next up previous
Next: Involving Users Up: USGCRP Previous: Useable information.

Credible prediction.

The latest USGCRP program plan has stated that ``Models of the Earth system will provide the only rigorous means for developing quantitative projections of the interactions of atmospheric composition, climate, sea level, terrestrial and marine ecosystems, agriculture, water resources, and the effects of human activities'' [CENR, 1994 p 67]. Earlier program plans explicitly claimed that such predictions are necessary to making sound policy. For example: ``Recognizing that effective and rational response strategies to environmental issues can be built only on sound scientific information, the overarching goal'' of the USGCRP is ``To gain a predictive understanding of the interactive... processes that regulate the total Earth system and, hence, establish the scientific basis for national and international policy formulation... .'' [CES, 1989 p 10]. Even today there is a clear implication that a more thorough and quantitative understanding of all the Earth's systems and their interactions is needed to make sound policy. (note: ``Models... . provide the predictive link between the physical Earth system and the human dimensions of global change, including economics, social structure and evolution, and resource use and management'' [CENR, 1994 p 67].) Brunner has shown this is not the case [1993]. Further, because the Earth is an open system, the promise of credible prediction of future climates is not likely to be fulfilled, not with a model that can be validated and verified [Oreskes, et al., 1994]. Nor is it clear that present approaches will be able to predict rapid shifts. Models of such complex systems are more useful as an aid to understanding than as predictors. The word ``credible''---meaning capable of being believed, believable---does not address whether anyone will believe the predictions nor whether the information will be used. As users have not been significantly involved in planning programs, neither useful results nor their use are guaranteed. Promising a policy-relevant (usable) model without involving users is the kind of insulation from real-world issues that Brown and Mikulski decry.

Are the promises of useful information and credible prediction examples of the temptation to avoid tough issues? In passing P.L. 101-606 calling for usable information Congress tried to take itself off the hook. That is, were the climate to shift tomorrow, Congress could say ``We asked to be kept informed. It's not our fault we aren't ready!'' Of course this is a hollow disclaimer for Congress has not faced the tough question of what information it needs to be ``ready.'' Similarly, in arguing that a thorough understanding of Earth systems is necessary for policy making the USGCRP essentially falls back on the Vannevar Bush model: Basic research leads to understanding and understanding leads to good decisions. Unfortunately, a comprehensive understanding of the Earth realized in a model is an extremely difficult goal, and will do more to rationalize decades of basic research than to support current decisions. However, the program can not evade its responsibility because by implying that a ``rigorous... quantitative prediction'' is necessary it steps outside science and into a very different field, policy. The implication is not believable because policy decisions are made---must be made---every day without a predictive model of the relevant phenomena. Such current decisions need support. Rather than supporting research the community wants to do (the Bush model), it might be wiser for the program to tackle the job of finding out what is needed for decisions being made now.

If the program is not on the right track, the question becomes: How to change direction gracefully; what will be acceptable to users? Recall that users such as Senator Mikulski seek research to address national strategic goals. Given the budget situation there is little hope of simply adding program elements or of increasing funding---for example, for modelling efforts to try to force a solution.

Suppose an oversight committee in the new Congress asks what useful information has been delivered for the several billions already spent? The following argument might be made: The USGCRP had to be started somehow, bootstrapped; scientists had to discover and delineate the problem before they could present it to decision makers. At the beginning they could not ask users what information they needed because the climate problem was not well-defined. This is all true but may be irrelevant.

Although useful results have been delivered not many are a direct consequence of the USGCRP per se and the related incremental spending of the last several years. Compared to what a billion dollars will buy in other areas of research there may not be billions of dollars worth of useful results. The bootstrapping argument may be undermined because after billions of dollars we can no longer claim to be at the beginning of the program. As many global change issues need illumination and related policy issues need support, it would be a major loss if the program fell into disfavor and were cut. Accordingly, consideration should be given to reorienting the USGCRP around the primary goal of prompt and continuing provision of useful information, as defined by those who will use it, and consistent with its legislative mandate. A less drastic alternative would be to rejustify the existing program on scientific grounds and set up a new program to provide information readily usable by officials from what is currently known about global change [Brunner, 1993]. A major activity would be to elicit the information needs of officials, and a major program element would be regular evaluation against such practical needs. The new program could be funded at a low level compared to the USGCRP because it would not need satellites or large data systems; if modest reductions were achieved in the rejustified USGCRP, the two programs could be conducted at the same cost as the original one.

This alternative recognizes two different activities: Generating new information by means of research versus providing readily usable information in response to urgent policy questions. The secret to providing useful information does not lie in short-term reorientation of research to answer urgent questions; unfortunately it does need to be said that because research is slow, urgent questions can be answered only by assessment of what is already known. Such assessments can be provided by analysts working at the policy/science interface and capable of interpreting both policy questions and scientific information. Consequently many more such interface analysts are needed both to answer and to elicit policy questions. (note: A potential source of jobs for some of the oversupply of PhDs in the natural and social sciences.) Research planning can be informed by studying the kinds of specific questions users ask today in order to generate useful information for tomorrow's questions. Participation of interface analysts would be critical to such planning.

To summarize the situation with respect to the USGCRP:

The prospects have been questionable because of the program's history of poor coupling with users, and because of the focus on large satellites and models which make it hard to inject policy learning into the program. However, positive program changes are proposed and the program is conducting good science [OTA, 1993].



next up previous
Next: Involving Users Up: USGCRP Previous: Useable information.



U.S. National Report to IUGG, 1991-1994
Rev. Geophys. Vol. 33 Suppl., © 1995 American Geophysical Union