Patrick J. Michaels • April 21, 2020
The U.S. Environmental Protection Agency’s (EPA) 2009 “Endangerment Finding” from carbon dioxide (CO2) and other greenhouse gases grants the agency a legal mandate that can have profound and far-reaching effects. The Finding is based largely on a Technical Support Document that relies heavily upon other mandated reports, the so-called National Assessments of global climate change impacts on the United States.
The extant Assessments at the time of the Endangerment Finding suffered from serious flaws. We document that using the climate models for the first Assessment, from 2000, provided less quantitative guidance than tables of random numbers—and that the chief scientist for that work knew of this problem.
All prospective climate impacts in the Endangerment Finding are generated by computer models that, with one exception, made systematic and dramatic errors over the climatically critical tropics. Best scientific practice would be to emphasize the working model, which has less warming in it than all of the others.
Instead, the EPA relied upon a community of wrong models.
New research compares what has been observed to what is forecast, and finds that warming in this cen- tury will be modest—near the lowest extreme of the prospective range given by the United Nations.
The previous administration justified its policy choices by calculating the Social Cost of Carbon [dioxide]. We interfaced their model with climate forecasts consistent with the observed history and enhanced the “fertilization” effect of increasing atmospheric concentrations of CO2. We find that making the warming and the vegetation response more consistent with real-world observations yields a negative cost under almost all modeled circumstances.
This constellation of unreliable models, poor scientific practice, and exaggerated estimates of the Social Cost of Carbon argue consistently and cogently for the EPA to reopen and then vacate its endangerment finding from carbon dioxide and other greenhouse gases.
The Environmental Protection Agency’s (EPA) 2009 “Endangerment Finding” is the basis for comprehensive regulations of carbon dioxide emissions under the Clean Air Act.1 It is based on another EPA report called a Technical Support Document (TSD), which, as this paper will show, is seriously flawed because it is based upon climate models that are making large systematic errors. Therefore, the subsequent impact models based on them are also unreliable. These project, among other things, changes in crop yields, ecosystems, and social systems caused by changes in climate. Further, the EPA did not follow best scientific practices in determining either the course of future climate or the cost of current and future carbon dioxide emissions.
Subsequent to the Endangerment Finding, the Obama administration justified its interventionist policies with its calculation of a figure known as the Social Cost of Carbon (SCC). The SCC is a monetary estimate of the damages supposedly caused by an incremental ton of carbon dioxide (CO2), the principal contributor to human-induced climate change, emitted in a given year. Discernible in neither economic nor meteorological data, social cost values are guesstimates produced by computer programs called “integrated assessment models” (IAMs). What IAMs “integrate” is a speculative model of how carbon dioxide emissions will change the climate with a speculative model of how climate change will affect consumption, GDP, and health. Only one of these models, known as FUND (for Framework for Uncertainty, Negotiation, and Distribution), contains an explicit term to account for the fertilization effects of increasing atmospheric carbon dioxide. All of the IAMs are based upon a warming that is too great—in which sensitivity to carbon dioxide is too high—based upon recently observed data.
Here, we will see that more realistic values for climate sensitivity and the fertilization effect of carbon dioxide, along with realistic discount rates, can produce a SCC that is negative, or something that could be considered a net benefit, certainly calling the concept of “endangerment” into question.
We will show that multiple and internally consistent lines of evidence, such as an inability to define serious costs with realistic model parameters, and two tragic flaws in the supporting “science summary” documents that underpin the original finding, should compel the EPA to reconsider and then vacate its 2009 Endangerment Finding.
History of the Endangerment Finding
The Endangerment Finding was produced in response to a 2007 Supreme Court decision, Massachusetts v. EPA, that the Clean Air Act Amendments of 1992 empowered the EPA to regulate emissions of carbon dioxide, if the agency found that they endangered human health and welfare.
The case against the EPA was originally brought by 12 states and several environmental advocacy organizations. It eventually came before the District of Columbia appellate court, which upheld an original decision in favor of the EPA’s choice not to regulate CO2, on the grounds of scientific uncertainty as to the amount and effects of climate change actually caused by increasing atmospheric concentrations.
The appellate court decision was split 2-1, with a vigorous dissenting opinion by Judge David S. Tatel. The Supreme Court granted a writ of Certiorari to Massachusetts v. EPA on June 26, 2006. Judge Tatel’s dissent was highly influential when the case came before the Court, argued on November 29, 2006. The decision, a narrow 5-4 verdict, was announced on April 2, 2007. The majority held that the Clean Air Act did grant the EPA authority to regulate greenhouse gas emissions as “air pollutants.”
The timing was late in President George W. Bush’s second term, which deferred substantive action. Incoming President Barack Obama did not, placing global warming as his second highest priority in his first inaugural address (behind national health care). His EPA issued a preliminary “finding of endangerment” less than three months later, on April 17, 2009, and a final finding was announced on December 7, 2009.2 That date coincided with the opening of the 15th Conference of the Parties to the United Nations 1992 Framework Convention on Climate Change (UNFCCC), also known as the Rio Climate Treaty. The 2009 meeting was supposed to adopt a new emissions reduction protocol to replace the failed 1997 Kyoto Protocol, but it failed to do so.
The Endangerment Finding is backed by a Technical Support Document, which itself was largely based on the Fourth Assessment Report (AR4) of global warming by the United Nations Intergovernmental Panel on Climate Change (IPCC)3 and the second National Assessment of climate change impacts on the United States by the U.S. Global Change Research Program (USGCRP).4
Both of the foundational documents for the TSD have only one method to predict future climate—a large series of computer simulations of global climate with enhanced greenhouse gases, known as general circulation models (GCMs) or Earth system models (ESMs). The output of these models is then used to drive “impact” models, which apply to the many aspects of life, projecting changes in migration, death, nutrition, mental illness, among others.
If the GCM/ESMs can be shown to be fatally flawed, then their prospective climate forecasts are useless. Therefore, any forecasts of changed impacts are worse then useless; they may even have negative utility, because impact models have their own error terms and those interact multiplicatively with the errors in the GCM/ESMs. The application of this error cascade to policy can provoke unneeded suffering and impose grotesque and gargantuan opportunity costs. In this eventuality, the Endangerment Finding becomes an exercise in expensive rhetoric.
Are The National Assessments Tainted?
The final version of the Technical Support Document, published on December 7, 2009, relied heavily upon a document published by the U.S Global Change Research Program, “Global Climate Change Impacts in the United States”, also known as the Second National Assessment (NA-2) report. Since then, it has become clear that the models used for prospective climate in that report made significant systematic errors in the three-dimensional tropical atmosphere with significant implications concerning the reliability of subsequent forecasts for temperature and precipitation over large expanses of the planet.
The first National Assessment (NA-1) was published in 2000. Thomas Karl, Director of the National Oceanic and Atmospheric Administration’s (NOAA) National Climatic Data Center (NCDC), Jerry Melilo of the Marine Biological Laboratory, and Thomas R. Peterson, also of NCDC led the National Assessment Synthesis Team (NAST). Like the subsequent NA-2, it had a fatal flaw.
The design of the NA-1 was similar to the succeeding Assessments. All future impacts were generated from the output of general circulation climate models. At the time NA-1 was under development, NAST had eight GCMs to choose from to describe 21st century climate. They chose two: a) the Canadian Climate Centre model, which produced the greatest temperature changes over the U.S. of all eight, and b) the model from the U.K. Hadley Center, a part of the Meteorological Office, which produced the largest precipitation changes.
In our peer review of the draft NA-1, my then-colleague Chip Knappenberger and I looked at how well those two models could do the simplest of tasks—simulating 10-year running means of coterminous U.S. temperature averages over the 20th century, such as, for example, 1900-1909, 1901-1910, and so on. They could not do it. The answers the models gave were worse than simply assuming the 20th century average value. In other words, the models added errors to the raw data. What they did was ex- actly analogous to a student taking a four-option multiple choice test and getting less than 25 percent correct. The NA-1 models actually supplied “negative knowledge”. You knew less about 20th century U.S. climate by using them.
I emailed Karl my result, as we were on good professional terms. He responded that indeed we were correct and that further, the models exhibited the same failure on running means of one, five, 10 (which we analyzed), 20, and 25 years. Using climate models that do not work to assess the effects of climate change (with subsequent policy recommendations) is a scientific malpractice.5 (Karl and Melilo were also on the team that assembled NA-2, along with Anthony Janetos of the environmental advocacy group World Resources Institute. NA-2 also had systematic problems.)
— Patrick Michaels
Systemic Flaws in the Second National Assessment
The 2009 Second National Assessment of climate change impacts in the United States (NA-2) is the principal U.S-specific reference in the Technical Support Document for the Endanger- ment Finding. The weakness of NA-2 was apparent in its draft form. As one of these authors (Michaels) and Chip Knappenberger noted in public comments submitted on behalf of the Cato Institute:
Of all of the “consensus” government or intergovernmental documents of this genre that [we] have reviewed in [our] 30+ years in this profession, there is no doubt that this is the absolute worst of all. Virtually every sentence can be contested or does not represent a complete survey of a relevant literature.
There is an overwhelming amount of misleading material in the CCSP’s [Climate Change Science Program’s] “Global Climate Change Impacts in the United States.” It is immediately obvious that the intent of the report is not to provide an accurate scientific assessment of the current and future impacts of climate change in the United States, but to confuse the reader by a loose handling of normal climate events (made seemingly more frequent, intense, and damaging simply by our growing population) presented as climate change events. Additionally, there is absolutely no effort made by the CCSP authors to include any dissenting opinion to their declarative statements, despite the peer-reviewed literature being full of legitimate and applicable reports that provide contrasting findings. Yet, quite brazenly, the CCSP authors claim to provide its readers—“U.S. policymakers and citizens” with the “best available science.” This proclamation is simply false.
The uniformed reader (i.e., the public, reporters, and policy-makers) upon reading this report will be [led] to believe that a terrible disaster is soon to befall the United States from human- induced climate change and that almost all of the impacts will be negative and devastating. Of course, if the purpose here is really not to produce an unbiased review of the impact of climate change on the United States, but a political document that will give cover for EPA’s decision to regulate carbon dioxide, then there is really no reason to go through the ruse of gathering comments from scientists knowledgeable about the issues, as the only science that is relevant is selected work that fits the au- thors’ preexisting paradigm.6
In 2012 we published an addendum in the form of a palimpsest that featured a point-by-point rebuttal of specific faulty statements in the NA-2.7 For example, under “Key Finding” #7, NA-2 says:
7. Risks to human health will increase. Harmful health impacts are related to increasing heat stress, waterborne diseases, poor air quality, extreme weather events, and diseases transmitted by insects and rodents. Reduced cold stress provides some benefits. Robust public health infrastructure can reduce the potential for negative impacts. (p.89)8
Our addendum notes:
7. Life expectancy and wealth are likely to continue to increase. There is little relationship between climate and life expectancy and wealth. Even under the most dire climate scenarios, people will be much healthier and wealthier in the year 2100 than they are today. (pp. 139-45, 158-61)9
We urge readers interested in documenting the problems between NA-2 and the endangerment finding to do their own side-by side comparisons of both NA-2 and our response, shown in Appendix 1.
Despite all of the problems noted above, the Endangerment Finding was declared final on December 7, 2009. The next section looks at some critical findings subsequent to its publication, which will be important to any attempt to vacate it.