by Judith Curry
Scientific information is important in many policy debates in the Pacific Northwest (salmon; wildfire severity; human activities and climate; genetically modified organisms; water scarcity). Science is essential in such policy debates, but I am concerned that policy-biased science is increasingly common.
Science should be objective and based on the best information available. Too often, however, scientific information presented to the public and decision-makers is infused with hidden policy preferences. Such science is termed normative, and it is a corruption of the practice of good science. Normative science is defined as “information that is developed, presented or interpreted based on an assumed, usually unstated, preference for a particular policy choice.”
Using normative science in policy deliberations is stealth advocacy. I use “stealth” because the average person reading or listening to such scientific statements is likely to be unaware of the underlying advocacy. Normative science is a corruption of science and should not be tolerated in the scientific community — without exception.
Let me illustrate with a current policy issue: “Should certain dams be removed to restore salmon runs?” Scientists can assess with some degree of confidence the likely effects of removing or maintaining a particular dam. Scientific information alone, however, is an insufficient justification for deciding to keep or remove a dam. There are biological consequences of dam removal (and maintenance), and those consequences may be substantial from a salmon perspective, but ecological consequences are but one of many elements that the public and decision-makers must weigh when making a policy choice.
Policymakers, not scientists, decide whether preserving salmon runs should trump flood protection, irrigated agriculture or electricity generation. As the public and decision-makers balance policy alternatives, what they need from scientists are facts and probabilities. What they do not need from scientists are their or their employer’s values and policy preferences masked within scientific information disguised as being policy neutral.
There are other common examples. In working with scientists, I often encounter value-laden terms like “degradation,” “improvement,” “good,” “poor,” “impact,” or “alien invasive.” Scientists should avoid these types of normative words in conveying scientific information. Such words imply a preferred ecological state, a desired condition, an accepted benchmark or a favored class of policy options. This is not science; it is a form of policy advocacy — subtle, sometimes unintentional, but it is patently stealth policy advocacy.
Consider the widespread use of concepts such as “ecosystem health.” It is normative science! “Ecosystem health” is a value-driven policy construct, but it is often passed off as science to unsuspecting policy-makers and the public. Think what the average person actually hears when scientific data or assessments are packaged or presented under the rubric of “ecosystem health.” Healthy is good. Any other state of the ecosystem must be unhealthy, hence, undesirable.
Scientific information must remain a cornerstone of public policy decisions, but I offer cautionary guidance to scientists: Get involved in policy deliberations, but play the appropriate role. Provide facts, probabilities and analysis, but avoid normative science. Scientists have much to offer the public and decision-makers but also have much to lose when they practice stealth policy advocacy.
This topic is about impact to people – your constitutents, my fellow citizens, my two kids – not just polar bears.
1) Most of the warming of the past 50 years is due to human activity, and extensive evidence supports this conclusion
2) Climate change is increasing the probability of extreme events, and in some cases may be strengthening their intensity or increasing their frequency (i.e. we are loading the dice towards more Sandy or blizzard type storms)
3) There is strong evidence that increases in some types of extremes are linked to human-induced climate change, notably extreme heat, coastal flooding, and heavy downpours. For other types of extremes, such as tornadoes, current evidence is much more limited.
His testimony closes with this statement from the recent AMS Policy Statement on Climate Change:
Avoiding this future warming will require a large and rapid reduction in global greenhouse emissions . . . Technological, economic, and policy choices in the near future will determine the extent of future impacts of climate change. Science-based decisions are seldom made in context of absolute certainty. National and international policy discussions should include consideration of the best ways to both adapt and mitigate climate change.
While I am not a fan of the AMS Statement on Climate Change (see my previous post here), it does reflect the deliberations and imprimatur of a professional society.
There is no question that Marshall Shepherd is an effective communicator; he effective uses metaphors and taps into the values and concerns of citizens. The twitter controversy surrounding Shepherd’s testimony is associated with his attribution of extreme weather events to AGW. Roger Pielke Jr has a post on this, excerpts:
Unfortunately, as is so often a case when leaders in the climate science community find themselves before an audience of policy makers, on extreme events they go rogue, saying all sorts of things with little or no scientific basis. Let’s take a step back. The science on climate change, extreme events and disaster costs is clear and unambiguous. You don’t need to take my word for it, you can find the science well summarized in the IPCC SREX. In a nutshell here is the state of the science (here I focus on the US as Shepherd did):
- US floods have not increased over a century or longer (same globally).
- US hurricane landfall frequency or intensity have not increased (in US for over a century or longer).
- US intense hurricane landfalls are currently in the longest drought (7 years+) ever documented.
- US tornadoes, especially the strongest ones, have not increased since at least 1950.
- US drought has decreased since the middle of the past century.
- US East Cost Winter Storms show no trends (here also).
- Disaster losses normalized for societal changes show no residual trends (US, other regions or globally).
- Trends in the costs of disasters are not a proxy for trends in climate phenomena.
In a twitter exchange (and also in his testimony), Shepherd refers to a paper by Trenberth that justifies his statements. The paper is titled Framing the Way to Relate Climate Extremes to Climate Change. Trenberth’s argument is basically: All weather events are affected by climate change because the environment in which they occur is warmer and moister than it used to be. I see two basic problems with Trenberth’s argument:
- There is no empirical evidence for a moister climate (see the paper by Von der Haar et al. [link]
- Warmer/moister does not necessarily imply more or more extreme weather events; weather events are controlled by atmospheric dynamics on scales from mesoscale to hemispheric.
One very clear finding from these studies is that one of the largest uncertainties about future climate relates to the choices that we and our children will make regarding energy use. The more dependent we are on CO2 emitting sources of energy, the more Earth’s climate will change. The decisions we make today will affect future generations and should motivate us to wisely use knowledge from climate science to reduce risks of harm from unnecessarily disruptive climate.