes identified previously in the Background (Section 2) of this proposal. Two previously designated respondents will be allocated to each of the six review papers, with respondents drawn from subject domains different from that of the paper. Breakout sessions, followed by reports in plenary session, will be deployed in realizing (iii) and (iv). Participants in each breakout session will be allocated beforehand to preserve balanced mixes of disciplines and policy/research personnel, once more in the interests of achieving seamless integration of the whole of the subject matter and maximizing the inter-disciplinary “literacy” of participants. Avenues for funding of TAUC beyond this proposal will be express objects of discussion in the breakout session associated with (iii). One breakout session and accompanying plenary session will be devoted to the task of assembling identified future issues for research into a synthesis paper. A central goal of the Workshop will be to achieve a comparative assessment of procedures and strategies in the US and EU.

 (f)        Workshop. Convenience for international participation and agency presence may best be served by having the Workshop in Washington, DC; however, hosting it on the campus of the University of Georgia would be cheaper. A budgetary-based decision on location will be made early in the project. Current plans are for project funds to cover entirely 20-24 participants (depending on location), with as many as up to 10-15 additional persons participating with support from other sources, if deemed appropriate (including participants from “affiliate countries”, such as China, Japan, South Africa, and Brazil, for example). Three products are sought: (i) a brief position paper for Science (in the manner of Kates et al (2001), ideally with the collective authorship of all Workshop participants); (ii) publication of the set of domain review papers and synthesis paper, which latter will enfold the Workshop discussion; and (iii) publication (eventually) of the TAUC Manifesto (under Program of Work #2).

 (g)       Publish Workshop Proceedings and Position Paper. Submission for publication of the revised domain review papers and the (Science) position paper within six months of the workshop is planned. The Workshop discussion will be used to provide feedback on the review papers; further internal review prior to submission will be the responsibility of the Workshop Organizing Committee. The PIs will assume responsibility for drafting the (short) position paper and the synthesis paper. The latter will be an expression of a future agenda for research — a set of “grand challenges” in making environmental decisions in spite of uncertainty — hence the key outcome of this first Work Program.

 Program of Work # 2: To Establish An Enduring Sustainable Network of Experts (TAUC)

The goal of this second Program is to launch TAUC and promote it within the wider context of UPEM II (and beyond).

 (a)       Facilitate (US) Agency to (EU) Agency Cooperation. Our purpose will be to establish the foundations of an enduring (government) agency-to-agency collaboration. Implementation will be through a visit of EPA officers (and the PIs) to Europe, possibly to institutions at several sites, but preferably as a single meeting at just one location (London, Brussels, or Copenhagen). The meeting/visit should conclude with the setting up of an Interim Steering Committee for the TAUC network. A further important outcome will be a commitment from the EU partners to match the funds sought in this proposal and to deploy them in support of the initial work program of TAUC (beyond the time-span of this proposal).

 (b)       Initiate Preparations for UPEM II. The option of a UPEM II, under the auspices of the International Water Association (IWA), while informally mentioned in Copenhagen (UPEM I, 2004), currently remains merely a possibility. Since the IWA Secretariat is in London, this approach — proposing to formalize and proceed with convening UPEM II (roughly within 24 months of the start of this proposal) — could be undertaken in conjunction with the foregoing task (a). Given private discussions at UPEM I, the European Commission’s Joint Research Centre (JRC) at Ispra, Italy, may have a strong interest in hosting UPEM II. This may also have implications for the funding decisions of task (a).

 (c)       Formalize “Articles of Association” and Establish Work Program for TAUC. Other models of the successful governance of networks are available for review, most notably the European Forum for Integrated Environmental Assessment (EFIEA) and the Resilience Alliance Network of Holling. The Interim Steering Committee will also be charged with assessing what to include in an initial portfolio of network activities, such as: preparation of proposals for subsequent funding; designing vehicles for providing advice to governments; setting up and monitoring case studies in making environmental decisions in spite of uncertainty; aligning TAUC functions so as to create synergies with other NSF programs, notably CLEANER and the Centers for Decision Making Under Uncertainty (in response to climate change); and designing Master Classes for training of young professionals in handling uncertainty at the interfaces among Science, Technology, Policy, Law, and Society.

 (d)       Publish TAUC Manifesto. The Manifesto will be a direct response to the grand challenges set out in the synthesis paper on future research (the product of Program of Work #1).

 (e)       Promote TAUC and Disseminate Its Vision at UPEM II. To ensure that due attention is given to this, we shall be looking to the allocation of EU matching funds (see task (a) above) specifically to support the TAUC contribution to UPEM II. This closing task must also effect the transition from the Interim Steering Committee to the form of governance for TAUC drawn up under the foregoing task (c).

 4          EXPECTED RESULTS AND SIGNIFICANCE

There is currently great interest in the challenges of handling uncertainty in the making of decisions in matters of environmental stewardship. Despite this, there is little cross-fertilization of thinking between the quantitative approaches to the analysis of uncertainty in the engineering, mathematical, and statistical sciences and the non-quantitative approaches residing in the subjects of, for example, law, anthropology, and psychology (Ayton and Fischer, 2005). Issues of uncertainty as “epistemic uncertainty”, or as “ignorance”, are becoming recognized as predominant and recalcitrant in the latter subjects and in the more philosophical treatments of environmental policy-making. Yet even the word “ignorance” can be highly disquieting to many in the community of engineers and natural scientists, many of whom barely accept the presence of uncertainty. In short, a coherent treatment of the subject of uncertainty in a fully well-rounded manner is absent. Just as conspicuous by its absence — in these times of global communication — is effective cross-fertilization between the US and EU countries of their respective experiences of handling uncertainty at the intersection between environmental science, policy, law and society.

An agenda for the future grand challenges of research in environmental decision-making in spite of uncertainty and the launching of a multi-disciplinary network of experts designed to begin responding to this agenda (as one of several possible responses) will be the two most important results of this project. This research agenda will be fashioned primarily according to the increasingly urgent needs of hard practice — not conceptual conjecture — from policy-making as it is presently occurring, and in its associated legal context. The network will be designated the TransAtlantic Uncertainty Colloquium (TAUC). Its backbone will be that of (US) Agency to (EU) Agency cooperation, thus to ensure it is policy-focused and, therefore, of both practical and conceptual significance. The importance of TAUC is that it must be a vehicle for regulator to speak to regulator, and for regulators to talk productively to policy and engineering/science researchers.

 5          GENERAL PROJECT INFORMATION

The curiosity and capacity to move amongst a variety of disciplines, and the skills of building, sustaining, and focusing cooperative, international networks on specific outcomes, lie at the heart of this proposal. PI Beck is based at the University of Georgia, but has a visiting appointment at the Imperial College of Science, Technology, and Medicine in London, UK, from where he leads an industry-focused network project on “Engineering for Sustainable Development”. He has also maintained long-term associations with the International Water Association (IWA) — whose Leading-Edge Program on Sustainability in the Water Sector he currently leads — and the International Institute for Applied Systems Analysis (IIASA; Laxenburg, Austria), for whom he co-chairs the Dynamics Systems Network Project on “Control Theory and Environmental Systems”. Beck led the International Task Force on “Forecasting Environmental Change” (1992-1998), which culminated in the monograph Environmental Foresight and Models: A Manifesto (Beck, 2002). Suffice it to say in the present context that PI Beck has substantial and extensive experience of participating in, initiating, and leading inter-disciplinary, internationally networked activities.

Likewise, PI Reckhow works at the interface between statistics, decision-theory, policy, and the environment, almost single-handedly bringing the issues of uncertainty to the attention of a very wide audience of stakeholders through his chairing of the 2001 NRC Committee reviewing the science base of the TMDL Program. He currently leads a similarly networked team of theoreticians and practitioners who will this year (2005) deliver a White Paper on Adaptive Implementation of watershed management within the context of the Clean Water Act. In his capacity as Director of the North Carolina Water Resources Research Institute he led one of the most successful of all such (virtual) networked state entities. In addition, Reckhow has been a primary motivating force in preparing the ground for implementing some of the NSF’s most substantial research networks (CLEANER and the Hydrologic Observatories). PI Oye formally combines an inter-disciplinary outlook in his joint appointment as Associate Professor of Political Science and of Engineering Systems at MIT. He has served two terms as Director of the MIT Center for International Studies, is now forming a Political Economy and Technology Policy Program within the Center, has been a guest scholar at the Brookings Institution, and has served as a consultant to the Institute for International Economics and the U.S. Trade Policy Coordinating Committee. Continuing in this international seeting, he is currently working with a team from the EU, Japan and the US on regulatory strategies of private firms.

PI Oye’s collaboration with Lawrence McCray — on knowledge assessment in areas marked by controversy and scientific uncertainty and on improving the adaptive capacity of public institutions — brings very considerable added value to this proposal. McCray is now working on Knowledge and Policy-making at MIT, having previously been head of regulatory reform at EPA, director of reform projects at the Executive Office of the President, and in senior management at the National Academy of Sciences/National Research Council, where — significantly — he served as project director for the NRC Red Book (1983).

Current NSF Support

PI Reckhow is Co-PI on “Development of a Plan to Achieve a CLEANER Neuse River Basin in North Carolina” (01/01/04 through 12/31/05).

PI Oye is Co-PI on the MIT IGERT “Assessing Implications of Emerging Technologies” (01/01/04 through 12/31/09).

 Figure 1: Project time-table, programs, and tasks