The Water Vapor Problem

The weak point in EPA greenhouse gas regulation

One of the best painters of water, in all of its states — Ivan Aivazovsky.

The U.S. Environmental Protection Agency (EPA), in its ruling last week that overturned its 2009 Endangerment Finding on greenhouse gases (GHGs), argued that in light of Supreme Court judgments since 2009, it was reinterpreting the language of the Clean Air Act (CAA):

With this final action, we return to fundamental principles governing decision-making within our democratic system: “Agencies have only those powers given to them by Congress,” West Virginia v. EPA, 597 U.S. 697, 723 (2022), and “the scope of an agency’s own power” is determined not by deference to asserted expertise, but by “the best reading of the statute,” which is fixed at the time of enactment. Loper Bright Enters. v. Raimondo, 603 U.S. 369, 400-01 (2024). [emphasis added]

However, under the U.S. Constitution responsibility for determining the “best reading” of a statute lies exclusively with the judiciary, and not executive branch agencies.

EPA cited the Supreme Court’s judgment in Loper Bright, which also explained this central tenet of the U.S. Constitution:

[C]ourts may not defer to an agency interpretation of the law simply because a statute is ambiguous.

The notion that courts determine what the law is dates to Marbury vs. Madison in 1803:

It is emphatically the duty of the Judicial Department to say what the law is.

By engaging in judicial review outside its remit, the EPA may simply be begging to be sued, forcing the issue to the Supreme Court.¹

It is of course conceivable that the Court might (a) decide that EPA was wrong procedurally to assume responsibility for judicial review of the CAA, but (b) substantively, the Court concurs with its de novo interpretation of the law in 2026, and thus overturns the 2008 Massachusetts vs. EPA ruling — which determined GHGs are pollutants and must be regulated by EPA, setting the stage for the 2009 Endangerment Finding.

It is particularly important that the 2008 Mass. vs EPA judgment explicitly rejected a de minimis argument — the notion that emissions from U.S. autos are too tiny to have a discernible effect on climate and thus are not worthy of regulation.

The court explained in Mass. vs. EPA:

A reduction in domestic emissions would slow the pace of global emissions increases, no matter what happens elsewhere. . . The risk of catastrophic harm, though remote, is nevertheless real. That risk would be reduced to some extent if petitioners received the relief they seek.

Thus, the Court clearly rejected the idea that a tiny magnitude of impacts avoided — the de minimis argument — is a valid argument against a finding of endangerment and the follow-on regulatory obligations. The Court explained:

Under the clear terms of the Clean Air Act, EPA can avoid taking further action only if it determines that greenhouse gases do not contribute to climate change or if it provides some reasonable explanation as to why it cannot or will not exercise its discretion to determine whether they do. [emphasis added]

In my understanding — remember that I’m a non-lawyer policy guy, caveat lector — the Supreme Court has already delivered its “best reading” of the CAA in Mass. vs. EPA. That precedent stands until overturned, even if subsequent Supreme Court rulings might seem to call it into question or an agency disagrees with the decision.

The Court rejected a de minimis argument to avoid regulation, however the argument was nonetheless used successfully in the 2009 Endangerment Finding and in its Technical Support Document to explain why EPA was not going to include water vapor in its basket of greenhouse gases to be regulated.

EPA in 2009 pulled a fast one. Water vapor is emitted in large quantities by human actions, including from automobiles, and these emissions have significant effects on local and regional climate. Scientifically, this is not at all controversial.²

In particular, research finds that auto exhaust contributes to these effects: “Combustion-derived water vapor (CDV) has significant impacts on urban climate and environment.” These effects have consequences for health and welfare. Under the logic of Mass. vs EPA, including water vapor in the list of regulated GHGs of the Endangerment Finding would have been a slam dunk.

My efforts to track down EPA’s 2009 justifications for excluding water vapor from its basket of regulated greenhouse gases quickly became comedic.

I started with the Endangerment Finding itself, specifically its Section IV.A.6.b:

The Administrator’s reasoning for excluding water vapor, was described in the Proposed Findings and is summarized here with additional information in Volume 10 of the Response to Comments document.

I dutifully went to Volume 10 of the Response to Comments:

The role of water vapor is discussed in Section 4(a) of the TSD, and in Volumes 2 and 9 of the Response to Comments document.

I chased the wild goose. The TSD and Volume 2 are silent on why water vapor was excluded from the Endangerment Finding.

I next went to Volume 9, which sent us back full circle, to the Endangerment Finding Section IV.A.6.b where I had started this search:

We disagree that the decision to exclude water vapor from the definition of air pollution in the Findings is arbitrary. Please refer to Section IV.A.6.b of the Findings for the Administrator’s rationale for water vapor’s exclusion from the definition of air pollution.

That was fun!

To the extent that the 2009 Endangerment Finding offers a justification for omitting water vapor from its basket of pollutants, it adopts a de minimis argument — That the effects of water vapor are too small to be worthy of regulation (emphases added):

Water produced as a byproduct of combustion at low altitudes has a negligible contribution to climate change³. . . The [IPCC] report also addressed anthropogenic contributions to water vapor arising from large scale irrigation, but assigned it a very low level of understanding . . .

However, as you’ve read above, the Supreme Court in Mass. vs. EPA rejected a de minimis argument. Yet the EPA in 2009 advanced one anyway to exclude water vapor. That exclusion has gone unchallenged ever since.

The omission of water vapor in the basket of regulated greenhouse gases is the weakest part of the 2009 Endangerment Finding, as it is totally inconsistent with EPA’s arguments about the inclusion of the other six gases.

As a practical matter, the regulation of water vapor would be a huge policy and political mess, imposing enormous costs across the U.S. economy. It would not be feasible, nor would it make policy sense.

I suppose that is why it was not included in EPA’s 2009 basket of gases. As a matter of simple politics — iron law and all that — the inclusion of water vapor would have stopped in its tracks the novel approach to GHG regulation conceived of in the 2009 Endangerment Finding.⁴

The EPA rescission ruling of last week argued correctly that water vapor should have been included in EPA’s 2009 GHG basket under the logic of the Endangerment Finding:

We further conclude that severing the endangerment and cause or contribution findings leads to untenable results and lacks any limiting principle. To illustrate the problem, the same logic would allow the EPA to issue emission standards for water vapor (H2O), another substance emitted by new motor vehicles and engines that is also considered a GHG. Considered in isolation, increased H2O concentrations in the atmosphere from all human activities can be said to endanger public health or welfare by resulting in rain that leads to slip-and-fall injuries, drownings, and damage to crops, livestock, and property, including through pools, rivers, and floodwater, although water vapor is not itself harmful and is necessary to sustain life. Also considered in isolation, CAA section 202 sources can be said to “contribute” to elevated H2O concentrations in the atmosphere from all anthropogenic sources, and these emissions of water vapor would thereby assertedly “contribute” to global climate effects similar to those attributed to other GHGs. CAA section 202(a)(1) does not contemplate prescribing emission standards for such an omnipresent, naturally occurring, and essential component of the ambient air because the text requires a unified analysis that ensures a nexus between the extent of contribution and the resulting danger. The logic of regulating water vapor appears absurd, but it is the same logic required to regulate GHGs under CAA section 202(a)(1). And the Administrator acknowledged in the Endangerment Finding that the statutory interpretation adopted in that action could support adding water vapor to the defined regulatory for “climate forcing” GHGs.

Is water vapor a pollutant under the language of the Clean Air Act?

Don’t take it from me, here are the guys over at Real Climate⁵ (emphasis added):

[T]he definition of pollutant in the Clean Air Act (CAA) clearly does include CO2 as well as water vapor . . . the issue is whether there is any evidence that these emissions might plausibly have adverse effects.

The Real Climate guys argue that water vapor’s effects are too small to be worth regulating, invoking a de minimis argument — exactly parallel to the 2026 EPA argument for why GHGs resulting from new vehicles do not meet the criteria of the CAA.

The Real Climate guys explain:

Basically, direct emission of water vapor has only a minimal impact on the increasing water vapor burden . . . Thus even if direct WV [water vapor] emissions were reduced to zero, that would be counteracted by a mere 0.05ºC 0.00005ºC change in global temperature.

Due to a math error, they crossed out that 0.05 C change in global temperature, but before correcting, they characterized an effect of that magnitude as “mere” and thus unworthy of regulation.

Compare what EPA said last week about the effects of GHG regulations on new vehicles:

[R]etaining a GHG emission standards program for vehicles and engines would result in a . . . 0.019 (0.012-0.027) °C impact on projected GMST through 2100.

If 0.05C is “mere,” then what is 0.027C? Ironically, on this point, Real Climate and the Trump administration’s EPA are in solid agreement.

I’ve often thought that a good way to get Congress to wake up and act on amending the CAA would be for an administration to promulgate regulations on water vapor, or alternatively, for some group to sue for water vapor’s inclusion in the basket of regulated GHGs — Thus forcing the issue.

Were that to happen, we’d see a bipartisan consensus emerge in no time on the need to update the Clean Air Act!

Interesting Postscript

Back in 2003, as GHG regulations were first being debated — well before the Endangerment Finding and Mass vs. EPA — I led a short piece in Science which raised this exact issue.

Admittedly, I was doing so in a way that drew upon the spirit of Jonathan Swift’s A Modest Proposal.

We focused on automobile water vapor emissions in response to an earlier article in Science arguing for a new commitment to hydrogen fuel cell vehicles. Here is what we argued:

Variation in water vapor affects local, regional, and global climates (1). Data on such effects are sparse because of complexities in the water vapor life cycle. However, our preliminary calculations indicate that a complete shift to fuel cell vehicles would do little to slow water vapor emissions, which presumably have increased perceptibly in some metropolitan locations through the growth in use of internal combustion engines. In some locations, changes in relative humidity related to human activity have arguably affected local and regional climate (2, 3). Depending on the fuel cell technologies actually employed, relative humidity in some locales might conceivably increase by an amount greater than with internal combustion engines. This increase could lead to shifts in local or regional precipitation or temperature patterns, with discernible effects on people and ecosystems.

The authors of the original piece responded with — what else? — a de minimis argument:

We are skeptical, however, that water vapor produced by combustion can have any important effect . . .

Not long after, I was up for promotion to full professor at the University of Colorado Boulder. A colleague told me (probably improperly), that a very high ranking official at the IPCC had written a letter to the university saying that I should be denied promotion to full professor due to this piece in Science, which was not helpful to the cause.

I was promoted anyway, becoming (I was told) at the time the youngest full professor on campus, and our 2003 argument has stood the test of time.

One quick request of you — Please click that “❤️ Like” before you go. More likes mean that THB rises in the Substack algorithm and gets in front of more readers. Thanks!

Comments, questions, insights welcomed! This is a complex legal and policy issue, and I invite everyone to participate in the discussion — experts and novices alike. I’d love especially to hear from constitutional and administrative law experts.

It is simple — You won’t find analyses like this one anywhere else. THB is made possible by its subscribers. So please consider a THB subscription at whatever level makes sense for you. I am appreciative!

1. It won’t take much begging. Environmental groups love to sue.

2. A scientific review of such effects goes beyond the scope of this post. Perhaps I’ll do one down the road. Interested readers can start with these articles:

On vehicle emissions of water vapor, some examples: Um, H. H. et al. (2007). Evaluation of the urban effect of long‐term relative humidity and the separation of temperature and water vapor effects. International Journal of Climatology: A Journal of the Royal Meteorological Society, 27(11), 1531-1542. Huang, X., & Song, J. (2023). Urban moisture and dry islands: spatiotemporal variation patterns and mechanisms of urban air humidity changes across the globe. Environmental Research Letters, 18(10), 103003. Liu, S. et al. (2022). Temporal variations of the contribution of combustion-derived water vapor to urban humidity during winter in Xi'an, China. Science of The Total Environment, 830, 154711.

More generally on athropogenic influences on water vapor and effects: Thiery, W. et al.(2020). Warming of hot extremes alleviated by expanding irrigation. Nature Communications, 11, 290; Yao, Y. et al. (2025). Impacts of irrigation expansion on moist-heat stress based on IRRMIP results. Nature Communications; De Hertog, S. et al. (2024). Effects of idealized land cover and land management changes on the atmospheric water cycle. Earth System Dynamics, 15, 265–288; Puma, M. J., and B. I. Cook. 2010. Effects of irrigation on global climate during the 20th century. Journal of Geophysical Research: Atmospheres 115.D16 (2010).; Ferguson, I. M., & Maxwell, R. M. (2012). Human impacts on terrestrial hydrology: climate change versus pumping and irrigation. Environmental Research Letters, 7(4), 044022.; Lo, M. H., & Famiglietti, J. S. (2013). Irrigation in California's Central Valley strengthens the southwestern US water cycle. Geophysical Research Letters, 40(2), 301-306..

3. Note how clever this language is — The EF says that water vapor emissions have a negligible effect on “climate change.” Whether or not that is true, the EF dodges the fact that water vapor emissions have well-documented effects on local and regional climate, with impacts on health and welfare, thus falling under the provisions of the CAA.

4. Of course, the fact that the logic of the 2009 EF leads to nonsensicle places does not mean that the regulation of carbon dioxide emissions is impossible — It just means that Congress would need to create legislation fit for purpose. I am soundling like a broken record . . .

5. One of the Real Climate authors of this comment is Gavin Schmidt, who was a government reviewer of the original 2009 Endangerment Finding.