Some Inconvenient Truths in Bringing Climate Science to the Judiciary
This paper translates scientific assessments of likelihood and confidence into legal standards like “more likely than not” and “beyond a reasonable doubt”. It speaks to a core challenge for the judiciary – to navigate the tension between law’s need for defined thresholds and science’s inherently probabilistic and context-dependent reasoning.
Climate science had been knocking on the courthouse doors for quite some time when the Supreme Court of the United States (SCOTUS) finally invited it into the realm of legal action in 2007 when the justices issued their ruling on Massachusetts v. EPA, 549 U.S., in April. By a 5-4 vote, the court found that the Clean Air Act of 1970, as amended in 1990, allowed the Environmental Protection Agency (EPA) to regulate greenhouse gases.
More specifically, greenhouse gases were declared pollutants because it was reasonable to anticipate that their emission into the atmosphere could “endanger public health or welfare.” To act on this finding, however, the EPA was required to show that concrete harm could “be traced to emitters and damages could be remedied by the courts.”
The EPA then spent two years assessing the science before publishing its first endangerment finding EF(2009) on December 7thof 2009. In their words, “The Administrator finds that the current and projected concentrations of the six key well-mixed greenhouse gases (GHGs)—carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), hydrofluorocarbons (HFCs), perfluorocarbons (PFCs), and sulfur hexafluoride (SF6)—in the atmosphere threaten the public health and welfare of current and future generations.”
The judicial door had been opened, but challenges remained. The EF(2009) finding withstood more than 20 separate attempts to have it nullified.
That all changed in 2025 when the federal government mounted more systematic attacks. One was designed to slam the door shut, and the second tried to cancel any attempts to educate the judiciary about the nuances of using climate science. The goal was to weaken the ability of the courts to judge the liabilities of private and public actors who knowingly ignored their legal and social responsibilities to ameliorate the climate risks of their actions.
Starting on Inauguration Day, the executive branch of the federal government attacked the EF(2009) by eroding the public’s confidence in climate science and climate scientists. Based on biased and immediately debunked assessments of the state of the science published by the EPA, the agency undertook a concerted campaign of disinformation designed to advance minimalist damage estimates, question the value of government actions to slow the pace of warming, and degrade the personal integrity of the entire worldwide scientific community.
Confident that its campaign was working, the EPA officially rescinded the 2009 Endangerment Finding early in 2026.
EPA Administrator Zeldin expressed his concern that the science on climate risk had not yet been settled and so had no business being communicated to the judiciary. Encouraged by those words, state Attorneys General quickly demanded that the Federal Judicial Center (FJC) and the National Academies of Science, Engineering, and Medicine (NASEM) excise Chapter 21from the 4th edition of the “Reference Manual on Scientific Evidence.” That new chapter had been written to summarize what the judiciary needed to know about climate change, but the FJC acquiesced.
Why is that so important? Because the chapter could have become the anchor of many efforts to help judges and lawyers and practicing lawyers understand what climate science can and cannot be expected to bring to legal arguments about (1) the applicability of the Clean Air Act in framing new and existing environmental policies and (2) the means by which court rulings can legitimately assign criminal and/or civil liability.
The second point above is italicized because it represents a second fundamental challenge for the judiciary. In 2007, SCOTUS used a “could reasonably be expected” legal threshold to support its decision. To me, they thereby applied an appropriate standard that reflects a specific confidence level to which climate science assessment can offer valuable and well-documented information.
It also seems to me that climate science can speak to a “beyond a reasonable doubt” threshold for criminal cases by presenting scientific knowledge that had been accepted with very high likelihood” (when rigorous statistical analyses were available) or “very high confidence” (when quality data was available and supported widely accepted process understanding) to the court. It can also speak to “more likely than not” thresholds with statistically supported 50–50 likelihood estimates or “medium confidence” assessments.
Well-researched documents, like the afore-mentioned Reference Manual and the integrated “Climate Science and Law for Judges” curriculum published by the Climate Judiciary Project, are vital resources for the judiciary. The latter was first published in 2023, but it is currently being updated in the face of strong public opposition from the Department of Justice.
Both of these documents are designed to bring the complexities of climate risk assessments to the fore. Specifically, they make it clear that likelihood and confidence assessments depend critically upon the social and economic contexts within which they have been conducted. “More likely than not” assessments can, for example, be accurate descriptions of reality for some plausible contexts and not for others. This means that legal liability can depend not only on what a defendant did or did not do to effect climatic dynamics, but also upon what other citizens did to alter the broader context.
Should, for example, a defendant whose actions have increased the likelihood or consequences of a flash flood be held completely accountable for catastrophic damages if the impacted community did not provide adequate early warning so that its citizens could get out of harm’s way?
None of this is to say that the legal system cannot accommodate competent descriptions of complex climate risks. I am arguing instead that accommodation is a complicated and locale-specific challenge that cannot be met successfully if the judiciary is not informed honestly about what to expect from the underlying science.
________________________
This article appeared in the The National Law Review, Volume XVI, Number 119, April 29, 2026.
_______________________
Gary Yohe, is the receiving agent of Gary W. Yohe, LLC and Huffington Foundation Professor of Economics and Environmental Studies (emeritus), Wesleyan University.