Many of us have an intuitive sense that scientific claims to knowledge are inherently trustworthy, and, as outsiders to any subfield, it is intellectually illegitimate to contest such claims. Along those lines, it is common to hear that we live in an era of a war on science, with intentional campaigns and societal forces unjustly undermining expert authority on scientific knowledge.

On the other hand, we also have an intuitive sense that certain claims, even if grounded in science, are more contestable, especially when they carry a valence of politicization. Additionally, there are numerous examples of scientific and science-adjacent experts demonstrating a lack of trustworthiness and reliability. These include the “replication crisis” in psychology (where only 36% of published results could be replicated), high-profile reversals of government dietary advice on topics such as recommended fat intake, and COVID-era reversals of mask-wearing guidelines.

But as an outsider to any subfield, when is it legitimate to contest and debate scientific claims, and when is it not? Daniel Sarewitz and Steve Rayner, two scholars of the relationship between science and society, have discussed several dimensions that help to identify the trustworthiness of claims to expertise and knowledge. Two aspects I find particularly useful are 1) how feedback-rich and falsifiable the knowledge is, and 2) how high the stakes are in terms of broader value-laden ramifications.

Trustworthiness and reliability increase when knowledge producers or experts receive rapid, repeated, real-world feedback and clear performance metrics. Sarewitz and Rayner give examples of pilots and surgeons as “practitioner” experts who occupy this space: their reliability and credibility are derived directly from the large number of feedback-rich trials on which they have built their expertise. Sarewitz and Rayner contrast this with “inappropriate expertise,” where supposed experts may be credentialed, but they seek to advise on matters in which they have not been practically able to accrue a demonstrably reliable track record. An example may be a group of international relations PhDs advising the government on whether to go to war.

Trustworthiness and reliability decrease when stakes are high and values are contested because, in those cases, knowledge generation is highly susceptible to cultural forces and thus motivated cognition. The broader incentive structure and/or often unconscious preferences lead knowledge producers (researchers, scientific institutions, universities, etc.) to begin with a strongly preferred broader conclusion and then gather evidence to support it.

When assessing the general trustworthiness of scientific or science-adjacent claims to knowledge, I have found it useful to arrange these two ideas (slightly reframed and rephrased) in a two-dimensional space:

Below, I put “How testable is the relevant real-world conclusion?” on the horizontal axis.

I put “How strong is the pull toward a preferred conclusion?” on the vertical axis.

At the bottom of the vertical axis, there is a great deal of susceptibility to motivated cognition. Human beings generate scientific knowledge, and human beings are far from being purely rational robots. Thus, scientific knowledge is influenced by ethical intuitions, culture, and peer pressure surrounding researchers, as well as the incentive structure of the scientific funding and publishing systems (see Bias and Science section of Clark et al. 2023).

This means that there will often be pulls towards certain themes and broader conclusions over others, independent of the evidence. Since constructing research papers involves numerous researcher degrees of freedom (such as the research question to ask, the methodological specifics of how to address the question, and the focus of the abstract and title of the paper), there is significant latitude for selection biases to favor preferred broader conclusions over nonpreferred ones.

With these two dimensions, we can broadly characterize scientific claims to knowledge into four quadrants (I, II, III, IV).

Subscribe now

Quadrant I: Claims are reliably trustworthy and uncontroversial but relatively inconsequential.

Quadrant I is the purest realm of the scientific method, where knowledge producers are motivated solely by the pursuit of truth. Here, they engage in an iterative cycle of posing testable hypotheses, deriving predictions, confronting these with reproducible observations/experiments, analyzing the results, and revising or discarding the hypotheses accordingly.

Examples in Quadrant I include the science facts that students learn and practice demonstrating in grade school, such as how dissolving salt in water lowers the freezing point of the water.

However, claims native to quadrant I can extend far beyond archetypal grade school science. The scientific method can be broadly applied to a wide array of claims outside of what are traditionally thought of as scientific domains. In this sense, science is distinct from the subject being studied. Someone can accumulate extensive knowledge in astronomy, earning multiple degrees along the way, without actively practicing science. Conversely, a person working in digital marketing who uses controlled A/B hypothesis testing is engaging in the scientific method. In this case, the digital marketer might have a stronger claim to being a scientist than the person knowledgeable about astronomy.

Quadrant II: Claims are contestable, but controversy remains academic.

In Quadrant II, broader conclusions are not subject to definitive tests either in principle or due to practical limits related to, for example, spatial scale or timescale. Thus, trustworthiness remains provisional and is based on evaluating the internal consistency of claims, indirect evidence, or mathematical elegance.

At the same time, there’s little incentive for the knowledge producers to desire specific conclusions since the broader moral or sociopolitical implications are small. That being said, researchers producing knowledge might still have personal incentives to prefer certain conclusions. This can be because of an idiosyncratic affinity for a particular idea (i.e., a ‘pet theory’). It can also come about due to the inertia in one’s career that may make it practically difficult or psychologically painful to relinquish a long-held position on some academic dispute. For example, if a professor has spent two decades under a particular school of thought and built their “lab” around it, they are inherently going to be more sympathetic to results that support their paradigm.

An example in Quadrant II might be the hypothesized extra, compactified spatial dimensions of string theory. It’s a well-structured idea supported by formalism, but direct, definitive testing is challenging, leading the arguments to focus mainly on theoretical plausibility.

Quadrant III: Foundational claims are trustworthy, but controversy arises from different frameworks implicitly emphasizing different values.

In Quadrant III, core foundational claims and, to a lesser extent, broader conclusions are testable. However, controversy arises due to disagreements on which evidence deserves the most weight and which conclusions to emphasize. Due to incentives within academic publishing or the implicit preferences of researchers, there may be strong publication biases, where certain broader conclusions are sampled much more frequently than others for reasons other than scientific merit.

An example might be the claim that raising the minimum wage helps poor people (keep in mind I’m using a broad definition of “science”). That claim may seem straightforward enough to be tractable, but a problem is that “helps” is a value-laden verb and is interpretable in many different ways. Empirical evidence might show that raising the minimum wage increases wages for those low-paid workers who are actually employed, but has conflicting effects on broader hours worked, employment levels, or small-business survival. One might argue (with data and models) that a minimum wage narrowly raises incomes in lower terciles in the short term but stifles economic growth in the long term, eventually harming society overall, including poor people. Whether this policy “helps” depends on which outcomes are prioritized, the specific populations considered, and the time frame of concern. These debates may seem to focus only on the technical aspects of the question, but they also incorporate ethical and value disputes that are inherently political. These political disputes are not so much about fundamental facts but rather are about “who gets what, when, and how” [Lasswell (1958) via Pielke Jr. (2004)].

Quadrant IV: From the perspective of desiring neat, straightforward answers, Quadrant IV is a mess.

Claims are reliably controversial, contestable, and difficult to adjudicate because they are framework and model-dependent, difficult to test, embed value-laden assumptions, and are strongly susceptible to personally and culturally-influenced motivated cognition by the experts producing the knowledge. The same suite of evidence or underlying facts can be legitimately assembled into coherent narratives that seem completely at odds with each other, and there is no straightforward way to adjudicate which emphasis is “correct” (See Sarewitz, 2004 for a full academic treatment of this issue).

An example could be the claim that society should respond to a deadly pandemic by shutting down schools for a year. The counterfactuals are highly uncertain, objectives are numerous and incommensurate (mortality, learning, equity, livelihoods). Rapid, controlled, clear tests don’t exist, and many interests influence the desirability of different conclusions.

Share

The problem of conflating Quadrant IV with Quadrant I

It should be clear from the descriptions above that claims native to Quadrant I can be taken at face value and considered reliably trustworthy, whereas claims native to Quadrant IV are inherently contestable and may be quite dubious if not properly contextualized. Arbitrarily dismissing a claim from Quadrant I can legitimately be referred to as science denial (unless you have devised some revealing new hypothesis test), while challenging or debating claims from Quadrant IV, even without special expertise, should be fair game.

As the examples above indicate, Quadrant IV often overlaps substantially with what we tend to classify as political claims (with scientific or technical underpinnings). Crucially, as stated above, politics is about deciding “who gets what, when, and how” [Lasswell (1958) via Pielke Jr. (2004)]. Thus, politics is not very susceptible to the scientific method. What hypothesis test would definitely tell us how to value various tradeoffs properly and what any individual, entity, or group “deserves”?

Most of us in liberal democracies believe that these political decisions should not be made decisively behind closed doors by technocratic expert panels but rather be discussed openly in public debates, op-eds, and through democratic decision-making. This would all be well and good if the distinction between Quadrant I claims and Quadrant IV claims were delineated clearly.

However, there is a large and seemingly increasing desire for ostensibly scientific institutions and expert bodies to overreach and use the epistemic authority granted to science by the qualities of Quadrant 1 to attempt to make authoritative statements in Quadrant IV. In its most extreme form, this amounts to dressing up political opinions as if they were scientific facts.

It is, of course, alluring to present political opinions as scientific facts if you see yourself as a scientist or believe that science inexorably supports your political views. This approach makes your stance appear not just as another opinion but as a privileged one rooted in objectivity and rationality.

Examples are too numerous to catalog, but one illustrative one in this context would be the paper “Paris Climate Agreement passes the cost-benefit test” published in Nature Communications in 2020. I have no problem with doing this kind of work, per se, as I have published on this exact same question (though with a more descriptive and, dare I say, accurate title). However, I recognize that these kinds of publications can vastly oversell their claims to authority. They blur the lines of epistemic authority because they are written in the language of science, utilizing equations and a formal third-person academic style (which gives the impression that the knowledge is independent of the human authors who wrote the paper). They are also published in “scientific” peer-reviewed journals, which creates the impression that broader conclusions have been extensively vetted and can thus be considered more or less “true” (even if provisional and subject to refinement).

However, a little reading and critical thinking reveal that broader conclusions of papers like this rely upon layers of stacked assumptions and embed disputed ethical and value questions. They involve attempting to quantify and weigh incomparable outcomes across different people, groups, and species over space and time while simultaneously projecting future socioeconomic, political, and technological changes.

For example, how should one weigh the possible long-term extinction of warm water corals against near-term increased human energy access and thus poverty reduction in low-income countries? Attempting to convert this question into a numerical one amounts to false precision and the fallacy of misplaced concreteness. As David Roberts put it when discussing a similar kind of modeling:

“These are social and ethical disputes being waged under cover of math, as though they are nothing but technical matters to be determined by ‘experts.’ But social and ethical judgments should be made in an open, transparent way, not buried in models as inscrutable parameters. I mean, we’re talking about how much we value our children and grandchildren. Surely that’s a matter for democratic discussion and debate!”

This paper is not an exception. Many disagreements in climate science and its downstream policy implications are presented as mainly technical, suggesting they could hypothetically be resolved by more or better science (or by focusing on ‘good’ science and dismissing ‘bad’ science). However, many of these disagreements are actually ethical or moral, leading to the emphasis of different facts and thus divergent broader conclusions.

For example, conventional climate science and climate policy, as represented by the United Nations Intergovernmental Panel on Climate Change, are heavily guided by the underlying goal of the United Nations Framework Convention on Climate Change of “avoiding dangerous anthropogenic interference with the climate system.” This framework emphasizes, for example, the precautionary principle applied specifically to climate change over cost-benefit analysis of various energy systems and their alternatives, as well as the intrinsic value of an unchanging climate over a centering of the relationship between energy and human welfare.

On this point, it was telling that in the foreword of the recent controversial United States Department of Energy (DOE) report on climate change impacts, Energy Secretary Chris Wright was explicit about reorienting the moral framing of climate change to be about the net effect of fossil-fueled economic and technological development on people, writing things like “Climate change is real, and it deserves attention. But it is not the greatest threat facing humanity. That distinction belongs to global energy poverty.”

Nevertheless, combatants in these debates like to argue that only their opponents‘ conclusions are heavily influenced by moral frameworks (or more pejoratively, “ideology”), whereas their own conclusions spring inexorably from the underlying scientific facts. This was emphasized in the Department of Energy report with sentiments in the forward accusing mainstream climate science of recommending “misguided policies based on fear rather than facts” and it was emphasized in various retorts to the DOE report, where the claim was that ideological bias caused the report to get the underlying facts wrong, which then led it to the wrong conclusions.

This all gives the impression that if we could just eliminate “ideological bias,” then pure “science” would illuminate a direct path forward. Ultimately, though, these are Quadrant IV discussions where a shared set of facts can be marshaled to support arguments for diametrically opposed broader conclusions. There is no such thing as eliminating ideological bias because all prescriptive recommendations on a course of action rest on some contestable moral framework, and many of the most important claims are very difficult to test.

Thus, it would be clarifying to focus on surfacing the ideological and moral disagreements that drive the pull towards different preferred conclusions and acknowledging why claims are difficult to adjudicate. In doing so, it would be easier to recognize that these Quadrant IV claims are inherently contestable and will always be.

Subscribe now

By Patrick Brown

I was recently on two episodes of the PBS show Energy Switch, hosted by Scott Tinker, alongside Andrew Revkin. The topic was extreme weather, disasters, and climate change.

I highlighted that elevated greenhouse gas concentrations, primarily from the burning of fossil fuels, are causing an increase in heatwaves and coastal flooding as well as a decrease in coldwaves. However, global fire activity, global floods, global meteorological droughts, global hurricane activity, global mid-latitude cyclones, and severe thunderstorms (and their subhazards) aren’t changing at nearly the pace (or sometimes even in the direction) many people think.

Subscribe now

Meanwhile, climate-sensitive outcomes, such as per capita deaths from extreme weather or damage per dollar exposed, have declined over time due to economic and technological development. That also means that outcomes from extreme weather today are dictated much more by societal arrangements and background economic development than by the severity of the hazard itself.

I think there’s a huge gap between what you’d read in the IPCC Working Group 1 chapter on extreme weather and what you see in a lot of media.

What is put in front of our eyeballs is largely Earth’s pre-existing extreme weather hazards interacting with vastly increased exposure over time and then disseminated to us much more efficiently than ever before through the internet and algorithms. Additionally, a major part of the perception is that activists in the climate movement have found it very useful to connect extreme weather to elevated greenhouse gas concentrations as a tool of advocacy.

You can watch the full discussion here:

S7 Ep3: Extreme Weather part 1 | Our experts dig through IPCC data to reveal surprising trends in extreme weather.

S7 Ep4: Extreme Weather part 2 | Floods, fires and storms are flat or down, but risk and damages have increased.

Additionally, we received 24 questions beforehand (many of which had to be cut due to time constraints in the episode). I have provided written responses with references/links that viewers seeking more detail may find helpful.

1) Why should our viewers care about extreme weather and its connection to climate?

• Extreme weather is consequential: it costs global society about 0.2% of global GDP annually (ref.), causes 20 million people to be forced to move annually (ref.), and still causes 10s of thousands of deaths annually (ref.).

• We only hear about and care about extreme weather because societies are exposed to them (there are people and structures in harm’s way) and are vulnerable to them (these people and structures can be adversely impacted by the weather) (ref.).

• We should care particularly about the connection to climate change when elevated greenhouse gas concentrations have a very clear direct relationship with changes in the extreme weather, like in the case of hotter heatwaves (ref.), milder cold waves (ref.), and higher coastal flooding (ref.) due to higher sea levels.

• Most of the motivation to facilitate continued decreases in vulnerability is independent of changes in extreme weather (ref.): it is not necessary for hurricanes to be worsening to want to reduce our vulnerability to them.

2) What’s the difference between climate and weather?

• The climate is the average and typical variability of all weather variables at any given place (ref.).

• Elevated greenhouse gas concentrations and the warming that results from that affect all weather to some degree.

• Contemporary climate change generally refers to how those averages and variability are being affected by elevated greenhouse gas concentrations, and we observe this by tracking changes in weather statistics over the long term (ref.).

3) Let’s start with a recent event, the Hurricane Helene flooding in NC. What happened?

Summary

• In North Carolina, heavy rainfall in the days and weeks leading up to the event and topographically enhanced rainfall rates (Total 2-day rainfall estimates were in the range of 12 to 20 inches in a large swath of western North Carolina (ref.)) led to severe flooding, causing $60 billion in damage (ref.) and killing 230 people (ref.) (making it the deadliest US hurricane since Katrina).

Hazard change

• The rainfall from Hurricane Helene was made about 10% more intense due to the warming since the Industrial Revolution (ref.), so places that saw 20 inches would have seen 18 inches in a preindustrial climate.

• The flooding was very likely enhanced by warming, but that does not make this flood unprecedented in recorded history - floods in 1791 (ref.) and 1916 (ref.) were comparable in severity.

Exposure change

• The population exposed to floods is increasing: Since the 1916 flood, the population of North Carolina has increased by over 300% (ref.), and there has been substantial housing development in the hills that experienced mudslides (e.g., Buncombe County) since the 1980s (ref.).

Vulnerability

Bad

• Aspects of the situation that made North Carolina more vulnerable than an ideal situation were aging and deteriorating water infrastructure; the US Army Corps of Engineers had previously classified 39% of the region’s dams as only in fair, poor, or unsatisfactory condition (ref.).

• Flooding occurred in a region with a large number of less resilient trailer homes and manufactured homes (ref.).

Good

• Standard disaster preparation and response should not be taken for granted, as they demonstrate many of the reasons that high-income societies are so much less vulnerable to extreme weather than they used to be and why they are less vulnerable than low-income places today (ref.).

• The flooding was very well forecast many days in advance, which allowed for effective evacuations and allowed for the prepositioning of many disaster resources, such as rescue teams, emergency vehicles, and supplies like fuel, food, water, cots, and blankets (ref.).

• The Humane Society was active in providing shelter, supplies, and veterinary services for pets (ref.), which demonstrates the luxury of resources in modern high-income countries compared to the past and to lower-income countries in the world today (ref. ref.).

4) For viewers who don’t know, what is the IPCC? Can we trust their science?

The IPCC

• The core of the Intergovernmental Panel on Climate Change is a group of academics and scientists who, on a part-time basis, synthesize the scientific literature on climate change in major reports released about every six years (ref.).

• Working Group I focuses on physical climate science, including extreme weather, Working Group II focuses on impacts and adaptation, and Working Group III focuses on emissions reductions (ref.).

Trustworthiness

• The UN IPCC is intertwined with the UNFCCC treaty, whose goal is to “avoid dangerous anthropogenic interference with the climate system,” and thus, it takes as its goal to reduce human impact rather than optimize human well-being.

• The IPCC does not do a holistic assessment of the costs and benefits of various energy and agricultural systems, but instead is oriented around the negative impacts of elevated greenhouse gas concentrations (ref.) and pathways to reduce greenhouse gas emissions.

• Working Group I is pretty even-handed (the rest of this document references it extensively) but has a slight negativity bias, and high-level summaries can oversimplify.

• Working Group II has a major negativity bias, often taking positive situations, such as the dramatic increase in crop yields over time or the decrease in human deaths from extreme weather over time, and constructing conclusions that suggest things are getting much worse over time (ref. ref.).

5) How can we be confident that the climate is warming? How much of it is human-caused?

The climate is warming

A consensus of measurements (ref.)

• We can measure warming in the atmosphere with a global network of meteorological stations, with satellites, and with weather balloons.

• We can measure warming in the ocean with ships, buoys, networks of floats, and satellites.

• We can measure melting ice with satellites.

• We can measure rising sea levels with satellites and tide gages.

Human contribution (ref.)

• The best estimate is that all of the global warming since the industrial revolution is caused by elevated greenhouse gas concentrations from human activities because the main natural drivers of global climate change on these timescales (changes in solar activity and changes in volcanic activity) should have caused no change or slight cooling since the industrial revolution.

6) In a warming climate, does the IPCC show more heat waves? Where, with what effects?

Heatwaves

• The easiest way to conceptualize contemporary climate change is that all temperatures everywhere are warmer (over land, about 2°C or 3.6°F warmer), so heat waves are warmer, cold waves are warmer, and unremarkable temperatures are warmer than they would have been without elevated greenhouse gas concentrations (ref.).

• If a recent day was 105°F, a day with the same weather setup in a preindustrial climate would have been closer to 101°F.

• Heatwaves over land are warming more slowly than cold waves, so it’s not the case that heatwaves are becoming disproportionately hotter than other less remarkable temperatures (ref.).

Effects

• The two most prominent societal impacts of heatwaves are probably their detrimental impacts on crop yields (which is contested, ref.) and human mortality (ref.).

• Reductions in vulnerability have far outpaced increases in the heatwave intensity, and thus, the result has been improving outcomes over time: increasing crop yields (ref.) and decreasing human mortality rates (ref.).

7) Is warming impacting precipitation patterns and amounts? Where and with what effects?

Patterns and amounts

• Fundamental theory tells us that annual mean precipitation should increase by about 2% per 1°C (1.8°F) of warming and that extreme precipitation (daily, hourly rain rates) should increase by about 7% per 1°C (1.8°F) of warming (ref.).

• Observed changes in global annual mean precipitation are not apparent yet (ref.), and local trends vary by location, with some locations showing increases and other locations showing decreases (ref.).

• Observed changes in extreme daily precipitation are appearing, with 9% of global stations showing statistically significant positive trends and 2% showing statistically significant negative trends (ref.). 89% of stations have not reached the threshold of statistical significance because the noise of natural weather variability is much larger than the signal induced from background warming (ref.).

Effects

• Precipitation affects floods and droughts (more below)

8) Do we see trends in storms, hurricanes, and/or cyclones? What are they?

• Short answer: No

Hurricane hazard

• We do not observe clear trends in measures of global hurricane occurrence or wind speed over the past several decades (ref.). However, models suggest that hurricane wind speeds should increase at a rate of approximately 2.5% per 1°C of warming (ref.).

• There is good evidence of decreasing hurricane occurrence since the 1800s (ref.)

• When hurricanes do occur, they tend to produce approximately 10-15% more rainfall (ref.) and generate storm surges about 8% larger than similar storms in the 1800s (sea level has risen 9 inches since 1880 (ref.), and a medium/large storm surge is around 10 feet (ref.).

Hurricane exposure and vulnerability change

• We see disproportionate growth in population and economic value exposed to hurricanes on the coast (ref.), but at the same time, modern technology, infrastructure, and building codes can drastically reduce vulnerability (ref.) in terms of proportional damage done, and lives lost (e.g., cyclone-related mortality in Bangladesh has declined by more than 100-fold over the past 40 years ref.).

Mid-latitude cyclone (low-pressure system) hazard

• Fundamental theory tells us that reduced equator-to-pole temperature differences weaken the potential energy these systems tap into and result in a reduction in their wind speeds (ref.), and we do observe that and project it to continue (ref.), but we still expect to see an increase in the most extreme daily rainfall rates of 7% per °C (ref.).

• The reduction in wind speeds, combined with an energy system more reliant on wind, introduces a new type of extreme weather that was not consequential before: the wind drought (ref.).

Mid-latitude cyclone (low-pressure system) exposure

• The US population has risen more than 200% in the last century (ref.), so many more people than ever before are exposed to the negative impacts of mid-latitude cyclones.

Storms: Severe Thunderstorms.

Tornado hazard

• In the US, accounting for reporting biases, historically observed trends in tornado counts vary by location, with some locations showing increases and others showing decreases; however, there is no overall change, and a decrease in the most severe tornadoes may be observed (refs. wherein).

• Projections suggest a 5% increase in U.S. supercell-hours (a proxy for the tornado threat) under 3°C of warming (refs. wherein).

Hail hazard

• In the US, historically observed trends vary by location, with some locations showing increases and other locations showing decreases (refs. wherein).

• Projections indicate that there will be increased hail formation in clouds, but falling through a warmer atmosphere will cause increased hail melting on the way down, resulting in uncertain future change (refs. wherein).

Severe thunderstorm exposure and vulnerability change

• In the US, the number of structures and population exposed to severe weather has increased by 400% since 1940, but at the same time, increases in structural quality and forecasting have drastically reduced the loss of life from severe weather (refs wherein).

9) Do we see trends in droughts? In floods? What are they?

• Short answer: No

Drought hazard

• Historically observed trends in droughts vary by metric and location, with some locations exhibiting increases and others showing decreases (ref., ref.).

• There are not historical observed trends in global metrics of droughts like surface soil moisture, though there appears to be a modest increase in the amount of area under drought as defined by some metrics like the Palmer Drought Severity Index (ref.).

• Projections suggest some locations should experience an increase in drought as measured by surface soil moisture, and other locations should experience a decrease in drought, and these changes are locally on the order of 2% per 1°C of warming (ref.).

Flood hazard

• More locations show observed increases in extreme daily precipitation than decreases (ref.), and the expectation is that the most extreme maximum daily precipitation over land should increase by about 7% per °C of global warming (ref.).

• Decreased winter snowpack decreases rapid spring melt floods, and the greater capacity of drier soils to soak up water potentially offsets this increase in extreme precipitation such that there has been little observed nor projected increases in floods as quantified by annual maximum streamflow (refs. wherein).

• Projections indicate that some locations should expect increases and others should expect decreases, and overall, under a constant population scenario and current flood protection, the percent of the global population exposed to flooding increases from 2.2% annually to 3.2% annually under 3°C (5.4°F) of warming (ref.).

Exposure and vulnerability change

• Globally, the number of structures exposed to floods has increased 400% since 1985 (partially because population growth in risky areas even outpaces background population growth), but at the same time, flood control systems work, and lives lost from floods have seen a dramatic decline (refs. wherein).

• When it comes to flood impacts (or any type of extreme weather), the vast majority of people would rather live in a modern industrialized world with both the warmed climate and the flood protection that comes with it than live in a preindustrial world with a cooler climate and without the flood protection.

10) Do we see trends in wildfires and weather that encourages them? What are they?

Weather that encourages them

• We have seen a global increase in the metrics designed to measure weather conducive to wildfires (like fire weather season length and the intensity of the Fire Weather Index) due to warming causing drying, and in spite of reduced extreme winds in many locations (ref.).

Wildfires

• Globally, we have observed a decline in fires over recent decades as measured by both area-burned (ref.) and mass-burned (ref.).

• Much of the observed decline is due to many historically fire-prone savannas in Africa being converted to farmland and grazing land (though there is some agricultural burning, the net effect has been a decrease in fire activity) (ref.).

• Even in locations of high concern where fire activity has increased over recent decades, like the Western US and California, fire activity remains well below where it was hundreds of years ago (ref.).

• Globally, the effects of ignitions, land use, and land management can and have overwhelmed the effects of climate change (ref.).

Wildfire exposure

• California’s population has increased by 200% since 1960, so that’s many more people to light fires (ref.) and many more structures to burn down.

• There has been a 60% increase in structures in the wildland-urban interface in the US West over the past 30 years (ref.).

11) Are these weather and extreme weather events outside historical norms?

• Extreme weather happens infrequently by definition, and we only have very high-quality observations for a lot of weather events in the satellite era, which started around 1980 (45 years).

• Fuzzier pictures can be estimated further back in time, and when we do that, most things we see today are not without precedent.

• Warming may well be increasing the intensity of many types of extreme weather. Still, the signal from warming is much smaller than the noise of natural variability (ref.). Thus, we tend not to see obvious trends in extreme weather (ref.) the way we do with, e.g., CO2 concentrations (ref.) or global average temperature (ref.).

• What is certainly outside of historical norms is reduced societal vulnerability to extreme weather.

12) To recap, which weather events show correlation to warming today, and which do not?

• Elevated greenhouse gas concentrations and the resulting warming affect all weather to some degree, but the direction of change often depends on the location and is sometimes contested.

• The extreme weather that we can most confidently connect to background warming are warming of extreme heat (by 2°C over land, ref.), warming of extreme cold (by more than 2°C over land, ref.), increases in the height of coastal flooding due to sea level rise (9 inches since 1880, ref.) and increases in extreme daily precipitation (10% higher, ref.).

• Floods, droughts, tropical cyclones, extratropical cyclones, severe thunderstorms, and wildfires are much more complicated, with both historical observations and projections not indicating a universal direction of change (ref., ref.).

13) Is correlation causation? What else could be driving changes besides warming?

• Locally, trends in a lot of extreme weather can be dominated by natural random unforced variability (e.g., El Niño, La Niña cycles, Atlantic Multidecadal variability, Pacific Decadal Variability) on decade-to-decade timescales (ref.).

• Extreme heat (ref.), extreme cold (ref.), wildfires, and floods are all affected significantly by local land use.

• In terms of societal outcomes (e.g., crop yields, water availability, death rates, malaria rates), there are almost always much bigger drivers of change than climate change (ref.).

14) Do other climate scientists and climate advocates agree with the IPCC on weather?

• There is a large gap between the messaging of IPCC WG1 Chapter 11 on extreme weather (especially the sections on observations) and the message conveyed by those in the so-called climate movement who see the connection between elevated greenhouse gas concentrations and weather disasters as the main tool in their toolbox to drive political change (ref.).

15) What do mainstream journalists say about extreme weather?

• Presently, since climate change has become a part of the culture war, sources such as The New York Times (ref., ref., ref.), LA Times (ref.), and The Guardian (ref.) have a tendency to cover all bad weather through the lens of climate change.

• Interestingly, the Los Angeles Times does not have a weather tab in its top navigation bar, but they do have a climate tab (ref.). Climate change is slow on human timescales, so as a news organization with a mandate for high-frequency coverage, that structural choice pre-commits the organization to covering basically all dramatic weather through a climate lens, even when the influence from elevated greenhouse gas concentrations is small and contested.

• Journalists will often use phrases like “climate-driven,” “climate-fueled,” or “supercharged by climate change,” where they leave out the actual size of the effect they are claiming, and the result is that the language greatly exaggerates the magnitude of the changes we are seeing.

• A particularly egregious example of this is when coverage explicitly contradicts the mainstream consensus, as represented by the IPCC, such as when extreme cold is portrayed as becoming more extreme rather than milder (ref.).

16) Should journalists be saying something else? What should the public know?

• What you’re seeing is largely the preexisting extreme weather of Earth’s climate, in some cases enhanced from its 19th-century versions, interacting with a vastly expanded amount of society exposed to this weather, amplified by the much more efficient dissemination of disaster coverage via the internet and algorithms. What tends to be left out is the vast decrease in vulnerability in terms of economic loss per unit of value exposed and lives lost per person exposed (ref.).

• Journalists should report that extreme weather-related outcomes are driven much more by societal arrangements and economic development than they are by the severity of the extreme weather.

• Journalists should be more honest about how fossil-fueled technological and economic development are responsible for the tremendous declines in vulnerability that we’ve seen and thus honestly report on the potential negative consequences of overly restrictive energy policies, especially for low-income countries.

17) What is attribution science? Why was it created, and how does it look at events?

Attribution science

• Attribution science attempts to quantify what aspects of some observed weather or climate phenomena can be attributed to elevated greenhouse gas concentrations (ref.).

• Attribution necessarily entails that you construct a counterfactual, typically with a climate model: What would things look like had greenhouse gas concentrations not been elevated above their preindustrial levels (ref.)?

• Running physics-based climate models, with and without greenhouse gas enhancement, tells us that we can be very confident that we can attribute observed global warming to elevated greenhouse gas concentrations (ref.).

• The same logic can apply to extreme weather at specific locations. But, there is much more uncertainty because the signal of climate change is small compared to the noise of natural variability. Thus, you are often making attribution statements without having observed trends in the weather phenomena (i.e., without having detected any climate change (ref., ref.).

• Regardless, extreme weather event attrition studies, like those conducted by World Weather Attribution, tend to show intensity changes in the range of -10 to 25%. For example, World Weather Attribution calculated that during the January 2025 LA wildfires, the Fire Weather Index was enhanced by 6% (ref.), and during Hurricane Helene, the rainfall was about 10% heavier (ref.) due to elevated greenhouse gas concentrations.

Why was it created?

• The originally stated motivation was to be able to credibly sue fossil fuel companies (ref.), and that motivation still animates many in the subfield today (ref.), according to their public statements (ref.).

18) It seems like the range of temperature forecasts by the best models in the world is still very significant. Is it possible to attribute some portion of a weather event to human-caused climate change, given that uncertainty?

• Many times, there is not a model consensus on the direction of change, so you might have 20 models (which are not independent), and 15 say warming increased the intensity of this event, and 5 say warming decreased its intensity (ref.) and it’s a subjective value judgment whether that is sufficient evidence for you to believe that warming increased its intensity. That uncertainty is rarely communicated in the media.

19) Are some institutions and the media relying on attribution science? Why?

• The climate movement finds extreme event attribution very useful from a messaging standpoint. A lot of this reporting is funded by environmental philanthropy and catering to audiences that share those environmental commitments (ref.).

• Stepping back, we’re talking about something like 3°C warming this century when there is, for example,

  • A 30°C difference between the coldest and warmest country capitals in the world (ref., ref.).

  • A 20°C difference between winter and summer temperatures for mid-latitude cities (ref.).

  • A 20°C difference between the current global temperature and some of the warmest periods in Earth’s history, like that of 100 million years ago (ref.).

  • A 20°C local difference between mid-latitude temperatures in, e.g., St. Louis and what they were in the Last Glacial Maximum 20,000 years ago (ref.).

  • A 4.5°C difference between Boston and Philadelphia (ref., ref.).

  • A 1-2°C difference between urban and rural areas within the metros of large cities.

  • There has always been a tacit conversation that goes, “3°C doesn’t sound like that much; how can we help people understand that it is?” (With the assumption that 3°C just has to be large). Connecting elevated greenhouse gas concentrations to more extreme weather, even if the effect is small in magnitude, and/or somewhat tenuous, is perceived to accomplish this goal (e.g., the hurricane coming out of the smokestack on the cover of An Inconvenient Truth) (ref.).

• You can see this in the way the media reported on the LA fires. World Weather Attribution came out with their analysis saying that the Fire Weather Index was 6% higher, and you get headlines like “Yes, you can blame climate change for the LA wildfires” (Science News, ref.). The same analysis could have been used to support a headline of “Climate change a minor factor for the LA wildfires.”

• Some politicians like attribution science because it puts the blame elsewhere and diverts attention away from local failures of disaster preparation or management.

• Some industries like attribution studies, like Climate Service Providers (CSPs) that use them to sell their information, and the insurance industry that can use them to justify increased premiums.

20) Is there a more valuable way to measure the impact of weather events? (damages)

• Noting increases in absolute damages and implying that this is because the weather is becoming more severe is blatantly misleading because increases in damages are caused primarily by increases in value exposed (ref.).

• Both human and economic vulnerability are declining dramatically - global average mortality and economic loss rates have dropped by 6.5 and nearly 5 times, respectively, from 1980–1989 to 2007–2016 (ref.).

• The drop in economic loss rates is very intuitive when you think about it: a trailer home or a poorly constructed home in a low-income country might get completely destroyed by an extreme weather event, whereas if there were larger, more valuable, more resilient homes, they might incur more absolute damage but less damage as a proportion of the value exposed.

  • 100% damage on $20 of value is only $20, whereas 50% damage on $100 of value is $50.

21) What are the factors increasing damages from weather events?

• Exposure growth dominates, both in the background population growth and disproportionate growth in high-risk areas (ref.).

• For the vast majority of extreme weather, intensity changes are calculated to be on the order of -10 to +20%, whereas the change in exposure is often on the order of an increase of +100% to +400% over the past 50 years. Exposure change tends to be 10 to 100 times larger than physical changes in the weather.

22) Should this understanding of weather and damages influence our response to climate?

• First, trying to influence extreme weather outcomes with energy policy is incredibly indirect and delayed (ref.).

• Focusing on changes in extreme weather can distract from more direct on-the-ground solutions that have to do with mitigating increases in exposure and facilitating continued decreases in vulnerability to extreme weather generally (ref.).

• We can mitigate increases in exposure by smarter zoning and allowing insurance premiums to reflect real risk (not artificially suppressing premiums).

• We can facilitate continued decreases in vulnerability globally by encouraging economic development in low-income countries (refs. wherein), which does entail increasing greenhouse gas emissions in the near term (ref.).

23) Final thoughts?

• In the absence of some very sci-fi future with supreme geoengineering, extreme weather is inevitable, but the evidence suggests that economic development will continue to decrease vulnerability faster than extreme weather is changing.

• Outcomes are dictated much more by the economic and technological capabilities of societies than they are by the severity of the weather: Mortality from extreme weather has declined radically over time, and it is many times higher for low-income countries than for high-income countries (ref.).

• Most of the motivation to continue to foster decreased vulnerability is independent of changes in extreme weather (ref.). I am not sure it matters much in terms of decisions whether the rainfall from Hurricane Helene was made 10 or 25% more intense or whether it made tropical cyclones, in general, become more or less likely.

24) What gives you hope?

• The empirical history of human innovation, both on the greenhouse gas emissions reduction side (ref.), as well as the decrease in vulnerability (ref.).

By Patrick Brown, Head of Climate Analytics, Interactive Brokers

I recently departed from The Breakthrough Institute to assume the role of Head of Climate Analytics at Interactive Brokers. Leaving The Breakthrough Institute was very difficult for me, as I have greatly appreciated the intellectual environment and believe that emphasizing the principles of Ecomodernism offers significant value to the energy and climate dialogue.

One of the aspects I valued most about being at Breakthrough was that it provided me the freedom and the opportunity to write about issues that I believe are pernicious and widespread in the development of knowledge on climate change.

In particular, I have written extensively about how social and professional incentives in academic climate science create self-reinforcing feedback loops that skew the overall output of academic literature toward negative predictions regarding future climate impacts while downplaying the potential adverse effects of restrictive energy policies. Essentially, although the ideal scenario is to have “evidence-based policy," I’ve argued that the reality involves a significant amount of “policy-based evidence," where various selection effects lead prudent researchers to frame their studies and results in ways that support their preexisting policy preferences, such as the temperature limits established by the Paris Agreement. Further, even when there is no such bias from the researchers involved, there is a strong tendency for scientific models to become “spherical cows” where the most relevant dynamics of a problem are assumed away or held constant because they are the most difficult to model.

However, identifying problems is much easier than solving them. I have felt the itch to work on projects that would improve the situation, but I have been unsure where my efforts would best be placed.

It was at this point that Thomas Peterffy began recruiting me to Interactive Brokers to work on prediction markets in the climate and energy sectors. While prediction markets may seem unconventional (much more on that below), I view them as a genuine opportunity to tackle many of the issues I have highlighted and ultimately to create some of the clearest representations of collective human knowledge regarding our energy and climate future.

The Problem of Prediction

Making predictions is a ubiquitous and crucial human activity—a part of decision-making in all walks of life, including business strategy, politics, and science. However, it has always been and continues to be remarkably challenging to do well.

For many people, the default strategy for obtaining the most reliable predictions is to seek out experts with specialized knowledge on the topic at hand. The assumption is that in-depth subject matter knowledge correlates well with forecasting abilities and that subject matter experts agree with one another, as they are presumed to be logical and rational, basing their reasoning on agreed-upon facts.

However, despite most people's intuition, expertise often fails to yield accurate predictions in practice. Studies have shown that confident, high-profile designated experts frequently perform no better than chance and, in many instances, worse than simple statistical models in their predictions. Forecasting errors arise from various sources, including cognitive biases, narrative fallacies, and the tendency to perceive patterns in randomness. Long-term prediction is particularly challenging due to path dependence, which causes small initial errors to accumulate over time and allows low-probability, high-impact events to shape the course of history. One challenge, therefore, is not merely to predict more effectively but to acknowledge the structural limits of prediction itself and design systems that explicitly recognize and incorporate inherent uncertainty rather than artificially side-stepping it.

Many of the same issues arise when it comes to expert predictions on matters specifically related to energy and climate. Well-understood fundamental physics makes certain aspects of climate change predictable, such as the global average temperature projected over decades in response to known changes in greenhouse gas and aerosol emissions. However, when economic and technological dynamics come into play, as they inevitably do, predictions can often be consistently and embarrassingly inaccurate.

But what is the alternative? One option might be to let the private sector evaluate climate risk. There has certainly been a significant increase in private climate consulting firms, or climate service providers, in recent years, many of which employ expert PhDs in energy and climate science. Theoretically, competition among firms should incentivize accuracy and lead to the distillation of the best information. However, these entities do not necessarily have the right incentives to provide the most accurate information either. They are also motivated to exaggerate the extent of change we are witnessing (and the accuracy of their predictive abilities) in order to sell their proprietary information. Competition should provide a check on this but a major problem with both academic and private climate service provider predictions is a lack of accountability due to the difficulty in verifying claims. This is partly because the predictions tend to be very difficult to falsify as they embed so many assumptions that are not considered to be part of the prediction.

In fact, the Intergovernmental Panel on Climate Change (IPCC) goes out of its way to protect itself from falsifiability by stating that it rarely makes climate predictions but instead makes climate projections that are based on assumptions about, for example, “future socio-economic and technological developments." The IPCC does not make probabilistic claims about these assumptions and simply states they “may or may not be realized.”

When forecasts are unfalsifiable, users of forecasts “...are not prepared to pay for quality they cannot verify, and sellers are unwilling to invest in quality they cannot demonstrate. Under these conditions, it is more rational for forecast providers to focus on presentation and the user-friendliness of their portals than on the accuracy of their forecasts.” (Roulston and Kaivanto, 2024). This applies to both private climate service providers and the production of academic publications, where there are strong incentives to spend more effort on the clean, aesthetically pleasing presentation of results (both visually and linguistically) than on improving unverifiable accuracy.

Therefore, three keys to improving the state of predictions are

1. Not blindly relying on perceived experts.

2. Ensuring that predictions are verifiable and thus falsifiable.

3. Motivating prediction producers to be as accurate as possible.

Prediction Markets for Climate and Energy

It turns out there is a mechanism that goes a long way toward meeting all three of the above criteria: Prediction Markets. Prediction markets enable participants to buy and sell contracts based on their expectations of the outcomes of well-defined events.

Prediction markets are not a new concept or something emerging from outside academia. Numerous studies have examined their accuracy, and they even have their own dedicated academic journal. This body of research demonstrates that prediction markets have consistently performed as well as or significantly better than alternatives, and they have even been used to identify which academic studies are of low quality.

Fundamentally, prediction market probabilities aggregate the collective wisdom of participants in the market, and it turns out that this crowdsourcing of information often provides our best representation of reality—in many cases better than the views expressed by designated experts. This is a main theme in James Surowiecki’s 2004 book, The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies, and Nations, as well as Cass Sunstein's 2006 book, Infotopia: How Many Minds Produce Knowledge. Perhaps the most famous antidote of the wisdom of crowds, detailed in Surowiecki’s book, is Sir Francis Galton's discovery in 1906 that the average estimate of an ox's weight made by 800 villagers was almost perfectly accurate and better than any of the individuals' guesses.

While crowds possess the potential for wisdom, Surowiecki articulates that the phenomenon of crowd wisdom materializes only when specific conditions are satisfied. These conditions include that individual contributions should be: 1) decentralized, enabling individuals to leverage local knowledge and unique information; 2) diverse, incorporating unconventional and eccentric viewpoints; and 3) relatively independent, ensuring that opinions are not overly influential upon one another.

When the conditions outlined by Surowiecki are not met, crowds do not exhibit the traits of wisdom but instead produce popular delusions and the madness of crowds. As a species that has relied on social cohesion for survival throughout our evolutionary history, humans are inherently predisposed to being heavily influenced by our peers. Thus, social contagion, information cascades, and a herd mentality can make group behavior much less logical than the reasoned judgment of any individual within the group. (Note that my ongoing references to Wikipedia represent a tacit endorsement of the crowdsourcing of information).

The situations where groups cannot be trusted arise when they are overly socially connected, which both stems from and contributes to a lack of diversity and independence. This presents a challenge for expert knowledge production, as designated experts often come from a relatively small number of elite PhD programs and spend their careers within rather insular groups. Moreover, the synthesis of academic research is frequently conducted by committees working closely together, which can further encourage herding and orthodoxy, especially when social incentives favor signaling that an individual expert belongs to a particular epistemic team rather than prioritizing the production of the most accurate information or predictions.

Prediction markets address these incentive problems by forcing clear questions and directly compensating participants for their predictive accuracy. They also empirically identify true experts through trial and error. One challenge of designating experts is that credentialed individuals can present plausible arguments on all sides of an issue, leading to the inevitable question of who the true experts are. Prediction markets select these people, not based on credentials and name recognition, but based on performance: those who continually make bad predictions will be incentivized to find ways to improve or exit the market. Furthermore, the financial incentives encourage all participants to invest time and resources into enhancing information and predictions.

The famous bet between economist Julian Simon and biologist Paul Ehrlich in 1980 on the future prices of five metals to settle a dispute over their differing views on human innovation and its effects on resource scarcity is an early example of quasi-prediction markets at work in the study of climate and energy. The bet shares characteristics with prediction markets in the environmental realm as it compelled Simon and Ehrlich to avoid vague proclamations (or academic obscurantism) and replace them with concrete claims or specific, measurable terms. Moreover, putting financial skin in the game ensured that there was a true penalty for being wrong and thus was high motivation for Simon and Ehrlich to put forth their best understanding of reality. The negotiation of terms between just two people with divergent views reveals some information, but prediction markets scale these dynamics by aggregating hundreds, thousands, or even millions of views.

So far, pilot programs in climate prediction markets have shown great promise, demonstrating accurate predictions on variables as diverse as UK monthly rainfall, average daily maximum temperature, annual wheat yields, monthly El Niño indices, annual Atlantic hurricanes, and annual US hurricane landfalls.

At Interactive Brokers, I will work on both high-level event questions of general interest and more local, specific event questions that have more direct applications to participants.

At a high level, for example, we will have event questions such as, “Will the annual global temperature breach the 2°C Paris Agreement temperature limit by 2040?” At first glance, this may seem like a narrow scientific calculation that is best left to experts, but initial impressions can be wrong. First, there is significant uncertainty in just modeling the physical climate system, so there is value in aggregating diverse opinions on that question alone. Second, answering this event question actually goes well beyond physics and requires information on the future trajectory of economics, technology, demographics, and geopolitics. It is thus a quintessential example of a question that would benefit from aggregating diverse information dispersed broadly across society in a prediction market.

The vision is that the market opinion on big-picture questions like these will provide the best available information to the public and decision-makers and help people understand where and how strong consensus is on various topics.

It democratizes participation so that those on both ends of the spectrum on climate change questions can no longer claim that their views are being marginalized. Those who are dismissive of the human origins of contemporary global warming are allowed to participate just as much as those who foresee the near-term collapse of society. Additionally, those within the climate movement should be supportive of the project, as research has shown that participating in climate prediction markets increases concern about global warming.

In the context of more local, practical applications, I hope that these markets will mitigate climate vulnerability by facilitating a more efficient resource allocation. For instance, if a contract indicating that “California will experience more than 1 million acres of area burned” in the coming year presents a significant probability prior to the commencement of the fire season, state and federal agencies could respond by increasing their hiring of seasonal firefighters or contracting for additional equipment.

These markets can also have the attribute of being extremely reactive to new important information. Thus, if the market for “Will a major hurricane make landfall in Miami-Dade County in 2025” shifts dramatically in the hours following a storm's formation, that information could be used to make decisions on evacuations, emergency declarations, and the prepositioning of disaster resources. These markets would not be alternatives to existing hurricane models but rather a way of combining information from multiple models (e.g., from traditional numerical weather prediction models with newer models like the GraphCast AI model created by Google’s Deepmind) via human minds into a single probability.

Like all human endeavors, there is a risk of bad actors manipulating the system for their own gain. However, singling out prediction markets as particularly susceptible to influences that undermine their predictive power is misguided, especially when compared to alternatives. For example, traditional climate information is shaped by a complex web of funding incentives, peer review gatekeeping, and institutional pressures that create their own types of bias—biases that are hard to detect and rectify. In contrast, climate prediction markets are open and decentralized, with prices determined by a broad range of independent participants rather than a small group of gatekeepers with aligned incentives. These markets are also no more vulnerable to manipulation than traditional financial markets, where significant sums of money are at stake, yet prices remain highly efficient. Critically, any attempt to artificially distort climate prediction markets presents a profit opportunity for informed participants, who can exploit and correct such manipulation by taking the opposite position. This self-correcting mechanism means that, over time, prediction markets should provide information that is more robust and resistant to bias than the alternatives.

Overall, prediction markets offer a way to discipline thinking by defining concrete event questions and they empirically select the best forecasters while continuously incentivizing improvements in accuracy. They also provide a means of aggregating diverse, widely distributed knowledge into a single probability or price.

I believe that scaling up these prediction markets in the climate and energy sector will bring clarity that has often been lacking from both academic and private sector climate predictions, ultimately leading to a better understanding of the world and more effective solutions to problems.

Subscribe now

Wildfires in California have taken center stage once again in 2025 and have inspired diverse arguments about wildfire causes and, thus, potential solutions. The influence of climate change often takes center stage in such discussions, even if it takes attention away from more immediate on-the-ground adaptation strategies that have a much more direct impact than greenhouse gas emission reduction policies.

Proactive fuel reduction treatments such as mechanical thinning and prescribed burning are one such adaptation strategy relevant to wildfires in California and western North America more generally. These treatments aim to reduce fuel loads to lower fire intensity and severity, thereby reducing their impact. We know that these treatments work under most circumstances, but thus far, there has been little research on whether these strategies work well enough to make a major difference as the climate continues to warm.

We recently published research focusing on just this question and found that there is substantial potential for fuel reduction to reduce wildfire intensity in California even as it continues to warm:

• Brown, P.T., Strenfel, S.J., Bagley, R.B., Clements, C.B. The Potential for Fuel Reduction to Reduce Wildfire Intensity in a Warming California (2025). Environmental Research Letters, 10.1088/1748-9326/adab86

I recently gave a short technical presentation on this research at the American Meteorological Society’s annual meeting and a longer, less technical presentation at Columbia University.

Background

Fire activity in western North America and in California used to be much more prevalent than it is today and probably reached a long-term minimum in the latter part of the 20th century.

Below is a record of fire occurrence from Swetnam et al. (2016) for a network of more than 800 sites in western North America, which shows the precipitous decline in fire frequency starting at the end of the 19th century.

Similarly, the findings of a more recent and more comprehensive paper, Parks et al. (2025), are summarized in the schematic diagram below that tells the same story:

As you can see in the above figures, fire activity hit a trough in western North America but has been ticking up over the past several decades. Perhaps ironically, one of the main causes of the current increase in fire activity is the antecedent decrease in fire activity caused by humans. The photos below from Graham et al. 2004, illustrate what is going on.

Prior to European colonization, indigenous people, along with natural fire regimes, had historically caused fires to occur roughly once per decade in many forests in western North America. This continuously cleared out much of the surface vegetation and smaller trees to the point where covered wagons could routinely traverse these forested landscapes when the Europeans arrived. However, by 1912, the US federal government halted these practices entirely, and in 1935, the “10 a.m. rule” was introduced, requiring that all fires be extinguished by 10 a.m. the following morning. This fire exclusion was initially successful but led to a buildup of vegetation, or “fuel,” in the forests. The map below, from the California Wildfire Taskforce, shows that much of the Sierra Nevada region in California, for example, has not seen fire in over 70 years, which is a major departure from its traditional fire frequency.

Once continuously warmer and drier conditions arrived on top of this fuel buildup, fire exclusion became less effective, and fires became more difficult to contain, leading to an increase in catastrophic uncontrollable fires since the 1980s.

The potential for fuel reduction to reduce wildfire intensity in a warming California

So, what’s the solution? One key approach is to bring conditions back to something closer to where they were in the 19th century by intentionally setting fire to the landscape under favorable conditions and also removing fuel mechanically. This results in less fuel to burn when wildfires occur, making them less severe and easier to contain.

Decades of on-the-ground firefighting experience show that if a forest is overstocked and untreated, a fire can easily escalate, but thinning and prescribed burning can limit fire intensity and ultimately help protect the forest and adjacent communities.

The above diagram represents the idealized theory for how this works, but there is also ample empirical evidence that it works. This is exemplified by meta-analyses of dozens of studies, but a picture says 1,000 words. Below is a photo of the after-effects of the Bootleg Fire in Oregon, which happened to traverse an area that had no fuel reduction, an area that had recently undergone only mechanical thinning, and an area that had recently undergone both mechanical thinning and prescribed burning.

We can see that the fire was most severe in the area with no treatment; it was less severe where there was only mechanical thinning, but the forest survived relatively unscathed where both thinning and prescribed fire had recently been implemented.

So we know these practices work, and we know their general direction of effectiveness, but our recent research focused on just how well they work, specifically in terms of offsetting the impact of climate warming over the remainder of the century.

Methods of our study

The starting point for our research was over 27,000 satellite observations of fire intensity within California state lines spanning the period 2012–2020, and our goal was to understand the relative influence of both fuel and temperature on these observations.

There were three fundamental steps:

1. Learn the relationships between environmental conditions (including both fuel and climate conditions) and wildfire intensity

2. Extrapolate those relationships in space

3. Make future projections under different warming and fuel reduction scenarios

First, we learn the relationships between observations of wildfire intensity and sixteen different environmental conditions. Many traditional regression methods assume that the influence of any predictor variable is independent of the influence of the other predictor variables and that their influence is monotonic, if not linear. However, it is well known that the influence of any one of these predictor variables will be highly conditional on the state of other predictor variables. Thus, rather than use traditional regression methods, we use machine learning models, specifically neural networks and random forests, in order to estimate the associations between the environmental conditions (or predictors) and wildfire intensity.

We confirm that these predictors do constrain wildfire intensity by testing the models on data that was held out of the training procedure.

After training the machine learning models, we extrapolate in space. Given a set of environmental conditions, the models predict what wildfire intensity would look like if a fire were to occur there at that time. Because we consistently have environmental data, we can produce maps of predicted wildfire intensity or “fire intensity potential” (FIP) maps. Below is an example from the afternoon of September 8, 2020, when large wildfires were erupting across California. This map shows where the model predicts the highest risk during that snapshot in time.

To create future projections, we placed the historical weather snapshot maps like the one above in different combinations of background warming and fuel reduction conditions to assess how changing the fuel and climate predictors might influence future statewide fire intensity potential. The background warming scenarios correspond to climate-model calculated warming from two greenhouse gas emissions scenarios—a slow emissions reductions scenario (SSP2-4.5) and fast emissions reductions scenario more in line with the Paris Agreement (SSP1-2.6).

When no fuel reduction is implemented (the top red and orange lines) we see that fire intensity potential goes up over the remainder of the century as conditions grow warmer and drier. However, the difference between the emissions scenarios isn’t very large. By mid-century, under the slower emissions reduction scenario, fire intensity potential would be about 14% higher than today, whereas faster emissions reductions would lower that increase to about 12% higher than today. By the end of the century, the gap grows a bit more, to around +25% versus +16% relative to today.

But if we scale up fuel reduction, as represented by the purple and green lines, that creates a much bigger effect. Conducting fuel reduction on 1.6 million acres per year (with a 5-year return interval), which represents 3% of the total domain, would cause a 12% reduction in fire intensity potential in 2050 relative to today and would still cause a reduction in fire intensity potential at the end of the century even under slow global greenhouse gas emissions reductions. Given that 1 million acres per year is the stated near-term goal in California, 1.6 million acres per year is within the range of plausibility. This is good news, and it is information that could be used to support these goals and facilitate their implementation in the face of multiple challenges, including funding constraints, workforce shortages, and bureaucratic and regulatory obstacles associated with the California Environmental Quality Act (CEQA) and the National Environmental Policy Act (NEPA).

Does this kind of research get enough attention?

This information, however, cannot be used to support these goals if it is not noticed.

This work represents the most recent iteration of research that I first published in Nature in September 2023. After publishing that paper, I wrote an op-ed critiquing that study and articulating my broader concerns about how high-profile, attention-grabbing climate science often focuses too narrowly on the most dramatic negative impacts of climate change (more here and here) at the expense of more actionable science on adaptation.

My core thesis was that customizing the research for a high-impact journal made it less useful than it would have been otherwise. Specifically, I pointed out that it was in my self-interest as a researcher to focus narrowly on the impact of climate change on a particular processed metric without considering other relevant dynamics that could offset the impact of climate change.

In much of the pushback that I got, I heard that I was wrong on this account and that, had my projections included more driving factors than just climate warming, it would have been even more attractive to a high-impact journal like Nature, not less.

The present paper represents a test of this hypothesis.

You’ll notice that the present paper was not, in fact, published in a high-profile venue, but this was not due to a lack of trying. The paper was rejected by Nature, as well as the other high-impact journals of Science, PNAS, Science Advances, and PNAS Nexus. Only once did it go to peer review in this series of submissions, and the rest were ‘desk rejected,’ meaning the editor subjectively decided that the paper was not worthy of their high-profile venue.

Is the reason that the present version landed in a lower-profile journal because it is somehow less adequate on technical grounds? No. In fact, this version of the research represents an improvement along at least seventeen dimensions, including more data, higher quality data, higher resolutions in space and time, and more sophisticated and elegant methods. The table below summarizes the improvements between the previous Nature paper (Brown et al. 2023) and the present Environmental Research Letters study.

So why was this version of the research more challenging to publish in a high-profile venue? I think the key difference lies in the framing. In accordance with what I argued in my piece in the Free Press (and elaborated on in the Chronicle of Higher Education), the path of least resistance to publishing high-impact climate science papers involves framing the research in a way that at least directionally supports global climate policies like the Paris Agreement:

“…the biases of the editors (and the reviewers they call upon to evaluate submissions) exert a major influence on the collective output of entire fields. They select what gets published from a large pool of entries, and in doing so, they also shape how research is conducted more broadly. Savvy researchers tailor their studies to maximize the likelihood that their work is accepted. I know this because I am one of them.

Here’s how it works.

The first thing the astute climate researcher knows is that his or her work should support the mainstream narrative—namely, that the effects of climate change are both pervasive and catastrophic and that the primary way to deal with them is not by employing practical adaptation measures like … better forest management or undergrounding power lines—but through policies like the Inflation Reduction Act, aimed at reducing greenhouse gas emissions.

So in my recent Nature paper, … I focused narrowly on the influence of climate change on extreme wildfire behavior. Make no mistake: that influence is very real. But there are also other factors that can be just as or more important, such as poor forest management”

The previous Nature paper (which focused narrowly on the negative impact of warming on fires) received considerable attention (even prior to my essay critiquing it) as it was covered by over 100 news outlets, including NPR, The LA Times, The San Francisco Chronicle, Yahoo News, the Japan Times, and The Times of India. This was mainly because of the built-in infrastructure that Nature has in place to widely and effectively disseminate the research they publish. Environmental Research Letters, the venue of the present paper, has no such infrastructure, and thus, the result has been that the present version of this work has been covered by a grand total of zero news outlets (as of February 2025).

So, the subsequent exclusion of the present paper from the high-impact literature is evidence in favor of my thesis that editors and reviewers mold the output of the high-impact literature to emphasize the influence of climate change over other (oftentimes more important) factors.

Which venue a paper is published in does not necessarily make a huge difference to the handful of other researchers within a sub-discipline (because they will tend to find it regardless), but it makes a huge difference for the amount of attention research receives by the public and, thus, how much the information becomes incorporated into the conventional wisdom of decision-makers.

The overall result of these dynamics is that the public and decision-makers receive the message that the main way to solve climate-related problems is through greenhouse gas emissions reductions rather than more direct on-the-ground adaptation strategies.

This is unfortunate because adaptation measures often have much higher leverage on improving outcomes, but they will be easiest to implement when they are appreciated as being effective.

By Patrick T. Brown

The United States Global Change Research Program’s (USGCRP) National Climate Assessment (NCA) could serve as a crucial resource for objectively and comprehensively understanding the impacts of climate change. However, in its current form, the framing, methodology, and communication strategies employed by the NCA make it far less useful than its potential. It is currently (correctly) perceived as a political document, and thus, the Trump administration is likely to either overhaul or scrap it. Ideally, the NCA could be transformed into a document recognized as valuable by both sides of the political aisle instead of merely reflecting the perspective of the party in office at the time of its creation. Below are some brief recommendations for producing such a document. It would be seen as less political, more resilient to administrative changes, and provide clearer, more balanced, and more useful information for decision-makers and the public.

1. A Broader Framing of Research Questions

Seeing Climate Change in Terms of Both Threats and Opportunities

The current NCA presents climate change almost as if it is categorically detrimental, but there is no physical law dictating that a warmer climate and elevated CO₂ levels are exclusively harmful. To take just one example, higher temperatures and increased CO₂ may pose threats to U.S. corn yields, but they can also create opportunities for boosting U.S. wheat yields. Similarly, warming will increase adverse health outcomes due to hot temperatures but decrease adverse health outcomes due to cold temperatures. An even-handed and useful assessment should evaluate the full spectrum of impacts on both the positive and negative sides of the ledger without being overly focused on seeking out and highlighting negative impacts.

Total Risk vs. Additional Risk

The current NCA tends to focus narrowly on “additional climate risk,” which can leave false impressions by not putting the impact of climate change in the context of other relevant causal factors. Using a “total risk” framework, as recommended by former NCA and IPCC author Brian O’Neill, solves this problem and avoids the confusing rhetorical ambiguity where the direction of a trend is conflated with the direction of an impact. Placing climate impacts in the context of other relevant factors—such as technological innovation, economic and social changes—enables decision-makers to see a more complete view of the state of the problem.

Share

2. Accounting for Exposure and Vulnerability

Going Beyond Physical Hazards

When it comes to understanding the economic and human impacts of extreme weather, the NCA should emphasize that disaster risk results from the interplay of three key elements:

• Physical Hazard: The frequency/intensity of physical weather and climate events (e.g., wildfires, extreme temperatures, precipitation events, droughts, hurricanes)

• Exposure: The amount of people and assets in harm’s way.

• Vulnerability: The susceptibility of people and assets to damage.

Previous NCA reports centered rising disaster costs with the implication that this was almost entirely due to increases in hazard frequency and intensity. However, increases in exposure likely dominate rising disaster costs, and thus, a useful report would acknowledge this. A NCA that systematically examines changes in exposure and vulnerability, in addition to changes in hazards, would provide a more accurate representation of total risk and thus provide more useful and actionable information to policymakers.

Acknowledging Adaptation and Resilience

Humanity has had tremendous success adapting to our often hostile climate but is nowhere near sufficiently adapted to its hazards (if it were, hurricanes, for example, would do no damage). Given this, the NCA should consider identifying sectors of society that would disproportionately benefit from increased resilience independent of the degree to which relevant hazards are changing. Enhanced infrastructure, early warning systems, and better urban planning can mitigate damage regardless of whether or how hurricanes and other hazards are increasing in frequency or strength. Identifying these gaps—sectors that would benefit the most from adaptation efforts—should be a priority within the NCA.

3. Synthesizing Knowledge with Systematic Reviews

Using Formal Guidelines for Literature Assessment

Assessment reports should attempt to avoid approaches where the author's subjective judgment, along with their social/professional affiliations, have a major influence on the narrative being presented and the research that is cited or promoted. Implementing established review frameworks—like the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)—can bring rigor and transparency to the research selection process by explicitly cataloging search terms and criteria for inclusion/exclusion (credit to Roger Pielke for this suggestion). Subsequent discussions that employ subjective analysis and value judgments regarding studies can follow afterward.

Systematically Delineate Between Different Kinds of Knowledge

Every section should conduct a systematic review of 1) observations/trends, 2) theoretical frameworks, and 3) modeling, accompanied by a thorough cataloging of the sensitivity of results to various observational datasets and models, as well as a discussion of competing theoretical explanations. The IPCC Working Group 1 Chapter 11 follows a format similar to this.

Addressing Publication Bias

The assessment should explicitly address potential selection effects within the published literature, particularly focusing on the concept of Publication Bias (commonly referred to as the file drawer effect). It is essential to evaluate whether the published literature accurately represents an unbiased sample of our comprehensive understanding of each topic. Should it be determined that the published literature is likely biased in a particular direction, the report should consider conducting fundamental original analyses aimed at uncovering null results or results that may not have been organically revealed through the research community’s inherent selection choices.

4. Improving Communication Style

Description Over Persuasion

An assessment aimed at informing should focus on presenting high-level observational metrics with as little pre-processing as possible before moving on to potential climate impacts. This style is exemplified by the Bulletin of the American Meteorological Society’s (BAMS) “State of the Climate” reports. Thus, prior to a discussion of the climate impacts on crop yields or health outcomes, changes in raw yields and health outcomes over the past should be shown.

Moving Away from Headline-Style Key Messages

The current “Key Messages” in the NCA often read more like calls to action than neutral statements of scientific findings—for example, “Climate Change Is Harming Human Health” “Systemic Racism and Discrimination Exacerbate Climate Impacts on Human Health”, “Climate Change Affects the Economy Directly,” “Self-Determination Is Key to Indigenous Peoples’ Resilience to Climate Change.” These do not necessarily provide the quantitative or comparative data that decision-makers need to understand the magnitude of the impact of climate change on various sectors of society. Replacing persuasive headings with neutral, data-focused summaries would enhance the report’s credibility and utility.

Reconsidering Frivolous Sections

An objective scientific report should focus on delivering robust and actionable information, not on promoting activism. For example, the current report includes a “ClimateXArt” section on “the understanding that, together, art and science move people to greater understanding and action.” An objective scientific report intended to inform rather than persuade would not include such a section.

5. Conclusion

The NCA can reform its current image as a political tool by presenting the total picture—considering both threats and opportunities, incorporating exposure and vulnerability into disaster cost analyses, leveraging systematic review methods, and communicating findings in an objective, data-centric manner. These reforms would not only make the report more useful but also allow it to be more robust to pendulum swings between administrations.

Subscribe now

By Patrick Brown

“Los Angeles is burning, and accelerating hydroclimate whiplash is the key climate connection.”

That was the first line of the UCLA Press Release on a recently-published Nature review paper Swain et al. (2025): Hydroclimate Volatility on a Warming Earth.

Thanks in no small part to the huge journalistic audience that lead author Dr. Daniel Swain commands, the “climate whiplash” vernacular was immediately adopted in international headlines covering the recent Los Angeles fires:

• The New Yorker: Climate Whiplash and Fire Come to L.A.

• ABC News: How hydroclimate whiplash contributed to the severity of California wildfires

• Newsweek: Why a Rain 'Whiplash' Is to Blame for Los Angeles Fires

• The LA Times: Intensifying climate ‘whiplash’ set the stage for devastating California fires

• The Guardian: Climate ‘whiplash’ events increasing exponentially around world

• Financial Times: Fire and floods: the rise of climate whiplash

• Grist: The ‘weather whiplash’ fueling the Los Angeles fires is becoming more common

This coverage is not entirely organic as organizations like Covering Climate Now—which advise journalists on how to frame stories through a climate lens—highlighted climate whiplash as a good talking point for the Los Angeles fires.

Regardless, the paper has demonstrated incredible reach and is in the 99.99th percentile in terms of online attention for all research (not just climate research) of a similar age.

But as is the case for so much high-profile climate science, there is a large gap between the impression conveyed by the coverage and the impression left from the observational data.

So what is “climate whiplash” and how much is it changing?

Dangerous, intense wildfires require dry vegetation. The idea behind the climate whiplash connection to the Los Angeles fires is that very wet winters in Southern California in 2022-2023 and 2023-2024 enabled a great deal of vegetation growth but that the very dry beginning of the 2024-2025 winter allowed that vegetation to dry out, resulting in a landscape primed for uncontrollable wildfires. Swain explains the mechanism in interviews with Adam Conover and Neil deGrasse Tyson.

The idea being conveyed is that these climate whiplash events are dramatically increasing not just in Southern California, but globally.

“Every fraction of a degree of warming speeds the growing destructive power of the transitions,” Swain said.

Taking a step back, the fundamental theory undergirding changing “hydroclimate” (think water cycle where we are considering not just how precipitation provides moisture but also how evaporation takes moisture) whiplash is nothing new. It is a basic fact of atmospheric physics that a warmer atmosphere can “hold” more water vapor (about 7% more per °C of warming). This warming influence on the water cycle has been discussed in detail since at least the 1980s (e.g., Manabe, & Wetherald, 1986)). At first, most research discussed a general intensification of the water cycle, typically emphasizing that already dry areas would get drier and already wet areas would get wetter as the globe warmed. However, by the mid-2000s, studies like Trenberth et al. (2003), Chou & Neelin (2004), Meehl et al. (2005), and Held & Soden (2006) began pointing out that the same physics (warmer atmosphere holds more moisture) can drive larger variability in the same place—heavier rain events juxtaposed with prolonged and/or more intense dry spells.

These concepts are taught regularly as a part of Climate Change 101 classes, including my own, and they are accepted as consensus climate science, articulated with “high confidence” in the IPCC’s most recent assessment report:

“A warmer climate increases moisture transport into weather systems, which, on average, makes wet seasons and events wetter (high confidence)”

“Warming over land drives an increase in atmospheric evaporative demand and the severity of droughts (high confidence).”

However, I like to point out that it is useful to break down lines of evidence in climate science into categories of

1) Historical observations/trends

2) Fundamental theory

3) Mathematical modeling

I know from teaching the “wet gets wetter, dry gets drier” concept that the evidence for increased variability in the same location is much stronger in the theory and modeling categories than it is in observations. This is important because observations should take precedence over the other two. Focusing on observations tells us a lot about how big of an effect we’re talking about (i.e., do we see major trends emerge through the noise of the observation system and natural variability?). Furthermore, a fundamental point of doing science is to explain observations. The canonical order of operations is that first you observe some phenomenon, and then you use the tools of theory and modeling to make sense of it.

Observations of year-to-year water cycle variability

So, let's privilege “observations” of water cycle variability. I put the word observations in quotes here because it is more appropriate to think of the available datasets as highly processed products that incorporate observations rather than thinking of the data itself as pure observations. The starting point for the Swain et al. (2025) analyses (what they end up calculating their whiplash metric on) is something called the Standardized Precipitation Evaporation Index (SPEI), which is a highly processed unitless drought index that incorporates both precipitation and potential evaporation from the Earth’s surface to account for the net balance between the two.

One issue with interpreting SPEI as “observations” in this context is that the “dry gets dryer” mechanism is actually built into the calculation of the index from the outset (evaporation is not measured directly but calculated to increase with increasing temperature). Therefore, using the SPEI to “observe” that dry times get drier as they warm is technically committing the logical fallacy of begging the question where the premise of an argument assumes the truth of the conclusion.

Leaving that issue aside, let’s look at “observations” of annual variability because that is what is most relevant to the Los Angeles fires. Also, unlike in Swain et al. (2025), I’ll restrict the analysis to the land surface because that is where impacts are relevant (as exemplified by Fig. 1 in Swain et al. (2025)). It’s not particularly consequential if deluges only increase over the ocean, and the entire motivation to look at net water cycle variability using the SPEI is that the land surface can dry out—we are not worried about drying out the ocean surface.

Swain et al. (2025) use two input datasets in their calculation of SPEI, but I’ll focus on the ERA5 dataset since it is widely considered to be the gold standard for global analysis of most climate variables after the year 1950. First, let’s look at climate whiplash events exactly as they are defined in Swain et al. (2025).

Below is the occurrence of all whiplash events and just wet-to-dry whiplash events in the grid box closest to where the Eaton Fire in Los Angeles originated, directly from the Swain et al. (2025) published data.

That’s not exactly compelling evidence for a major increase. But these events are rare (and thus noisy), so to aggregate over more data, let’s look at the frequency of occurrence over a larger area. Below is the fraction of land in a larger region over southern California that experienced annual wet-to-dry whiplash events.

Still, not exactly compelling evidence for a major increase. So perhaps it was a stretch to try to shoehorn the results of the paper into the news cycle around the Los Angeles fires, but are the results still compelling globally? Below is the change in overall whiplash frequency of 12-month whiplash events between the first and most recent 30-year time periods over land using the ERA5 dataset.

So, over all global land, at the timescale that is most relevant to the Los Angeles fires (annual), in the premiere observational dataset (ERA5), using Swain et al. (2025)’s own data, we have seen a long-term decrease in whiplash frequency (this, by the way, is acknowledged in passing in the text of Swain et al. (2025) on page 37).

Let’s pause for a second to recall the first line of the UCLA press release (“Los Angeles is burning, and accelerating hydroclimate whiplash is the key climate connection.”) and the global news coverage it generated. Would any reader of this coverage have any idea about the incredibly important caveats above? Not that I can tell.

How large are changes in annual water cycle variability?

Now, to be fair, Swain et al. (2025) purport to show evidence of increasing whiplash frequency at multiple timescales, spatial extents (over the ocean, for example), and in other datasets.

However, highlighting changes in arbitrarily-defined “event” frequency without reporting changes in “event” magnitude is misleading, and it goes against one of the core recommendations of the National Academies of Sciences 2016 report on Attribution of Extreme Weather Events in the Context of Climate Change. As Ted Shepherd recently put it in his presentation to the committee responsible for the next such report: “Frequency is the more impressive number, but magnitude is perhaps the more physically interpretable number.”

So, let’s look at how large these changes in water cycle variability are (i.e., changes in magnitude) using the most conventional measure of the magnitude of variability, the standard deviation.

Below is a demonstration of what we might see if variability was dramatically increasing (i.e., what is conveyed by the coverage of the climate whiplash study). The black line is data with increasing variability (left y-axis), and the light blue bars are the standard deviation of that data for each decade (right y-axis). Whiplash events technically require swings to be very close in time but overall variability and whiplash frequency are clearly related.

So again, we see no long-term increase in water cycle variability at the location and timescale relevant to the Los Angeles fires.

Below is the same change in annual SPEI variability but over the entire land surface.

Rather incredibly, we have seen an overall decrease in annual scale water cycle variability (quantified using SPEI12) over global land since the 1980s.

How large are changes in annual precipitation variability?

Unfortunately, it is still difficult to interpret changes in the magnitude of SPEI variability because it is a unitless index. So, to get a better intuition of changes in water cycle magnitude variability using something with physical units, let’s look at the change in annual precipitation variability.

Changes in precipitation variability differ from the specific whiplash metric used in Swain et al. (2025), but using annual precipitation variability is certainly in line with how the research was presented and received. For example, in the context of the Los Angeles fires, prominent climate scientist Andrew Dessler cited the whiplash research when saying, “We are confident that climate change is making rainfall more variable.”

Uncertainty in global precipitation “observations” is quite large because of changes in the spatial footprint and/or density of the underlying rain gauge observations, the introduction of new satellite data or retrieval algorithms over time, uncertainty stemming from models used to calculate precipitation, or statistical uncertainty associated with the extrapolation to locations without observations. Because of these issues, many groups have created different observational datasets with different methods and assumptions. They generally attempt to create more-or-less homogenized data with continuous records over as large a spatial area and as high a spatial resolution as possible, but they all come with a laundry list of caveats. (precipitation is just one input into the SPEI calculation, so all these caveats apply to the SPEI calculations above as well).

Regardless, if the change in variability was large, you'd hope to see it manifest in climate science’s conventionally-used precipitation datasets. Below is the annual precipitation and its variability in eight such precipitation datasets for the Los Angeles region (the closest non-ocean grid point to the Eaton Fire in each dataset).

Forget “how large is the magnitude of change;” we see no coherent evidence of increased annual precipitation variability over Los Angeles in the first place.

But again, Swain et al. (2025) was about global changes. Below are global maps of change in annual precipitation variability (expressed as percent change in local standard deviation per decade) for the same datasets since the 1980s.

And finally, below is the global average of the local rainfall variability for each decade.

Again, discussing the change in magnitude is almost irrelevant when we see no coherent evidence of increased annual precipitation variability.

Summary

So, let’s recap. At the annual timescale that is most relevant to the Los Angeles fires…

• Figure 1: There is no clear increase in overall whiplash occurrence or wet-to-dry whiplash occurrence in Los Angeles in the premiere observational dataset (ERA5) using data directly from Swain et al. (2025).

• Figure 2: There is no clear increase in overall whiplash occurrence or wet-to-dry whiplash occurrence over southern California in the premiere observational dataset (ERA5) using data directly from Swain et al. (2025).

• Figure 3: There is a long-term decrease in whiplash events globally over land (where it matters) in the premiere observational dataset (ERA5) using data directly from Swain et al. (2025).

• Figure 5: There is no increase in the variability (standard deviation) of annual SPEI either for all months or centered on January) in the Los Angeles grid point using the pre-existing standard SPEI dataset.

• Figure 6: There has been a decrease in the variability (standard deviation) of SPEI over global land since the 1980s using the pre-existing standard SPEI dataset.

• Figure 7: There is no agreement on the direction of change (if any) in annual precipitation variability (standard deviation) over the Los Angeles area across eight different precipitation datasets.

• Figures 8 and 9: There is no agreement on the direction of change (if any) in annual precipitation variability (standard deviation) globally across eight different precipitation datasets.

While “climate whiplash events” may be increasing in frequency under most of the very specific, selected definitions used and datasets investigated in Swain et al. (2025), the general idea that annual precipitation (or more generally, the water cycle, which includes evaporation) is becoming dramatically more variable is not supported when a broader set of datasets and definitions are used.

Would a reader of Swain et al. (2025), or especially its coverage, have any idea about the weakness of its broader conclusions or the lack of robustness of its results to different definitions and datasets? Almost certainly not, and I contend that this is a major problem for public understanding and trust in climate science.

Why don't we see a robust increase in water cycle variability given the strong theory underpinning “wet gets wetter, dry gets drier”? For one thing, the theoretical size of the effect is known to be quite small relative to natural, unforced variability, making it inherently difficult to detect. For example, we see in Figure 7 above that year-to-year rainfall in Los Angeles naturally varies by as much as 300%, yet the signal we are looking for is one to two orders of magnitude less than this. It is also apparently the case that observational uncertainty is larger than the signal (or there would not be such disagreement between datasets). Physically, perhaps increasing mean precipitation is offsetting the increase in calculated evaporation in the SPEI index, reducing its variability. Maybe reduced temperature variability (via arctic amplification) is reducing calculated evaporation variability.

I don’t know the full answer, but these would be great research questions to identify and outline in a Nature review like Swain et al. (2025). Unfortunately, Swain et al. (2025) missed this opportunity because the paper seemed so focused on assembling evidence in favor of increasing water cycle variability that contradictory evidence was never presented or seriously grappled with.

My main discomfort with Swain et al. (2025) and its rollout is that it appears that the primary goal was to create and disseminate the “climate whiplash” meme rather than conduct a truly rigorous evaluation of the evidence, including countervailing evidence. Ultimately, this makes the research a much larger advance in marketing than an advance in science.

Subscribe now

A version of this article was originally published by City Journal, you can read that here.

Devastating images are emerging out of the Los Angeles area depicting the immense destruction of several fires that are raging through communities this week. These fires are likely to be the costliest wildfires in US history, overtaking the 2018 Camp Fire that destroyed Paradise, California.

Many are looking for a clear culprit to blame for this disaster, with climate change being one of the most popular. This disaster, however, is the result of a confluence of natural and human-caused factors. Climate change may be playing a role, but a narrow focus on it could be counterproductive as it distracts from more direct measures to reduce fire risk and impact.

Is climate change the culprit?

Fire danger is a product of meteorological and fuel conditions. You also need an ignition—you can't start a fire without a spark. Global warming from fossil fuel burning raises temperatures relatively uniformly and thus affects all aspects of meteorology relevant to fires to some degree, though often not in the same direction.

These fires are being driven by a particularly intense Santa Ana wind storm with widespread hurricane-force winds and gusts over 80 mph in some places. These winds are the primary meteorological driver of the disaster: they provide the oxygen needed for burning, drive flames forward, spread embers, and can make aerial firefighting impossible.

Santa Ana winds are an innate part of southern California's climate, and there is little evidence that climate change will make them worse. If anything, we expect Santa Ana winds to become less intense/frequent as the climate changes.

Fires need something to burn, so the state of vegetation (the fuels for fires), and in particular how dry that vegetation is, is the other key ingredient for fires. The Los Angeles area has received much less rain than normal for this time of year, which makes vegetation more flammable. However, there is little evidence that warming is a primary driver of a lack of precipitation like what we have seen this year.

However, when Los Angeles does see droughts like this, a warmer climate means a drier atmosphere, which works to dry vegetation and makes vegetation more flammable.

Is vegetation the problem?

Then, there are the characteristics of the vegetation on the ground, such as how much fuel there is for fires. The situation in Southern California Chaparral brush is different than in the forests to the north (where a major problem is fire exclusion and the century-long build-up of overstocked forests). Nevertheless, mechanical brush removal and prescribed burning can still be used to reduce vegetation and fire danger, though it can come at a cost to ecosystems.

So, how effective would proactive vegetation reduction be in the face of continued warming?Our research seeks to answer this question. The procedure takes conditions over the course of a fire, warms the background climate, and simulates vegetation reduction to see how future warming and fuel treatments would simultaneously affect a fire. Below is what this looks like when applied to the weather in the Los Angeles area this week.

Our modeling indicates that warming is indeed contributing to increased wildfire intensity in the region, with a further enhancement by mid-century.

It is important to note that two different greenhouse gas emissions scenarios are shown. One is a slow emissions reduction scenario that roughly represents current global policies (SSP2-4.5), and one represents fast emissions reductions that are much more aligned with the Paris Agreement (SSP1-2.6). The movement from the slow emissions reduction scenario to the Paris Agreement scenario represents major differences in the global energy and agricultural economies, technology adoption, geopolitics, etc. Yet, these differences translate into only marginal changes in 2050 fire intensity (from +7.2% to +5.5%).

When it comes to addressing wildfire risk, reductions in greenhouse gas emissions are often heralded as the primary lever we can pull. However, trying to influence weather-related outcomes through energy policy is incredibly indirect and there are often much more direct solutions at hand.

In that vein, the blue bar represents fire intensity in 2050 under slow emissions reductions (4.5) but with proactive vegetation reduction. We see that even in a situation of 2050 warmth, under a slow emissions reduction scenario, vegetation reduction would reduce fire intensity by about 15% relative to today.

This is evidence that California’s goal to increase hazardous fuels reduction treatments to 1 million acres per year, along with the federal government’s target of reducing fuels on 50 million acres over a decade, are laudable and would be helpful in combating the growing wildfire threat.

However, actually achieving these goals requires overcoming multiple challenges, including funding constraints, workforce shortages, and logistical issues related to complex land ownership patterns.

Finally, there are bureaucratic and regulatory obstacles associated with the California Environmental Quality Act (CEQA) on California state land and the National Environmental Policy Act (NEPA) on federal land. The NEPA process, in particular, is associated with significant obstacles to fuel reduction treatments. The Forest Service writes the most NEPA reviews of any federal agency and is the most likely to be sued, resulting in litigation that delays project implication by about 3 years on average.

Human Ignitions

All these fires were lit by people, and thus, the huge population of southern California provides many ignition sources that would not be there naturally. Common causes are equipment use (sparks from chainsaws, mowers, etc.), sparks from vehicles, ATVs, dirt bikes, smoking, campfires (or fires in homeless encampments), BBQs, fireworks, and Arson. Thus, Increasing public awareness of fire safety and red flag warnings should help reduce these ignitions, and this is actually the case as human ignitions have probably decreased over the past several decades.

There is also the issue of utility-caused fires. Southern California Edison has preemptively shut off power in several regions to reduce the risk of powerline-caused fires. This is not an ideal solution, but it does prevent ignitions. In the longer term, we can continue to reduce vegetation around power lines, bury distribution lines, and install powerlines that automatically de-energize when they make contact with an object.

Home hardening & firefighting.

Given that extreme fire weather conditions like this will inevitably occur and given that human ignitions cannot be eliminated, protection at the structural level is also important. Houses are much more resilient to fires if they have no vegetation within 5 feet of the house, and vegetation is fire-resistant and sparse from 5 feet to 100 feet. Building codes and “home hardening” also make a difference. Things like non-combustible roofing materials (e.g., metal, tile, or asphalt shingles), ember-resistant vents with mesh screens, and fire-resistant materials for siding (like stucco, fiber cement, or metal) have been shown to be effective in the lab as well as real-world settings. There is also a synergistic effect, with increased overall effectiveness, as more houses in a neighborhood adopt these practices.

Also, it goes almost without saying that well-resourced firefighting (personal as well as air and ground equipment), as well as high-quality fire-weather forecasting, are critical to slowing down and ultimately containing fires like these.

Overall

Devastating events like this leave many to search for a simple villain to blame, but the reality is often much more complicated. Climate change may be contributing to some increase in fire danger in this case, but its lack of influence on the high winds and drought means that it does not deserve central billing. Furthermore, the effects of global emissions reductions on fire activity are quite indirect and delayed.

Fire suppression and the long-term build-up of fuels are not an issue in brush landscapes the same way that they are in the Northern California Forests, but our modeling indicates that vegetation reduction would still be able to appreciably reduce fire danger in these landscapes.

Apart from those measures, the main way these types of events can be mitigated in the future is via further reductions in human ignitions, potentially increased firefighting resources/technology, and enhanced “home hardening” measures within fire-prone communities.

Ultimately, we live on a planet that is often hostile to our well-being no matter what we do. Southern California is a region that has been very fire-prone throughout human history, and it will continue to be so indefinitely. Thus, devastation from natural disasters cannot be completely avoided, and we are often left with partial measures that are only able to reduce rather than eliminate risk.

Weather and climate extremes—such as high temperatures, floods, droughts, tropical cyclones, extratropical cyclones, and severe thunderstorms—have always threatened both human and natural systems. Given their significant impacts, there is considerable interest in how human-caused climate change influences these extremes. This is the focus of the relatively new discipline of Extreme Event Attribution (EEA).

Over the past couple of decades, there has been an explosion in EEA studies focusing on (or, “triggered by”) some prior notable weather or climate extreme. Non-peer-reviewed reports from World Weather Attribution (e.g., here, here, and here) represent some of the most notable examples of these kinds of analyses, and many similar studies also populate the peer-reviewed literature. The Bulletin of the American Meteorological Society’s “Explaining Extreme Events From a Climate Perspective” annual series compiles such studies, as does the Sabin Center for Climate Change Law, and they are also synthesized in reports like those from the IPCC (IPCC WG1 AR6 Chapter 11.2.3) and the United States National Climate Assessment.

The collective output of these kinds of studies certainly gives the impression that human-caused climate change is drastically changing the frequency and intensity of all kinds of weather extremes. Indeed, Carbon Brief recently published an extensive summary of the science of EEA studies, which begins with the proclamation, “As global temperatures rise, extreme weather events are becoming more intense and more frequent all around the world.”

This is seemingly supported by hard data from Carbon Brief’s extensive catalog of nearly every published EEA study. Their database includes over 600 EEA studies on about 750 extreme weather events and trends. They report that approximately 75% of these events were intensified or made more probable due to climate change, while only 9% were made less intense or less likely as a result of climate change. This skew towards increased intensity or likelihood is evident across various event types, appearing in ten of their twelve categories.

However, these numbers cannot be taken as an accurate quantification of the influence of climate change on extreme weather because they are heavily influenced by a cascade of selection biases originating from the physical climate system, as well as researcher and media incentives. Identifying and understanding these biases is a prerequisite for properly interpreting the collective output of EEA studies and, thus, what implications they hold for general scientific understanding, as well as political and legal questions.

Share

An Apparent Discrepancy?

One might conclude from the collective output of EEA studies that there is strong evidence indicating an increase in the frequency and intensity of extreme weather. Yet, this conclusion seems to be in tension with the more comprehensive evaluations of extreme weather changes found in the Intergovernmental Panel of Climate Change’s (IPCC) Working Group 1, Chapter 11, and Chapter 12, which are noticeably more reserved in their conclusions on identifying and attributing shifts in these extremes (see also Roger Pielke Jr.’s Weather Attribution Alchemy series).

The IPCC indicates that extreme heat over land is rising at a rate that is roughly equal to, or just below, the mean warming rate for land. This increase, however, is balanced by a decrease in extreme cold. Consequently, there is no substantial global net rise in the occurrence or intensity of extreme temperatures. Furthermore, the IPCC observes that there are currently not detectable globally coherent trends in inland flooding. Drought conditions are variable, with some types of droughts found to be increasing in specific areas, yet there is not evidence of any global trend in meteorological droughts characterized by precipitation deficits. Trends in tropical and extratropical cyclones, as well as severe thunderstorms, all show mixed results with no clear long-term increase.

Overall, the lack of a strong influence of climate change relative to natural variability is exemplified by the IPCC’s Table 12.12, which uses a standardized methodology for the detection of change to assess when (if ever) the climate change signal will emerge from the noise of natural variability for a number of “Climatic Impact Drivers.” The color corresponds to the highest level of confidence reported in any region, and white cells indicate that evidence is lacking or that any climate change influence is not yet detectable above natural variability.

It is notable, especially in the context of the aforementioned Carbon Brief catalog of EEA studies, that there are white cells (climate change not detected) over the historical period for

• river flood

• heavy precipitation and pluvial flood

• aridity

• hydrological drought

• agricultural and ecological drought

• fireweather

• mean wind speed

• severe wind storm

• tropical cyclone

• hail

So what is going on?

Cause of the Apparent Discrepancy

The large apparent discrepancy between the size of the influence of human-caused climate change on extreme weather reported in EEA studies (like those compiled by Carbon Brief) compared to more comprehensive systematic analyses (like those compiled by the IPCC) can, in large part, be attributed to the many layers of Selection Biases that influence the EEA literature's collective output.

Selection Bias is a broad term that refers to any bias that arises from a process that selects data for analysis in a way that fails to ensure that data is representative of the broader population that the study wishes to describe.

The type of selection bias that is perhaps most familiar to most people (and a slightly more serious version of the above cartoon) is the Nonresponse Bias of polling, where individuals who do not respond to a poll differ systematically from those who do respond. For example, if Republican voters are less likely than Democratic voters to respond to an election poll, then the poll will have a sampling bias in the sense that the sample (people who responded to the poll) is biased towards Democratic responses and is thus not fully representative of the population that will end up voting in the election.

Selection biases in the context of EEA studies include those associated with the physical climate system itself, those concerning proclivities and incentives facing researchers/journals, and those concerning the proclivities and incentives facing the media. They include Occurance Bias, Choice Biases, Publication Bias, and Media Coverage Bias.

Occurrence Bias is a bias introduced by the physical climate system. Since EEA studies tend to be triggered by extreme events that have actually occurred, there is reason to believe that these studies will disproportionately sample events that are more likely than average to be exacerbated by climate change because the events occurred in the first place. Essentially, extreme events that are more likely to occur under climate change—and thus more likely to be observed—are going to be overrepresented in EEA studies, and extreme events that are less likely to occur under climate change—and thus less likely to be observed—are going to be underrepresented in EEA studies.

The map below illustrates this phenomenon. It shows changes in the magnitude of extreme drought under climate change. Specifically, it shows the fractional change in the intensity of once-per-50-year droughts (as quantified by monthly soil moisture) between a preindustrial and 21st-century run (SSP2-4.5 emissions) of the highly-regarded NCAR CESM2 Climate Model. Blue areas represent locations where the model simulates that extreme droughts become less frequent and intense with enhanced greenhouse gas concentrations, and red areas represent locations where the model simulates that extreme droughts become more frequent and intense with enhanced greenhouse gas concentrations. It is notable that overall, this model simulates that warming decreases the frequency and intensity of extreme drought in more locations than it increases it (consistent with soil moistening under warming simulated by other models).

Now, here’s the kicker: The dots show locations where once-per-50-year droughts actually occurred in the 21st-century simulation and thus represent events that would plausibly trigger EEA studies.

What do you notice about where the dots are compared to where the red is? That’s right; the simulated EEA studies overwhelmingly sample areas where droughts are getting more intense and more frequent by the very nature that those are the types of droughts that are more likely to occur in the warming climate. The result is that the EEA sample is majorly biased: warming decreased the intensity of once-per-50-year droughts by about 1% overall, but it increased their intensity within the EEA sample by 18%! Thus, if you just relied on the EEA sample, you would come away with an incorrect impression not only on the magnitude of change in extreme droughts but also on the sign of the direction of change!

Choice Bias arises when researchers use prior knowledge to choose events for EEA studies that are more likely to have been made more severe by climate change. A clear example of Choice Bias pervading the Carbon Brief database is there have been 3.6 times more studies on extreme heat than there have been on extreme winter weather (205 vs. 57). Another example would be the dearth of EEA studies on extratropical cyclones (the kinds of low-pressure systems with cold and warm fronts that are responsible for most of the dramatic weather outside of the tropics). The IPCC states that the number of extratropical cyclones associated with intense surface wind speeds is expected to decrease strongly in the Northern Hemisphere with warming. Yet, it is relatively rare for EEA attribution studies to be done on these types of systems, which results in an exclusion of this good news from the EEA literature.

Publication Bias could be playing a role, too, where researchers are more likely to submit, and journals are likely to publish studies that report significant effects on salient events compared to studies that find null effects.

Finally, the climate reporting media ecosystem is characterized by actors whose explicit mission is to raise awareness of the negative impacts of climate change, and thus, there will be a natural Media Coverage Bias with a tendency to selectively highlight EEA studies where climate change is found to be a larger driver than EEA studies that do not reach such a conclusion.

These selection biases are apparent at the aggregate level, but there is also strong evidence of their presence in individual studies. My self-critique of my 2023 Nature paper (which was an EEA study on wildfires) invoked my own Choice Biases, Publication Bias, and Media Coverage Bias.

A more recent specific example suggestive of many of these dynamics is a study, Gilford et al. (2024), titled “Human-caused ocean warming has intensified recent hurricanes”. This study was conducted by three researchers at Climate Central, which summarizes the study’s findings with the following infographic:

Essentially, they claim that climate change is enhancing the intensity of all hurricanes and that the enhancement is quite large: Storms today are calculated to be an entire Category stronger than they would have been in a preindustrial climate.

This is a huge effect, and thus, if it were real, it is reasonable to expect to see clear long-term trends in metrics of tropical cyclone (hurricane) intensity like the accumulated number of major (Category 3+) hurricane days or the accumulated cyclone energy from all tropical cyclones (which is proportional to the square of hurricane windspeed accumulated over their lifetimes). However, any long-term trends in such metrics are subtle at best, both globally and over the North Atlantic.

So, this is a microcosm of the aforementioned apparent discrepancy between more broad quantifications of changes in extremes and their associated EEA counterparts, and again, I’d argue there are several selection biases at play affecting the production and dissemination of the EEA study.

Let’s start with Choice Bias on methodology. Human-caused warming changes the environment in some ways that work to enhance hurricanes and in other ways that diminish them. The main way that hurricanes are enhanced is via the increase in sea surface temperatures (which provides the fundamental fuel for hurricanes), and the main way that hurricanes are diminished is via changes in atmospheric wind shear and humidity. The net result of these countervailing factors pulling in opposite directions is that we expect fewer hurricanes overall, but when hurricanes are able to form, they can be stronger than they would otherwise. These factors, though, are small relative to natural random variability, and thus, they are difficult to detect in observations.

However, the Climate Central researchers made the methodological choice to largely exclude the influence of factors that diminish hurricane development from the study. First, it uses a “storyline approach” that takes the occurrence of hurricanes as a given. In other words, although there is good evidence that warming should reduce hurricane numbers, the methodology implicitly holds hurricane numbers fixed. Second, the main reported results are only for the effect of warmed sea surface temperatures on hurricane strength, neglecting the countervailing effects of changes in the atmosphere that would reduce hurricane strength (The study includes a half-hearted and methodologically convoluted attempt to account for the countervailing influences of tropical cyclone strength in one sensitivity that cuts the influence of climate change on wind speeds by more than a third. However this result is not reported in the abstract, main figure, or the study’s infographic).

Climate Central’s headline on their study was “Climate change increased wind speeds for every 2024 Atlantic hurricane,” but this headline would be more accurate if it was immediately followed by “...if we focus exclusively on those aspects of climate which we know intensify hurricanes (sea surface temperatures) and ignore or downplay those aspects of climate change which would work to diminish hurricanes.”

Are these Choice Biases in event type and methodology an accident? There are many reasons to believe they are not.

The research paper itself spells out that the motivation of the study is to “connect the dots” between climate change and hurricanes because “landfalling hurricanes with high intensities—can act as ‘focusing events’ that draw public attention” and that “Increased attention during and in wake of storms creates opportunities for public and private discourse around climate and disaster preparedness.”

It is also relevant that the study was funded by the Bezos Earth Fund, The Schmidt Family Foundation, and the CO₂ Foundation, all of which have missions in line with the eventual conclusions of the study (but no potential conflicts of interest were reported by the study's authors). For example, according to the CO₂ Foundation’s website, its mission is to “execute impactful grant-making and communication about the urgent societal risks from extreme weather…”

Then, there is the extensive media coverage of this study. It was picked up by 134 news outlets and ranked in the 99.95th percentile of research articles (across all journals) of similar age in terms of online attention. Further, it was immediately incorporated into seven Wikipedia articles (likely having high leverage on AI queries, which would make its findings indistinguishable from scientific “fact”). This is affected by the aforementioned Media Coverage Bias, but it is also undoubtedly directly influenced by the efforts of Climate Central, which is explicitly an advocacy organization whose self-described specialty is media placement and dissemination. In a recent fundraising email, they led with this theme: “The influence our analysis had on media coverage of Hurricanes Helene and Milton before they even made landfall gives just a taste of what we can achieve with your support.”

Motivations behind the skew of EEA studies

The above sheds light on the reasons for certain choice biases in a particular study, but there is plenty of evidence that these selection biases are pervasive in the EEA field. After all, Dr. Myles Allen essentially founded the field with the motivation of answering the question, “Will it ever be possible to sue anyone for damaging the climate?”. This same motivation seems to animate many of the most high-profile scientists in the field today, like Allen’s protege, Dr. Friederike Otto (co-founder and leader of World Weather Attribution). She and her organization are frequently cited as bringing the necessary intellectual authority to credibly sue fossil fuel companies. She states the motivation of her work explicitly: “Attributing extreme weather events to climate change, as I do through my work as a climatologist, means we can hold countries and companies to account for their inaction.”

Similarly, Dr. Emily Theokritoff—a research associate at Grantham, who is working on an initiative to publish rapid impact attribution studies about extreme weather events, similar to World Weather Attribution—told Carbon Brief that “The aim is to recharge the field, start a conversation about climate losses and damages, and help people understand how climate change is making life more dangerous and more expensive.”

Given the explicitly stated motivation of those in the EEA field, it is quite reasonable to suppose that there are major selection biases at play, and thus, it is not at all surprising that the collective output of the EEA field would look so different from more broad comprehensive assessments.

Implications of Selection Biases in EEA Studies

Returning to the apparent discrepancy laid out above, the IPCC Chapters 11 and Chapter 12 are addressing the question:

“What is the impact of human-caused climate change² on extreme weather¹?”

Whereas the collective output of the EEA field appears to be asking the same question, but it implicitly contains two important footnotes:

¹ By extreme weather, we mean mostly the kinds of extreme weather that we already know are made worse by human-caused climate change and not the kinds of extreme weather that we know are ameliorated by human-caused climate change.

² By human-caused climate change, we mean mostly the sub-elements of human-caused climate change that we already know make the kind of extreme weather we are studying worse.

Appreciating these footnotes is critical if we want to communicate our best scientific understanding of the issue to the public and to reliably inform political and legal questions.

The general belief that climate change is greatly impacting extreme weather appears to have grown alongside the surge in EEA studies, reflected in a marked increase in Google searches for “climate change extreme weather."

These EEA studies also carry practical legal and policy implications as they frequently relate to assessed monetary damages from extreme events which means they can be used to bolster support for more stringent greenhouse gas emissions reductions via their inclusion in the calculation of the social cost of carbon. EEA studies are also frequently discussed in the context of the Loss and Damage Mechanism of the United Nations Framework Convention on Climate Change which would mean they could be used to justify climate reparations payments across borders.

Finally, as alluded to above, EEA studies can potentially be used in direct legal actions against entities such as fossil fuel companies. In 2017, two lawyers wrote a Carbon Brief guest post stating, “we expect that attribution science will provide crucial evidence that will help courts determine liability for climate change-related harm.” Four years later, the authors of a study on climate litigation wrote a Carbon Brief guest post explaining how attribution science can be “translated into legal causality.” They wrote: “Attribution can bridge the gap identified by judges between a general understanding that human-induced climate change has many negative impacts and providing concrete evidence of the role of climate change at a specific location for a specific extreme event that already has led or will lead to damages.”

Given the broad scientific and practical implications of EEA studies, it is crucial that the selection biases highlighted here are appreciated and ultimately fought against so that the output of the EEA field can more accurately reflect the overall influence of climate change on extreme weather.

By Patrick Brown

The election of Donald Trump for a second stint as US president has stimulated a great deal of reflection from those on the left side of the political aisle. A recurring theme has been that the policies and cultural issues emphasized by the left wing of the democratic party are unpopular with the average voter and thus are weighing down the Democratic Party as a whole. An increasingly embraced prescription has thus been to reject the overreach of the far left wing of the party. 

In a parallel discussion, scientists and their institutions (universities, government research labs, professional societies, journals) are also reflecting on the election results, especially what they signal in terms of public trust in science and expertise. Along with declining trust in other societal institutions, trust in science has deteriorated, with about 1/4th of US adults and 1/3rd of self-identified Republicans now reporting that they have little or no confidence that scientists act in the best interest of the public. Most science is conducted within the university setting, and trust in universities has cratered among Republicans, with only 31% saying that universities have a positive impact on the nation, compared with 74% of Democrats. 

A primary reason why Republicans distrust science is that they perceive that universities and institutional science are fused with the Democratic party. Surveys of party affiliation among academics more-or-less bear this out, as Democrat faculty outnumber Republicans by much more than 10 to 1 in most departments. Accordingly, there is a substantial concern, even emanating from within the academy, that the research output of universities is heavily skewed toward findings that support left-leaning worldviews. Along these lines, Harvard professor Roland Fryer recently expressed grave concerns that research that could be used to support right-leaning worldviews is likely to be considered too controversial for publication, saying, “Realistically, either journal editors are refusing to publish controversial results, or academics are too cowardly even to do the research.” Survey data indicates that 34 percent of professors feel pressured by peers to avoid “controversial research,” with 91 percent reporting being at least “somewhat likely” to self-censor in academic publications, meetings, presentations, or on social media. 

It is in this context of declining public trust in academia and science, especially among Republicans, in which academics and researchers are reflecting on how best to conduct themselves during a second Trump presidency. There are a myriad of dynamics to consider, but lessons can certainly be taken from what transpired in the aftermath of the first Trump election in 2016, which has directly impacted the current levels of mistrust in universities and science. 

President-elect Trump’s persona, characterized by bluster and a propensity to lie and “bullshit,” make him personally offensive to many of us who value curiosity, rationality, and careful methodical thinking. Thus, the natural impulse of most academic and scientific institutions has been to define themselves in opposition to Trump (and to conservative parties more generally). The result, however, has been that during the first Trump presidency, the line between political advocacy and academic science, which had already been blurry, became virtually invisible. Universities and groups of experts embraced the general tone of manichaeism, making no secret which side of the political aisle was to be seen as good and which bad. Prominent examples of this were the 2017 "The March for Science” and the world’s most prestigious scientific journal, Nature, officially endorsing Joe Biden for president in 2020 as well as Kamala Harris for President in 2024. A particularly salient example of political bias was the letter from health experts in the middle of the COVID-19 pandemic, which abruptly reversed previous messaging and endorsed an exception for social gatherings, so long as it was for the progressive cause of the moment.

It is clear to me that Trump does not embrace the core values of science, so I understand the scientific community's impulse to stand in opposition to him. We are already seeing this instinct take hold again, as the first reaction from Nature was an article headlined, “Scientists around the world expressed disappointment and alarm as Republican Donald Trump has secured a second term.”

However, if we, as a scientific community, want to regain public trust, we must follow a similar strategy to the recent self-reflection of sects of the Democratic party. That is, we should reject the urge to overreach. Many Democrats are now critiquing the extreme activist wing of the party for overreaching and alienating the general public on social issues. The scientific community can draw a parallel critique towards scientists that overstate scientific consensus and assert an all-encompassing authority on policy decisions.

The Overreaching Tendencies of Climate Science

“…the lessons that ought to have followed the [first Trump] election—lessons about the importance of understanding other Americans, the necessity of resisting tribalism, and the centrality of the free exchange of ideas to a democratic society—have not been learned. Instead, a new consensus has emerged in the press, but perhaps especially at this paper: that truth isn’t a process of collective discovery, but an orthodoxy already known to an enlightened few whose job is to inform everyone else.”

The above statement by Bari Weiss refers to the attitudes in elite left-of-center media, but it could just as easily have been said about attitudes in elite scientific and academic institutions. The idea that climate scientists engage in a special way of thinking and have superior knowledge of how global society should be arranged—what Mike Hulme identifies as gnosticism—is smug and condescending. Not coincidentally, it has fueled public resentment, as 61% of Republicans and 32% of Democrats now say research scientists feel superior to others. Not only has this superiority complex hurt science’s credibility, it is seriously flawed. This attitude has contributed to the scientization of the climate change problem, in which a complex, value-laden, and contentious political issue is misleadingly presented as purely scientific or technical. This dynamic fosters the depoliticization of the issue, as people mistakenly imagine that science is capable of acting as a supreme authority, capable of unequivocally dictating policies without the need for the discursive processes inherent to traditional political debates. 

This scientization and depoliticization is clear in U.S. Climate Envoy John Kerry’s justification of  the political positions of the Biden Administration:

“Nothing that we are doing, nothing that President Biden has sought to do, has any political motivation or ideological rationale. It’s entirely a reaction to science, to the mathematics and physics that explain what is happening [to the climate],”

Kerry is not a scientist, but this sentiment was also ascendant at the highest levels of leadership within scientific institutions during the first Trump presidency. For example, editor-in-chief of Science, Holden Thorpe, more-or-less, affirmed Kerry’s stance on science and politics when defending Nature’s decision to officially endorse Joe Biden in the 2020 election. Thorpe stated, without political endorsements from scientific institutions, people would have “the permission to say things like ‘climate change may be real, but I don't think we should have government regulation to deal with it,’ which is unacceptable.” (my emphasis).

Similarly, the editor-in-chief of Nature, which regularly publishes articles that go quite far in their support of particular policies, stated unequivocally that “When it comes to science, Nature does not have a preferred narrative.” This all aligns with the notion that science can not only inform but dictate policy and was exemplified by official condemnations of Trump’s first exit from the UN Paris Agreement from the highest levels of scientific authority, such as from the American Association of the Advancement of Science. But is it legitimate to state that exiting the Paris Agreement is actually anti-scientific? 

What Can Science Say on the Climate Change Problem?

In this context, it is relevant to review what claims can truly be considered “anti-scientific” in climate policy debates. Under the strictest definitions, the scientific method involves making observations, forming hypotheses about the underlying causes of those observations, and then meticulously designing experiments to test the validity of the hypotheses. The strongest hypotheses, and those that get conferred the status of being more-or-less true, are those that have withstood all attempts to falsify them to date. Unfortunately, the types of questions where it is feasible to design neat tests of clear hypotheses are rather narrow and tend to be limited to fundamental aspects of physics, chemistry, and biology. Outside of these areas, published scientific studies often resemble detective-like inductive inference, assembling the most parsimonious story from the available evidence. 

Claims in climate policy are on a spectrum from being quite amenable to the scientific method to being almost completely outside its realm. One way to help delineate different parts of the spectrum is to decompose the predominant narrative on the climate problem into four parts:

1. Climate change is happening, 

2. Contemporary climate change is caused almost entirely by humans, 

3. The negative impacts of climate change are larger than the positive impacts

4. The positive impacts of reducing emissions rapidly enough to adhere to the Paris Agreement temperature limits are larger than the negative impacts.

Although climate science is heavily constrained by the impossibility of using copies of Earth to test any hypotheses about the causes and consequences of global climate change, statements 1 and 2 are essentially narrow enough to be addressed by components of the scientific method. We can test, for example, that gases like carbon dioxide absorb and reemit radiation at wavelengths that cause them to have a greenhouse effect. We can observe the increase in the greenhouse effect directly using satellites and surface instruments, and we can document the enhanced greenhouse effect’s predicted consequences on temperatures through independent observations and instrumental networks. Finally, we have physics-based mathematical models that allow us to conduct simulated hypothesis tests of what the world would look like with and without increased greenhouse gas concentrations. The oft-quoted “97% to 99% scientific consensus on global warming” applies to statements 1 and 2. Thus, when Trump has done things like call the entire concept of climate change a hoax, this can legitimately be called out as science denial.

Beyond statements 1 and 2, however, things become more fuzzy. Statement 3—on the net negative impacts of climate change—veers from a descriptive claim of reality to a normative claim of desirability, invoking values outside of the traditional realm of science. Most would agree that the first-order negative impacts from carbon emissions (e.g., increases in extreme heat, sea level rise, increases in floods and droughts in some regions, enhanced fire weather conditions, and possibly stronger, though less frequent hurricanes) outweigh the first-order positive impacts (from, for example, reduced deaths from cold spells, carbon dioxide fertilization of plants, or the expansion of habitable climates and agricultural zones in higher latitudes). However, what is more relevant to discussions of policy like the Paris Agreement (Statement 4) is the net effect of industrialization via fossil fuels. 

Consider that historical warming of ~1.4°C since the Industrial Revolution has coincided with almost all climate-sensitive aspects of global human society trending in positive directions in recent decades, as technology and economic development have outpaced negative climate impacts. Consider also that despite these positive trends globally, tremendous inequality persists today, with 3 billion people still in extreme energy poverty, and the observed average mortality from floods, droughts, and storms 15 times higher for low-energy use countries than for high-energy use countries. Even in high-energy use countries like the U.S., many families at the lower end of the income distribution are forced to endure harmful indoor temperatures or even sacrifice medicine or food in order to pay for energy costs. 

Thus, any policy, like those in the Paris Agreement, that set hard limits on emissions restricts energy options, creating strong potential to raise costs and harm human resilience to climate change as well as quality of life overall. 

When we take this more holistic view of the problem and acknowledge the messy trade-offs involved, we see that this is a political controversy with technical underpinnings that can't actually be resolved by technical means, because the ultimate target decision (not to mention many of the technical aspects themselves) is subject to different framings and interpretations based on value-based political and ethical positions. The issue involves technological, economic, ethical, philosophical, and moral questions far outside neat scientific hypothesis testing. Therefore, science is not and cannot be a substitute for traditional democratic argumentation and deliberation. 

Thus, it will not be strictly “anti-science,” when the Trump administration withdraws from the Paris Agreement for a second time, or when it implements policies more friendly to domestic fossil fuel production. 

Delineate Between Scientific Claims and Political Preferences.

As mentioned above, researchers and academics skew overwhelmingly towards the left side of the political aisle, another example of the diploma divide in American politics. For those on the left, this is often interpreted rather self-servingly that “reality has a well-known liberal bias.” expounded on at book length in Chris Mooney’s The Republican Brain: The Science of Why They Deny Science—and Reality. But if dogma vs. reason were the best way to describe the two factions in American politics, it wouldn't be the left that tends to oppose mainstream scientific views on issues such as GMOs, nuclear power, alternative medicine, blank-slate accounts of human nature, or biological sex.

A better explanation of the political alignments in U.S. politics is that people tend to embrace conclusions—scientific or otherwise—that support themes, ideologies, and narratives that are preexisting components of their worldview. To that end, the left/right divide in the U.S. seems to cut along themes quite similar to how left and right were initially defined during the French Revolution; people differ in how they perceive the desirability and validity of social hierarchies. It just so happens that mainstream climate policy, represented most prominently by the Paris Agreement, aligns very well with several aspects of the more egalitarian preferences and the globalism preferred by the left, and it is in conflict with the social hierarchical preferences and nationalism preferred by the right. This is why populist/nationalist leaders who are more articulate than Trump, like Argentina’s Javier Milei, oppose UN plans, not on technical grounds but on claims that the plans are “super-national socialist government program that aims to solve the problems of modernity with solutions that undermine the sovereignty of nation-states and violate the right to life and property of individuals.” 

There are various hypotheses as to why scientists and academics hold left-leaning views, but I am skeptical of the self-serving notion that more credentialed people hold these views because the views are somehow fundamentally more true. While it may be the case that most academic scientists engage in research that entails quantitative rigor constrained by empirical data on narrow research questions, scientists often conflate this day-job activity with the notion that their political worldview is also arrived at through pure empiricism and rationality. On the contrary, I would argue that scientists arrive at their political opinions in mostly the same way everybody else does: through a combination of their natural temperament, the culture they are embedded in, personal experience, and the resulting incomplete and oversimplified models of the world. More self-reflection on this reality by prominent scientists should act as a counterweight against the desire to present themselves as neutral or purely scientific while subtly promoting a specific political agenda (what Roger Pielke Jr. has labeled Stealth Issue Advocacy). 

Regain Public Trust by Rejecting Overreach.

“Entangling the sciences in political disputes in which differing views of nature, society, and government are implicated has not resolved or narrowed those disputes, but has cast doubt upon the trustworthiness and reliability of the sciences and experts who presume to advise on these matters.”

-Steve Rayner and Daniel Sarewitz

The public understands correctly that many political questions are moral/ethical, involve tradeoffs, and are very difficult to answer with strict scientific hypothesis testing. In other words, the public knows that science cannot fully determine whether to lock down for a prolonged period of time during a pandemic, because science cannot calculate the optimal tradeoff between, for example, the livelihoods of small business owners and broader public health risks. Similarly, science cannot determine the optimal rate of decarbonization of the global economy, because that would involve definitively weighing all the costs and benefits of fossil fuel use that involve technological, economic, ethical, philosophical, and moral questions far outside of anything that can be answered with the scientific method. 

When the scientific community overreaches by presenting their policy preferences as scientifically mandated, they confirm public suspicions, especially on the right, that a technocratic elite is attempting to dictate societal choices without proper democratic deliberation. Specifically, defining scientific institutions as opposing a popular political movement, like that behind Trump, does not convince the public that the popular political movement is wrong, but rather, it convinces the public that establishment science is fused with one side of the political aisle, and thus does not deserve trust. 

As citizens, scientists have just as much right as anybody to try to persuade fellow humans to adopt their political views, but they shouldn't do so under the guise that their political preferences are derived from some bedrock truth. In the context of scientific claims on climate change, Bill Maher recently articulated that “I really hate it when people try to manipulate me, try to move me in a certain direction; just tell me the truth.” Scientific institutions should take this idea to heart if they want to regain credibility, and there are encouraging signs that they are in fact doing this. 

Thus, the proper response to this election and the distrust of scientific institutions is not to double down on the idea that science is an authority that opposes Trump and the worldviews of the right. Instead, institutional science should respond differently this time. It should be much more humble about the limits of inquiry amenable to the scientific method and be honest that they don't dictate the correctness of various policies. That is the only way that trust in science was ever justified in the first place, and it is the only way to operate in the world during a second Trump presidency and beyond. 

Subscribe now

By Patrick Brown

In Part 1, I outlined the recent rise of the Degrowth Movement, where its supporters argue that the immense and urgent challenge of climate change necessitates the dismantling of capitalism and the deliberate reduction and redistribution of global wealth. I critiqued the Degrowth perspective for missing the critical role that economic productivity and technological progress plays in facilitating general climate adaptation. I showed that higher-income countries tend to be the ones that are most economically free and thus capitalistic, and it is these countries that are best suited to build infrastructure, respond to climate disasters, and have systems in place that are generally robust to our often hostile climate. I argued that capitalistic competition undergirds the technological innovation that causes direct climate adaptation, such as driving forward air conditioning, and supports the large tax base that produces nominally government-supported adaptation efforts like public infrastructure, early warning systems, and disaster response efforts.

In Part 2, I make the case that, just as capitalism is an asset for climate adaptation, it is also a key to transitioning to a sustainable low-carbon economy while simultaneously facilitating a high material standard of living.

Part 2: Mitigation of CO2 Emissions

One useful framework for thinking about drivers of global CO2 emission is the Kaya Identity, which breaks down human-caused CO2 emissions (from the energy system) into four fundamental components:

1. Human Population

2. Economic productivity (typically represented by GDP per capita)

3. Energy efficiency of the economy (energy use per unit of GDP)

4. Carbon intensity of energy (CO2 emissions per unit of energy).

Share

These four terms are multiplied so, theoretically, if any one of them were to go to zero, then human-caused CO2 emissions associated with the energy system would also go to zero. Thus, we could eliminate anthropogenic CO2 emissions and largely halt climate change by reducing the human population to zero, economic production to zero, by magically producing goods and services without energy, or by harnessing energy without emitting CO2.

This last term is the only one that could plausibly and desirably go to zero. However, the Degrowth Movement also heavily emphasizes putting downward pressure on all the other terms, especially economic productivity driven by capitalism. This emphasis is not without some logic because, historically, global CO2 emissions have been increasing, and they have been driven up primarily by an increase in population and economic productivity (this is despite increases in energy efficiency and reductions in carbon intensity of energy).

However, attempting to force down economic productivity in order to reduce CO2 emissions is parochial and misguided because it is actually economic productivity itself that puts downward pressure on all the other terms, making a sustainable society more plausible in the long term.

Let’s look at this in more detail, going through the four terms one by one.

Population

In order to achieve an indefinitely sustainable society on earth, it is fair to argue that the human population would need to stabilize at some level. Some environmentalists have long advanced the notion that the human population will increase exponentially as long as it has resources to consume, much like bacteria in a petri dish. In this model, population can only be stabilized through the blunt force of running out of resources, or it can be stabilized preemptively via top-down coercive policies that discourage reproduction.

But this model is wrong. Humans do not simply multiply uncontrollably until they consume all resources. In fact, It turns out that as societies become materially better off, fertility rates tend to decline naturally. This phenomenon apparently results from a multitude of factors associated with economic productivity, including increased access to education, especially for women, improved healthcare, and reduced child mortality rates. But capitalism has also played a more direct role as it is the most decentralized and capitalist economies that tends to have the lowest fertility rates.

One reason for this is that in capitalist systems, women have far greater motivation and incentives to invest in their human capital or career prospects. The disproportionate purchasing power that results from disproportionate investment in professional development motivates many women to prioritize careers over having larger families.

In addition to cultural drivers, the more proximate mechanism that greatly facilitates lower fertility rates—birth control largely in the form of oral contraception—has been driven forward by capitalism. The first birth control pill, Envoid, was developed in the 1950s by the G.D. Searle & Company. Since the introduction of Enovid, innovation in contraception has continued to advance as pharmaceutical companies have developed lower-dose birth control pills with reduced side effects. Bayer and Merck have developed IUDs that compete with birth control pills, and recently, there has been innovation in non-hormonal options, male contraception, and digital health technologies for tracking fertility. These innovations were not the result of a government mandate or concern for the environment, but rather resulted from private firms correctly identifying market opportunities to generate profit.

Energy Efficiency of the Economy

Increasing the energy efficiency of the economy means using less energy to produce the same amount of economic output, and this can be achieved through technological advancements, more efficient production processes, and a larger share of GDP coming from services rather than physical goods.

In a decentralized capitalistic economy, private firms seek to maximize profits by competing to offer goods and services at the lowest possible price. In this system, a penny saved is a penny earned, so a key lever for increasing profit is to reduce input costs to the production process. Since energy costs money, firms have an incentive to reduce direct energy expenditure as well as the embedded energy in the physical inputs to their production process. Furthermore, consumers have an incentive to demand products that are energy efficient. Nobody wants a TV that will double their monthly electricity bill.

The story of the modern smartphone, summarized by Andrew McAfee in More from Less, demonstrates this phenomenon. He highlights that the smartphone is a single compact device that replaces 13 out of 15 devices advertised in a 1991 Radio Shack ad.

This consolidation represents not only a reduction in physical materials, but also in the energy required to produce these services–not to mention the simultaneously increased quality of the services. Additionally, the brick-and-mortar RadioShack itself has been put out of business by more efficient means of shopping for and obtaining these services.

The creative destruction of all of these now-obsolete devices, and their retailer RadioShack, was not driven by a top-down mandate on energy efficiency, but rather by entrepreneurs like Steve Jobs (in the case of the smartphone) and Jeff Bezos (in the case of the demise of RadioShack) operating in a competitive capitalistic environment that demands innovation in efficiency.

These dynamics even foster reductions in energy use via the reuse of goods. Profit-seeking companies like Facebook have developed services like Facebook Marketplace that greatly facilitate the transfer of items between individuals, effectively extending their lifetimes and reducing waste and the demand to create new items.

The above are specific examples, but they are representative of broader trends observable at the aggregate economic level. In the highest-income countries, which also tend to be the most economically free, we have seen peaks and subsequent decreases in energy consumption not only per unit of GDP, but also on a per-person basis.

The motivation for creating the iPhone, Amazon, Facebook Marketplace (as well as the motivation to participate in Facebook Marketplace), and widespread efficiency innovations across all sectors of free capitalistic economies is not primarily a desire to reduce environmental impacts. Yet, the result has been an increase in efficiency so great that many more people have access to much higher-quality goods and services, even as absolute resource and energy use has decreased.

These innovations came about from people acting in their own economic self-interest within an institutional environment where private property and freely fluctuating prices signal supply and demand and, thus, the most efficient use of fundamentally scarce resources. Therefore, this progress comes about because of, not in spite of, economic self-interest. Yet in top-down communist regimes advocated for within the Degrowth movement, these price signals and economic self-interest would be abolished or severely thwarted, relying instead on technocratic control of production through coercion. As examples like the Soviet Union, China in the middle of the 20th century, North Korea, East Germany, and Venezuela have shown, this has historically proven to result in tremendous inefficiencies and a lack of incentive to innovate.

Carbon Intensity of Energy

Carbon intensity of energy can be reduced by deploying carbon capture and storage technologies and shifting from burning fossil fuels to energy sources like nuclear, hydro, wind, solar, and geothermal. Carbon intensity of energy is the only term in the Kaya Identity that could plausibly and desirably go to zero, and thus, it must play a central role in the discussion.

Global carbon intensity of energy has been decreasing gradually since the 1960s. The overall environment in which this is fostered is summarized in More from Less as having the four elements of 1) capitalism, 2) technological progress, 3) an informed public, and 4) responsive government. Publicly funded science led to the discovery and subsequent public awareness of the climate change problem, which in turn has led to democratically elected officials implementing government policies supporting R&D and commercialization of energy technologies that reduce the carbon intensity of energy. Ultimately, however, the technological progress necessary to offer low-carbon energy at affordable prices is and will be driven by capitalism.

In competitive capitalistic environments, energy producers attempt to reduce costs and improve efficiency to improve their prospects of securing government contracts and partnerships with utilities. Venture capitalists, motivated by the potential for high returns, fund these efforts, betting on various technologies' potential to meet future energy demands. As Thomas Friedman eloquently put it, “It's very hard to get a generation living today to make major sacrifices for a generation yet to be born, and that’s why you also need the profit motive.”

Examples of these dynamics abound. In wind energy, companies like Vestas, General Electric, and Siemens Gamesa are racing to produce turbines that capture more energy from the wind while reducing the cost per megawatt-hour (MWh).

In solar energy, companies like First Solar, LONGi Green Energy, and JinkoSolar are battling to produce ever-cheaper and more efficient panels. Even though it is in tension with the core ideology of the Chinese Communist Party, the Chinese-based solar industry has embraced a for-profit model where private firms compete against each other in order to motivate innovation.

Advanced geothermal energy is characterized by rapid innovation from companies like Fervo Energy, which uses horizontal drilling to access new resources, and Eavor Technologies, which developed a closed-loop system that eliminates the need for natural underground water reservoirs. TerraPower, Kairos Energy, and X-Energy are examples of competitors in advanced nuclear reactors, vying to create smaller, safer, and more efficient designs.

These innovations extend beyond just energy. Companies like CarbonCure (which injects captured CO2 into concrete for permanent mineralization) and Solidia Technologies (which cures cement with CO2 instead of water, cutting energy use and emissions) are competing to reduce the carbon footprint of cement production. Similarly, the electric vehicle (EV) industry features intense competition among Tesla, Rivian, and Lucid Motors. Rivian, backed by Amazon, emphasizes electric trucks and SUVs, while Lucid targets the luxury market with its Air sedan. Ford's Mustang Mach-E and F-150 Lightning, alongside GM's electrification plan for 2035, demonstrate traditional automakers striving for market share in the EV sector.

Just like in the realm of energy efficiency of the economy, competition for market dominance, funding, and contracts—driven ultimately by the profit motive—is responsible for producing the innovation we see in clean energy. And this explains why the most decentralized and capitalist economies tend to have the lowest carbon intensities of energy.

Economic Productivity (GDP per capita)

The last term on the Kaya Identity that we will investigate is economic productivity or GDP per capita. As we saw in Part 1, higher economic freedom is associated with higher economic productivity.

Societies characterized by elevated levels of economic freedom and productivity are largely responsible for fostering the technological innovations that enhance the energy efficiency of the economy while simultaneously reducing the carbon intensity of energy. A high material standard of living is a prerequisite for the aforementioned innovations, as it allows for human energy to be directed towards these higher-level goals rather than being exclusively focused on meeting basic material needs. However, after a certain standard of living is achieved, further GDP growth can be achieved while simultaneously reducing CO2 emissions (an example of the Environmental Kuznets Curve).

In the United States, for example, CO2 emissions peaked in 2007 and have been on a downward trend since, despite continued productivity growth. The US is not alone in this pattern and is one of at least 30 high-income countries that have achieved economic growth while reducing CO2 emissions in recent years.

Once a technology has been invented, it can be widely adopted without being invented again. Thus, groundbreaking ideas and innovations have widespread and lasting consequences, and it is in humanity’s best interest to maximize our collective ideas. There is more than sufficient reason to alleviate poverty for its own sake, but an additional benefit of poverty alleviation is that it will liberate billions of minds from solely focusing on the basic necessities of life. Some of those minds can then contribute to innovations in energy efficiency and zero-carbon technology, greatly expanding the possibilities of breakthrough technologies that might be adopted globally to reduce greenhouse gas emissions. In this sense, there is no reason to believe that global economic productivity is currently “sufficient.” Furthermore, as we have seen, economic productivity is critical for putting downward pressure on the three other components of the Kaya identity: Stabilizing population, increasing energy efficiency, and decarbonizing the energy system.

But economic productivity also seems to have breaks built into itself. Contrary to a prime argument advanced by the Degrowth Movement, the capitalistic system that undergirds economic productivity growth does not inherently require endless growth. As Harry Saunders has shown, growth is not a tenet of capitalism in any formal mathematical sense, and in the real world, we observe that once societies reach a certain level of productivity, their growth rates actually begin to decline.

If human appetites for goods and services were insatiable, we would see the highest-income countries having the highest growth rates. But as people become materially better off, they choose to work fewer hours, effectively sacrificing purchasing power and consumption for leisure time. This allows for GDP growth itself to slow without top-down coercion from a centralized authority.

Conclusion

Within the Degrowth Movement, it is presumed that if humanity is left to its own devices, decentralized economic activity will result in ever-expanding resource extraction rates and ever-expanding greenhouse gas emissions. Thus, the supposed solution is to centralize control of the global economy in order to force people (ultimately through coercion) to not engage in transactions they want to engage in. Public Choice theory explains why it is a serious error to assume that a global state, given centralized control of the economy, would act in a way that optimizes social outcomes. Those in government, just like those outside of it, tend to act in self-interested ways. When control of the economy is centralized, efforts of the ambitious and talented tend to go towards gaining political power or favor with those in authority, rather than on out-innovating competitors in the markeplace.

This is ultimately why capitalism is so successful, not only at increasing productivity overall, but also at facilitating climate change mitigation: it channels natural human self-interest in a way that is socially beneficial. These benefits are relatively straightforward and intuitive in the realm of climate adaptation, but we see here that they also extend to the reduction of greenhouse gas emissions. Specifically, private property, markets, and prices actually work to reduce fertility rates, improve energy efficiency, and reduce the carbon intensity of energy. Thus, decentralized economies do not inherently result in indefinite climate change, and instead, they may be the best system available to stabilize the climate while supporting the continued improvement of human welfare.

Subscribe now

On September 23rd, the New York Times ran a David Wallace-Wells column called “Our Adaptation to Global Warming Is Largely Fictional” featuring “a new, eye-opening paper,” entitled “Are We Adapting to Climate Change?” The paper purports to show that there is limited evidence of human adaptation to climate change across a wide range of sectors, including human deaths, agricultural productivity, crime, conflict, economic output, and damages from flooding and tropical cyclones.

The paper supposedly counters the Pollyannaish view that “adaptation has been a hugely underappreciated success story.”  

Count me as someone who thinks that “adaptation has been a hugely underappreciated success story.” However, this is not just some assumption that is “taken for granted” but is instead based on examining large-scale empirical trends in climate-sensitive outcomes. For example, as it has warmed over the past several decades, we have observed almost universally higher crop yields, translating to more calories available per person, and a reduction in death rates from malnutrition and famines. Access to safe drinking water has increased, while the prevalence of climate-sensitive diseases like malaria has decreased. Furthermore, mortality rates due to extreme temperatures—both cold and heat—have declined, as have deaths from natural disasters. Finally, we have seen precipitous declines in damage from natural disasters as a fraction of the value exposed. 

So what’s going on? Are we adapting to climate change or not? The “new, eye-opening" paper argues we have not, but it does so in a misleading way.

Share

First, let me offer what I think are intuitive, common-sense definitions of adapting to climate change:

• Adaptation: Climate is changing, and we are adjusting fast enough so that outcomes stay the same over time.

• No adaptation: Climate is changing, and we are not adjusting; thus, outcomes are getting worse over time.

• Mal-adaptation: Climate is changing, and we are changing in ways that exacerbate climate change problems, worsening outcomes more than they would have been from climate change alone.

• Super-adaptation: Climate is changing, and we are changing even faster (not necessarily because of climate change), so outcomes are improving over time.

One important nuance is that society's exposure to various climates and extreme weather hazards is also changing over time. After all, there are three times more people on the planet than there were in 1950. So, a population increase in hot parts of the world might mean more deaths from extreme heat simply because there are more people exposed to extreme heat, and that wouldn't necessarily imply mal-adaptation to climate change. Thus, studies on changes in climate-related outcomes over time should express outcomes in rates like “deaths per number of people exposed.” When studies define adaptation in this broad (I’d argue common sense) way, they overwhelmingly show super-adaptation. For example, empirical evidence of declining global vulnerability to climate-related hazards showed clear globally declining death rates and economic damage rates for floods, droughts, extreme temperatures, and extreme winds. 

However, this broad definition of adaptation is not common in the academic literature, where adaptation tends to be narrowly defined as only those technologies or behaviors that are more beneficial in a changed climate than they are in an unchanged one. For example, according to some papers, if technology increases crop yields equally in the current and future climates, it should not be considered adaptation. Definitions like this allow for striking claims like the IPCC’s that adaptation is supposedly “insufficient to offset the negative impacts of climate change” for crop yields despite the fact that crop yields have been increasing steadily as the climate has warmed.

So is this narrow definition of climate adaptation the reason why the aforementioned study covered by the New York Times finds such minimal effect of adaptation? 

Not quite. In the paper, they say that they take a broad view of adaptation, but are able to present a negative view of our adaptive capacities by focusing on the relative changes in adaptation within each decade:

“To quantify the net effect of these and any other adaptive actions that could have taken place, we estimate whether the sensitivity of a range of societal outcomes to a fixed change in climate has changed over time.”

So, in this case, it’s not an intentionally narrow definition that accounts for the small effectiveness of adaptation that they calculate. Instead, the apparent discrepancy comes down to the word sensitivity above, which is doing a tremendous amount of work. In the paper, sensitivity refers to the response of an outcome for a given climate fluctuation, but critically, it is relative to the decade average for that outcome. 

Share

Let’s illustrate with an example. The paper finds maladaptation in some outcomes, like Brazilian soybean yields. This may sound odd to people familiar with South American agriculture because Brazilian soybean yields have tripled since the 1960s as Brazil has warmed.

This has been driven by growing global demand stimulating productivity improvements and thus enhancing supply. Specifically, Brazilian soybean yields have benefited from improved crop varieties, including genetically modified versions that are more resistant to pests, diseases, and adverse weather conditions, fertilizer use, precise irrigation management, and the mechanization of farming. These technological advances have overwhelmed any negative impact from climate change, so how does the paper find mal-adaptation?

Well, in the 1970s, Brazil used to produce 1.5 tonnes/hectare of soybeans for growing seasons that were normal and 1.42 tonnes/hectare of soybeans for growing seasons that were 1°C hotter than normal (26°C) or a 5% decline. 

By the 2010s, Brazil’s average productivity had doubled to 3.0 tonnes/hectare of soybeans for normal growing seasons but was reduced to 2.67 tonnes/hectare for growing seasons that were 1°C hotter year than normal (27°C). 

So that's an 88% increase in output for a given +1°C variation! However, the paper’s methodology defines sensitivities as being relative to each decade. Since 11% is larger than 5%, it concludes that we have become less adapted to climate change over time. 

The New York Times article writes that, 

“On three-quarters of the impacts they studied, our vulnerability to warming hasn’t improved at all, meaning that a given climate event would be just as damaging as 50 years ago — perhaps more damaging.” 

But this statement will mislead the vast majority of readers. The “more damaging” phrase is only true if it is interpreted as “more damaging relative to the constantly rising background level,” it’s not “more damaging” than 50 years ago in any absolute sense. The same is true for the other outcomes they study, like death rates from floods, and extreme temperatures or economic damage. When decreases in vulnerability emerge over decades, they are not allowed to count as adaptation and thus will be neglected by this study.

Investigating the question of whether relative sensitivities for adaptive capacity are changing over time is a fine exercise for an academic paper that explicitly acknowledges the long-term improvement in outcomes. But shifting the goal posts of what adaptation means to present an overly negative portrayal is something else. Further, the use of the word adaptation only serves to confuse because a definition of adaptation that is intuitive to most people would simply measure if we are becoming more or less vulnerable to given weather-related extremes compared to the past. 

This common denigration of adaptation does no service to us because it misinforms the public and policymakers and diverts attention away from repeatable success stories. Adaptation to the climate is a function of decisions we collectively make and should remain a priority for humanity. The improvements in adaptation of the past decades are repeatable, but only if we understand them as success stories, and they are not denigrated to sell a narrative that our society is collapsing under the weight of climate change.

Subscribe now

By Patrick Brown

Mainstream economic discussions, whether in the context of elections or central bank interest rates, take for granted that the underlying goal is to facilitate and stimulate the overall economy. However, there has long been a faction of the environmental left that rejects this goal outright. Undergirded by themes of catastrophic resource depletion articulated by Thomas Malthus in the 1700s and later in the 1960s and 1970s in Paul Ehrlich’s The Population Bomb, and in the Club of Rome’s The Limits to Growth, these circles have placed economic growth firmly in the crosshairs, framing it not as a measure of society's standard of living but rather as wasteful self-destructive consumption.

These views have persisted largely outside of the mainstream, but they are enjoying a resurgence in recent years. In the current context of a great deal of hyperbole on climate change impacts, the idea that reducing greenhouse gas emissions should trump all other concerns and that economic growth is a major enemy in this fight has become fashionable for academics and climate activists. These arguments are often intertwined with attacks on capitalism, which is considered the system responsible for unsustainable growth and all sorts of immorality. This theme was apparent in the title of Naomi Klein’s popular 2015 book Capitalism vs. The Climate, but momentum has been building since then, represented by Kate Raworth’s 2017 Doughnut Economics, Jason Hickel’s 2020 Less is More, and Kohei Saito's 2024 Slow Down. Falling under the umbrella of the Degrowth movement, Hickel and Saito argue for a kind of world communism that is aimed explicitly at dismantling capitalism and reducing global economic activity. Though these ideas may seem quite extreme to the average person, they are receiving increasingly favorable coverage in venues such as the New York Times and Nature, and they are being bolstered by a perception of credibility from a rapidly expanding field of academic research (that is, in fact, quite flawed). 

Because the hostility towards both economic productivity (framed as wasteful consumption) and personal economic freedom in the form of capitalism seem to be ever in the background and enjoying a resurgence, it’s worth revisiting, from first principles, why both economic productivity and capitalism are assets, not liabilities in the context of general human material well-being and in the context of climate. 

In Part 1, I’ll look at the case of adapting to climate and climate change, specifically how private property, markets, and trade (i.e., capitalism) tend to be great at facilitating productivity, and this, in turn, results in the societies that are least vulnerable to climate and climate change. In Part 2, I’ll look at how capitalism and the economic and technological progress it propels are essential for the eventual elimination of greenhouse gas emissions.

Share

Part 1: Adaptation to Climate

The degrowth movement asserts that we can only hope to address our climate challenges by reducing economic activity to reduce greenhouse gas emissions and other resource draws on the earth, typically through some form of global control of the economy. However, this would be incredibly counterproductive for climate adaptation as it would dismantle the primary phenomena responsible for increasing the human material standard of living that fortifies us against our often hostile climate.

Economic productivity is typically measured with gross domestic product (GDP) per capita (the total per-person monetary value of all goods and services produced). GDP per capita is a popular target of critics. It is certainly an imperfect measure of total human well-being, but it represents a good proxy for a society's material standard of living. In fact, GDP per capita is strongly correlated with many outcomes almost universally considered socially desirable, including higher life expectancy, lower child mortality, higher educational attainment, fewer working hours, and higher self-reported life satisfaction.

 Furthermore, we see a strong relationship between GDP per capita and reduced vulnerability to climate over space and time. Over time, almost all climate-sensitive aspects of society have been trending in generally positive directions as GDP per capita has increased. Over space, death rates from natural disasters like floods, droughts, and storms are 15 times higher for low-productivity countries than they are for high-productivity countries, and disaster damages as a percentage of exposed GDP are much lower in high-productivity nations. Finally, there is a strong across-country relationship between GDP per capita and measures of climate change adaptive capacity, such as the University of Notre Dame’s climate adaptation Readiness Score.

While the underlying productivity of a society is the foundation, governments play a critical role in providing the fundamental infrastructure and systems related to energy, transport, communication, health, and education that are upstream of resilience to the climate. The well-resourced governments of high-productivity economies are also able to provide many services that are more directly related to climate adaptation. For example, the US has the luxury of supporting advanced weather and climate prediction centers like the National Center for Environmental Prediction (NCEP); it has agencies like the Federal Emergency Management Agency (FEMA), which coordinates disaster preparedness, response, and recovery efforts; it has agencies like the Army Corps of Engineers which designs, builds, and maintains large-scale flood protection; it has agencies like the U.S. Forest Service Fire and Aviation Management Program which fights forest fires, and it has agencies like the United States Department of Agriculture which spends billions annually on research and development (and should probably spend more) that has supported breakthroughs in crop varieties, precision farming technologies, and improved fertilizers. These are all nominally government programs, but they are ultimately tied to the productivity of the private sector that funds these endeavors. Furthermore, all these programs buy the tools they use (vehicles, airplanes, computers, software, construction materials, bulldozers, etc.) from the private sector, and much of the public R&D is commercialized and implemented via private entities.

Share

The Facilitator of Economic Productivity

Throughout the vast majority of human history, material poverty and extreme sensitivity to the whims of the climate were the universal state of the world. As measured in GDP per capita, there was virtually no economic growth until the commercial revolution in the late Middle Ages, when economic trade began to better facilitate productivity and innovation. As the 19th-century economist Frédéric Bastiat noted, “In the absence of trade and production, life itself would be impossible. Nature does not supply us with ready-made products, but only with the materials from which we must fashion the means of our existence.”

In recent centuries and decades, humanity has been rapidly becoming materially much better off. The average person alive today is approximately five times richer than the average person in 1950, with parallel declines in the share of the global population in extreme poverty. So, what is the engine behind this explosion in economic productivity? As Adam Smith famously observed, productivity is greatly enhanced by the system that we now call capitalism. In particular, in a system where individuals and firms know they can own, control, and disproportionately benefit from disproportionate productivity, they are incentivized to work harder, think more creatively, and invest their money and their time (including investment in themselves) more wisely. Even Karl Marx emphasized that capitalist competition amongst private firms is largely responsible for stimulating the technological change that increases productivity (in terms of production per hour of labor).

In recent history, examples such as the United States, Germany, Switzerland, and Japan show the success of mixed economies or “welfare-state capitalism” with significant government intervention but robust private sectors embracing capitalism. In stark contrast, fully planned command economies have a track record of failure. Some of the 20th century's more famous examples include the Soviet Union’s economic stagnation, the economic and human disaster of Mao Zedong's centralized attempt to collectivize Chinese agriculture, and the contrast of economic success in capitalist West Germany compared to socialist East Germany. More recently, Venezuela’s expansion of socialism has been marked by economic decline and crisis.

Of course, various economic systems spanning the range from more market-oriented to those with much stronger levels of state intervention and control have been implemented, and more laissez-faire is not necessarily better. South Korea, Taiwan, and Singapore provide examples of rapid economic development in the latter half of the 20th under autocratic rulers who used formal economic planning. However, this planning relied on incentives and support not dictatorial mandates, and the vast majority of the economies in these countries remained under private control. 

Overall, the field of developmental economics has continually identified strong institutions that enforce property rights and allow for open markets and free trade as critical facilitators of productivity. In systems where a government ensures that property rights are secure, individuals are free to specialize in the areas where they have the most skill or interest, and this specialization leads to greater efficiency and productivity as people focus on what they do best with an eye to producing what their fellow citizens desire most. Competition is the key to driving efficiency and innovation, so governments also play a critical role by inhibiting monopolies and price fixing. 

Critically, freely fluctuating prices reflect people’s needs and realistic production constraints without explicit calculation. What’s produced is determined by the prices people are willing to pay for products, and prices for raw materials, labor, loans, rent, and other costs determine the most efficient way of producing those products. These same prices determine how much each person has to spend in the market. Prices vary and thus reflect changing conditions. When a desirable product becomes scarce, its price rises, which stimulates more resources to flow towards producing that product (or substitutes for that product), reducing the scarcity. 

These dynamics are why astrongassociation exists between a country's degree of economic freedom and its economic productivity, and there is good evidence that this relationship is causal. Below, the current across-country relationship is visualized using GDP per capita and economic freedom estimated by the Fraser Institute’s Economic Freedom Index (which gives points for, e.g., the security of private property rights, freedom to trade internationally, and penalties for, e.g., government ownership of industry and overregulation of business).

Finally, since economic freedom is associated with economic productivity and we saw above that economic productivity is associated with climate adaptation readiness, it is not surprising that there is also a strong across-country relationship between economic freedom, and climate adaptation readiness.

Overall, the three-way positive association between economic productivity, economic freedom, and climate change adaptation readiness indicates that the market-based, high-productivity societies of the world are those that not only have the best current protection from the climate but those that will be the most agile in their adaptation to climate change.

The Agility of Capitalism in Climate Adaptation

In his defense of a command economy, Karl Marx articulated that only that which is planned is rational. When it came to attempts to implement Marx’s ideas, the most consequential hurdle has been finding benevolent autocrats and planners. However, more relevant to the issue of climate adaptation was Marx’s supposition that technocratic planners could have sufficient information and knowledge to be effectively rational, even if they were benevolent. In contrast, Ludwig von Mises argued in 1920 that “socialism is the abolition of rational economy,” making the case that without private ownership and free-market pricing mechanisms, there is no way of calculating the most “rational” allocation of resources. A survey of the landscape of climate adaptive measures in productive economies lends much more support to von Mises than to Marx.

For example, as Anderson et al. discuss in their 2018 paper The Critical Role of markets in climate adaptation, markets in agriculture effectively signal emerging climate risks, prompting adjustments such as changes in cropping patterns, irrigation practices, and even the geographic shift of agricultural production to less vulnerable areas. If certain crops become less viable in one region, they can be grown elsewhere and traded on the global market, mediating any effect on the total food supply. This is a specific example of the more general principle of Ricardian comparative advantage, where, in this case, capitalism naturally facilitates the efficient distribution of production across the globe, allowing regions to specialize in the goods and services for which they are best suited. 

Prices for labor (i.e., wages) also facilitate adaptation to climate change. As new challenges arise, the demand for certain skills and occupations changes, creating incentives for workers to acquire new skills and transition into industries critical for climate adaptation. For example, in agriculture, the need for agronomists and environmental scientists to develop and implement new climate-resilient practices may increase, driving labor toward these areas.

These dynamics tend to be mediated through private firms. John Deere, for example, has been behind the development of numerous machines that have improved crop productivity, including recent precision agriculture aided by GPS-guided tractors and other “smart farming” equipment that allows farmers to better optimize planting, irrigation, and harvesting. Firms like Nutrien and Yara International have developed advanced fertilizers that increase crop productivity, and companies like Bayer have innovated in gene editing, facilitating the creation of drought-resistant and pest-resistant crops.

When it comes to the infrastructure that is critical for climate resilience, basics like steel were famously developed by icons of capitalism like Andrew Carnegie. More recently, private firms like LafargeHolcim and CEMEX are behind the development of high-performance concrete, Saint-Gobain is behind the production of tempered glass five times stronger than normal glass, and buildings, bridges, dikes, dams, etc., are assembled at relatively low cost with the help of Caterpillar and Komatsu.

Air conditioning has been critical for human protection from heat, and this technology, too, has been greatly pushed forward by private-sector competition. While working for the Buffalo Forge Company in 1902, Willis Carrier invented the first modern air conditioning system and then founded Carrier Engineering Corporation. They innovated sufficiently to produce products suitable for cooling large spaces. However, Frigidaire was the first to introduce a unit small enough for residences in 1929, and its design was supplanted by one from General Electric in the early 1930s. Other improvements by General Motors primed the technology for an explosion in adoption in the US in the middle of the century. Demand for cooling systems was large due mostly to the climate not being tuned to optimum human temperatures (not necessarily climate change), so companies like Carrier, Trane, Daikin, and Lennox entered the market, competing to improve efficiency and affordability. Through this competition, innovations like variable refrigerant flow (VRF) systems, smart thermostats, and energy-efficient compressors were developed.

These innovations have helped make air conditioning accessible to a broad population, contributing to widespread adoption and protection from extreme heat that has undoubtedly saved an enormous number of lives. A common critique of capitalism is that it leaves the poor behind, but as Joseph Schumpeter argued, the record of capitalism is one of unprecedented economic growth that benefits the poor the most. As he put it “The capitalist achievement does not typically consist in providing more silk stockings for queens but in bringing them within steady reach of factory girls in return for steadily decreasing amounts of effort.” Echoing this sentiment, Matthew Khan makes the same case in the context of climate adaptation via air conditioning:

“The market for air conditioning is an example. As outdoor temperatures rise, people will increasingly demand more powerful air conditioning. This aggregate demand creates a market for firms to design products to meet this demand. Such mass production leads to economies of scale, which lowers the cost per unit produced. As a result, lower-income people can afford these goods. As lower-income people purchase such adaptation-friendly products, they suffer less from extreme weather events. Academics using big data can rigorously test whether the damage caused by such extreme weather is shrinking for poorer people. If this is the case, then this is direct evidence that capitalism is playing a productive role in protecting the most vulnerable people from emerging threats.”

Climate adaptation has been a resounding success in the modern era of rapid capitalistic economic growth, but we still have a very long way to go. Critics of capitalism, including those in the degrowth movement, often advocate for revolution on the grounds that the status quo is imperfect. They frequently echo Jean-Jacques Rousseau’s famous sentiment that “Man is naturally good, and it is by his institutions alone that he becomes evil.” In putting forward their proposed alternatives, however, they are often committing the Nirvana Fallacy, where the supposed choice is between the imperfect status quo and a perfect utopian solution rather than the choice being between alternative real-world arrangements. 

But we live in the real world, and thus, we must compare systems as they actually operate without being overly indulgent in our fantasies about how they might ideally operate. In the real world, Bastiat is correct that poverty is the base state of humanity, and it would appear that Smith is correct that a critical component for alleviating that poverty is to channel natural human self-interest through systems that rely on private ownership and freely fluctuating prices that represent realistic constraints on supply and demand. Historically, this formula has pushed great progress, and there is scant evidence that the anti-capitalist degrowth model would do anything other than rapidly unwind this progress and return us to a state of greater material poverty and vulnerability to the climate.

In Part 2, I will shift the focus to the equally important topic of greenhouse gas emissions mitigation, which is critical for the stabilization of the climate. I will argue that market-driven technological innovation seen in capitalist economies is necessary for the production of an energy and agriculture system that produces no CO2 emissions and that trying to achieve climate goals through global top-down controls on population and/or economic activity is misguided. 

By Patrick Brown

Wildfire season is ramping up across the United States, and nowhere is that more apparent than in northern California, where the Park Fire has been burning since July 24th. As of today, August 14th, the Park Fire has burned nearly 430,000 acres (672 square miles), or about 65% of the size of the state of Rhode Island. It is still only 39% contained and has destroyed over 600 structures.

The Park Fire currently stands as California’s fourth-largest fire since meticulous record-keeping began in the 1980s, and by itself, it has burned more area than that from all California fires in the calendar years of either 2022 or 2023

Throughout its lifetime, it has exhibited extreme fire behavior, like giant rotating plumes of smoke, and its severity will undoubtedly be catastrophic, with widespread tree mortality in many of the forests it has burned.

Large catastrophic wildfires like the Park Fire are now typically covered through the lens of climate change as this tends to be the most salient angle for consumers of the news. This is the same reason why researchers themselves are incentivized to frame their wildfire-related research questions in a way that centers climate change. But wildfire outcomes are the result of many factors, including how we manage the landscape. Our research, discussed below, indicates that land management choices have more leverage on wildfires like the Park Fire than climate policy does over the next several decades.

Climate Change is Just One of Many Drivers of Wildfire Behavior

Climate change is just one element of many that affect wildfire behavior and outcomes, and focusing exclusively on it distracts from practical solutions to increasing wildfire danger.

First, fires need to be ignited, and 84% of wildfires in the U.S. are started by people. The Park Fire was apparently deliberately started by a man who pushed a burning car into a ravine. Once started, wildfire behavior is well known to result from environmental influences that can be divided into the categories of topography, weather, and fuels (i.e., the live and dead vegetation that fires burn)

The Wildfire Behavior Triangle describes the three main categorical drivers of wildfire behavior.

Humans have little effect on topography, but can affect the weather by emitting greenhouse gasses and warming the global climate. Summers in the region where the Park Fire is currently burning have warmed at about 0.8°F per decade since 1975, and according to climate models, we expect them to warm about 1.5°F between now and mid-century. This warming decreases the moisture in fuels, causing fires to burn more intensely and spread faster.

The other major way humans influence wildfire behavior is through our influence on the vegetation that fuels fires. First, we fight wildfires primarily by interrupting their fuel source.

We use bulldozers and heavy equipment, combined with setting intentional fires, to “contain” fires by breaking up the continuity of their fuel. The Park Fire, for example, has had hundreds of bulldozers and many thousands of personnel deployed, creating these lines to contain the fire

This way of breaking up the fuel continuity in an effort to fight an active fire might be considered reactive fuel reduction, but we can also proactively manage the landscape to reduce fuels.

The Park Fire is currently burning in forests at two ends of the extreme of the degree of human intervention on the landscape, and neither is ideal for reducing wildfire danger. One end of the extreme is essentially the lack of active management on the US Forest Service land in Lassen National Forest, and the other is management optimized for timber harvests on the private land owned by Sierra Pacific Industries (SPI)

Forests Managed for Timber Production Exacerbate Wildfire Danger.

The forests managed for intensive timber production in the Sierra Pacific Industries (SPI) territory are marked by the ~20-acre patches of clear-cut areas. These “group selection” patches are meant to promote the natural regeneration of trees from nearby mature live trees while maintaining habitat for various species and reducing the aesthetic issues associated with clearcutting larger swaths

On the one hand, these patches do allow for firefighters and flame retardants to have good access to an active firefront. However, open patches also tend to allow more sun and wind at the surface, which can facilitate faster fire spread, especially if dry brush is allowed to build up. Additionally, relatively tightly packed, younger tree stands (designed to maximize revenue from timber) are particularly sensitive to high-severity catastrophic fires. While SPI has engaged in some thinning practices and put in strategic fuel breaks designed for wildfire mitigation, this is not the same as managing forests specifically to reduce the risk of catastrophic fires where, e.g., only smaller trees less than 2 feet in diameter are removed.

Unmanaged Forests Exacerbate Wildfire Danger

On the other end of the spectrum is the non-management exemplified by the portion of the Park Fire burning within the boundaries of the Lassen National Forest. Here, we see the manifestation of ill-advised policies starting in the early 20th century that sought to exclude all fires from the landscape. These policies have led to a large build-up of tree and vegetation density in much of America’s forests

This exacerbates wildfire danger for several reasons. Densely packed vegetation means increased competition for finite resources (like water, sunlight, and nutrients) that make ecosystems more susceptible to disease and sensitive to drought, which increases their proclivity to burn. It also makes fires more likely to climb into the canopies of forests, where they can kill the entire forest.

The image below shows the consequences of this. In the 1800s, frequent fires sustained open Ponderosa Pine forests northwest of Colorado Springs, but a century of fire exclusion produced dense forests by the year 2000 that were devastated by the Hayman Fire in 2002. The forest has yet to recover

More vegetation also means fires burn more intensely, produce more smoke, a huge detriment to human health, and create more ash, which harms municipal water supplies. Finally, a densely packed forest makes it more difficult for firefighters and fire retardants to access fires.

This legacy of fire exclusion is affecting the Park Fire, whose active northeastern flank has been burning in a region that has not experienced a fire in about a century. This is a similar setting to that of the devastating Dixie Fire of 2021, which burned nearly 1 million acres, caused over a billion dollars in damage, and resulted in widespread tree mortality

Specifically, the active northeastern edge of the Park Fire, which is the main area of concern for further growth (pink arrow), is in a region with very high tree density and, thus, very high amounts of fuel available for the fire

This tightly-packed forest with ample fuels was identified by the California Wildfire Taskforce several years ago as having significant wildfire hazard potential and being prone to high-severity fires that will kill most of the trees

Rather than hastily removing fuels under critical fire-weather conditions to contain an active blaze, these fuels can be removed proactively with mechanical thinning (using machines to remove vegetation) and prescribed burning (intentionally setting fire to the landscape under much more mild conditions when the fire can be controlled).

The idea is that mechanical thinning to remove trees with smaller diameters and prescribed burning to remove live and dead surface vegetation will reduce fire intensity, produce less smoke, and be less likely to kill mature trees. Some wood products can be sold from mechanical thinning to offset the cost of conducting these treatments, but this is quite different from the aforementioned “group selection” land management specifically for timber harvesting.

The above diagram represents the theory for why this type of fuel reduction should work, but there is also ample empirical evidence that it works in practice. This is exemplified by meta-analyses of dozens of studies, but a picture says 1,000 words. Below is a photo of the after-effects of the Bootleg Fire in Oregon, which happened to traverse an area that had no fuel reduction, an area that had recently undergone only mechanical thinning, and an area that had recently undergone both mechanical thinning and prescribed burning.

The effects on severity were stark. Without fuel treatment, the fire killed the vast majority of mature trees, devastating the forest. The forest came out in better shape in the area that experienced only mechanical thinning, but the most benefit was seen where both mechanical thinning and prescribed burning occurred.

We see this same effect currently in the firefighting operations on the Park Fire, where areas of minimized surface vegetation and large, widely-spaced trees are experiencing only low severity fire.

How Much Would Fuel Reduction Reduce the Intensity of the Park Fire?

The above constitutes evidence that fuel reduction makes a difference, but how much of a difference does it make, and how does this compare to the other effect that gets the most attention—climate change? We are conducting original research that seeks to answer these questions.

We use machine learning to untangle the complex relationships between the weather, topographic, and fuel variables and the resulting fire intensity using over 27,000 satellite observations of fires. Once the relationships are quantified, we can predict fire intensity for any conditions, like, say, over the next couple of days across the entire state of California. We can also take conditions over the course of a fire and warm the background climate to see what we expect fire intensities to look like in the future and/or simulate fuel reduction treatments to see how future warming and fuel treatments would simultaneously affect a fire over its lifetime.

Below, this procedure is applied to the Park Fire. The black line shows the calculated potential for fire intensity over the Park Fire's lifetime (using a weather forecast for a few days in the future). We see tremendous variability driven by changes in weather, topography, and underlying fuels over the course of the fire.

In particular, we see daily peaks in fire intensity as it gets warmer and drier in the afternoon, especially when this coincides with strong winds. Also, the amount of fuel and how dry that fuel is are huge drivers of fire intensity potential

In addition to the black line, which represents the severity of the Park Fire under today’s climate and fuel conditions, the colored lines represent the same Park Fire progression in 2050 temperature conditions, with and without hazardous fuel reduction (fuel reduction is represented using 10 characteristics shown in Extended Data Table 2, here).

In 2050, the temperature conditions will be about 1.5°F warmer, which translates into drier fuels more conducive for higher-intensity fires. The red and magenta lines show the situation without hazardous fuel reduction. We see that the red and magenta Fire Intensity Potential lines are almost universally higher than the black Fire Intensity Potential line, indicating that the warming in the 2050s does increase fire intensity. Sometimes this climate change component is quite large. For example, on August 5th, this increase was about 25% from around 400 megawatts to around 500 megawatts. However, hourly and daily weather variability is much larger than this climate change signal.

It is important to note that two different greenhouse gas emissions scenarios are shown. One is a slow emissions reduction scenario that roughly represents current global policies (SSP2-4.5), and one represents fast emissions reductions that are much more aligned with the Paris Agreement (SSP1-2.6). The movement from the slow emissions reduction scenario to the Paris Agreement scenario represents major differences in the global energy and agricultural economies, technology adoption, geopolitics, etc. Yet, these differences translate into only marginal changes in Park Fire’s 2050 intensity.

When it comes to addressing the wildfire crisis, greenhouse gas emissions reductions are often heralded as the primary lever we can pull. However, trying to influence weather-related outcomes through energy policy is incredibly indirect and there are often much more direct solutions at hand.

In that vein, the blue line represents the progression of the Park Fire in 2050 under slow emissions reductions (4.5) but with hazardous fuel reduction. Not only is the blue line always below the red and magenta lines, meaning that it would reduce fire intensity over the fire's entire lifetime, but it is also typically below the black line representing the Park Fire in today’s climate.

This means that even in a situation of 2050 warmth, under a slow emissions reduction scenario, hazardous fuel reduction would reduce the Park Fire’s intensity relative to today.

We Need Reform to Scale up Hazardous Fuel Reduction Treatments.

The above is specific to the Park Fire but we find that these results are generally representative for the state of California as a whole over a range of conditions. We also calculate that these types of fuel reductions, though expensive, are nonetheless very likely to induce a net economic benefit.

This is evidence that California’s goal to increase hazardous fuels reduction treatments to 1 million acres per year, along with the federal government’s target of reducing fuels on 50 million acres over a decade, are laudable and would be helpful in combating the growing wildfire threat.

However, actually achieving these goals requires overcoming multiple challenges, including funding constraints, workforce shortages, and logistical issues related to complex land ownership patterns.

Finally, there are bureaucratic and regulatory obstacles associated with the California Environmental Quality Act (CEQA) on California state land and the National Environmental Policy Act (NEPA) on federal land. The NEPA process, in particular, is associated with significant obstacles to fuel reduction treatments. The Forest Service writes the most NEPA reviews of any federal agency and is the most likely to be sued, resulting in litigation that delays project implication by about 3 years on average.

The recently adopted California Vegetation Treatment Program (CalVTP) is designed to reduce the time and resources needed for individual project alignment with CEQA, and there is momentum for reform of NEPA to increase the pace and scale of hazardous fuels reduction treatments on federal land.

These reforms offer some promise, but as it stands, actually achieving the articulated acreage targets for fuel reduction will continue to face significant funding, coordination, political, and legal hurdles. All efforts should be made to overcome these barriers because the evidence shows that hazardous fuel reduction would have reduced the intensity of the Park Fire and wildfires more generally. Further, we show that we could have a future of reduced wildfire danger despite the inevitable warming that is in the pipeline and regardless of the specifics of global climate policy. Ultimately, achieving this future involves shifting away from land management schemes that couple fire exclusion with a hands-off approach or those that seek to maximize timber production and towards management plans that are explicitly designed to optimize forest health and wildfire resilience.

Subscribe now

By Patrick Brown

Fire season is ramping up into full gear in the western United States (and Canada), with widespread impacts on lives, property, and health via exposure to smoke. These worsening impacts have become routine in recent years and are part of a trend over recent decades. In the continental US, for example, approximately 2 million acres of land burned in a typical year in the 1960s through the 1990s, but that number has more than tripled to around 7 million acres per year over the last couple of decades. 

Humans start about 85% of wildfires, and thus the expansion of settlements into more fire-prone areas has played a role. However, the two main drivers of the increase in wildfire activity over the past several decades are a warming and drying climate and the long-term build-up of excess vegetation in much of our forests. 

One primary strategy for dealing with the increase in wildfire danger is to engage in proactive land management—namely, prescribed burning and mechanical thinning to reduce hazardous fuels and mitigate wildfire danger. 

The case for doing this, often referred to as Forest Restoration, includes “activities that…return forest composition and structure to a more natural state…” The argument goes that we have erred by excluding fires and introducing invasive grass species, which have created unnatural, unhealthy conditions; therefore, we must actively return the landscape to more natural conditions. 

In reality, the fire regimes in the U.S. West prior to modern fire exclusion policies were heavily influenced by Indigenous burning practices. However, these practices, too, get implicitly categorized as being essentially natural due to perceptions of Indigenous people living in harmony with nature.

Ironically, the case against proactive land management is also made in the name of naturalness. Environmental groups that seek to stop active land management projects typically do so on the grounds that human intervention moves the landscape from a perceived ideal natural state to a less ideal unnatural state. Many environmentalists argue that we must neither manage forests nor prevent fires and rather let “nature” simply return to its “natural order.”

Both the case for “natural” passive land management and the case against “unnatural” active land management mistake the core reason for wildfire mitigation: to protect human lives, property, and health while balancing other interests like habitat maintenance for certain species, recreation, sustainable timber harvesting, etc.

“Natural” is an arbitrary category and not necessarily good

Using the degree of naturalness as an argument for or against active land management is an application of the naturalistic fallacy, which asserts that what is natural is inherently good. 

The naturalistic fallacy is omnipresent in contemporary discussions around energy, environment, and food issues. For instance, “chemicals” and genetically modified organisms are often vilified for being "unnatural," while organic farming is praised despite its lower resource efficiency and higher land use. Similarly, renewable energy sources like wind and solar get favored over nuclear power in many circles due to their perceived naturalness, despite nuclear power's lower land and material footprint and higher reliability.

But the naturalistic fallacy is a fallacy for a reason. First, it is not at all clear whether it is useful or even possible to definitively distinguish between what is natural and unnatural. Humans are one branch on the tree of life, just like every other species. Our bodies are made of materials from the earth and powered by sunlight, just like every other species. We rearrange and recombine materials from our environment for our own self-interest, just like every other species. This manipulation results in things like sunblock, cement dams, and modern skyscrapers. But it is difficult to argue that a skyscraper is fundamentally different from a bird's nest or a spider’s web unless you arbitrarily define activities conducted by one particular species—humans—as being somehow unnatural.

Even if a clean line could be drawn between unnatural humans and the natural world, in what sense should nature deserve the reputation of being inherently good? Cancer, viral infections, heart disease, volcanic eruptions, tsunamis, and earthquakes all undoubtedly cause great suffering to conscious creatures who take no solace in the fact that these things are natural. 

Even the parts of nature that are presented as being “in harmony”, like a basic ecosystem food web are not so obviously good when interrogated. The food web gives the impression that every creature plays its part in maintaining a benevolent and perpetual balance. This creates the idea that each organism's role is to sacrifice for the greater good of the ecosystem. 

However, the true underlying dynamic of an ecosystem is an unrelenting competition between individuals of the same species and between species for finite resources. The existence of carnivores is often taken for granted, but it’s worth reflecting on what carnivores are actually doing: Competition for resources within the ecosystem is so intense that species evolved to kill and consume each other to forcibly take their resources and incorporate them into their own bodies. The prey do not consent to this fate; they live in a constant state of resisting being eaten throughout their lives until, one day, they fail and are eaten or they die from some other “natural” cause. Far from considering themselves to be noble components of the ecosystem, they would desperately want out of this system if given the choice.

This is not to say that nature should be disdained, but rather to make the point that nature itself is amoral and does not deserve a default status as being synonymous with “good.” Wildfires are natural in the sense that they have been present since soon after the appearance of land plants about 400 million years ago. Some species have been selected to take advantage of them. But none of this means that wildfires are desirable to humans today.

Ultimately we decide what we consider to be good through innate emotion and higher reason. We seek to persuade each other of our own versions of what is good through stories and logic, but people fundamentally disagree on values and priorities, so consensus is impossible. 

Actively manage the land to optimize desired outcomes, not to optimize naturalness.

Even if we will never reach a universal consensus on the optimal balance of tradeoffs in land management decisions, some principles will have wider endorsement than others. As people, it is not surprising that we would want to prioritize human interests that enjoy broad support, including water quality for municipal use, habitat maintenance and pest management (inevitably favoring some species that we prefer over others), aesthetics for recreation and tourism, timber harvesting, and wildfire mitigation.

Previous forest management regimes sought to maximize human benefits by excluding fire completely, but this has backfired. In 1911, the federal government adopted an official policy that put an end to the controlled burns that had been practiced by many Indigenous peoples, and in 1935, the “10:00 am” policy was put in place that sought to make sure all fires were extinguished by mid-morning the day after they were detected. Initially, this worked extremely well, as the amount of acreage burned fell from close to 50 million acres per year in the US in the early part of the century to the aforementioned 2 million acres per year in the late part of the century. But there were unintended consequences. 

The lack of fires meant that locations that would previously burn perhaps once per decade have now gone over a century without fire, leading to a buildup of small trees and brush. For example, in the 1800s, there were about 50 trees per acre in U.S. West Ponderosa pine forests, but now that number is approximately 200 trees per acre. The buildup of “hazardous fuels” for fires is like the stretching of a rubber band— potential energy has built up, ready to be released all at once the moment a fire inevitably occurs. Compounding the problem is that more vegetation means increased competition for finite resources (like water and sunlight) that make ecosystems more susceptible to disease and sensitive to drought enhanced by a warming climate. It also means that fires are more difficult to fight because it’s more difficult for firefighters to access them. And that fires burn more intensely and produce more smoke harming human health, and ash harming municipal water supplies. Finally, this buildup increases the likelihood that fires will climb into the canopies of forests and are thus more likely to kill mature and old-growth trees.

Mechanical thinning combined with prescribed burns for wildfire mitigation has become a priority because, for the most part, wildfires are bad for human interests. Not only is there an unprecedented number of people and structures at risk of being directly affected by wildfires (structures adjacent to the wilderness have increased in the US by 60% from just the 1990s to 2020), but we are continuously learning more about the negative impacts of wildfire smoke on human health and mortality and other indirect effects that lead to enormous economic burdens of hundreds of billions of dollars per year.

In this context, we must be extremely careful about reverting to the more “natural” fire regime that existed prior to the era of fire exclusion. Fire activity in the late 20th century likely marked a minimum compared to at least the past 3,000 years. As one study on prehistoric fires in California put it:

The idea that US wildfire area of approximately two million ha annually is extreme is certainly a 20th or 21st century perspective. Skies were likely smoky much of the summer and fall in California during the prehistoric period.

But going forward, the situation is unique. We have an excess of hazardous fuels on the landscape, the climate is becoming more conducive to catastrophic wildfires, and roughly two orders of magnitude more people live in the U.S. West than there were prior to European colonization. 

We can and should fight fire with prescribed fire (i.e., controlled burns) because it reduces subsequent wildfire intensity and it reduces the amount of smoke released compared to an uncontrolled wildfire. Not only that, but prescribed burns are done during conditions when prevailing winds will minimize human exposure to smoke. For this to work optimally, prescribed fire must often be paired with prior mechanical thinning of the landscape to physically remove brush and younger small-diameter trees so that the fire safely stays low and does not climb into the canopies of forests. 

Conducting these hazardous fuel treatments at scale will reduce wildfire danger, and our motivation for embarking on this project is the enormously detrimental impact of wildfires on people, our health, our property, and our economy. Our motivation is not based on adhering to a misguided notion that the most natural state of things is inherently the best state.

Subscribe now

By Patrick Brown

Climate change adaptation continues to receive increasing attention both in international and domestic contexts. In these discussions, we are continuously told that humanity was more-or-less sufficiently adapted to the historical climate, but now, the rapid and unprecedented rate of climate change is the primary motivation to invest in increased societal resilience to weather and climate hazards. We are also told that climate change adaptation is not going well, since the climate is changing much faster than we are adapting to it. 

Both of these points are seriously flawed. Humanity has never been close to sufficiently adapted to any climate. Our historical climate, far from being benign and nurturing, was indifferent and often hostile to our well-being, and this is why history is rife with examples of devastating climate impacts on society. 

At the same time, humanity has had tremendous success strengthening our defenses against the climate's inherent hostility faster than negative impacts from climate change have materialized. This climate adaptation (as opposed to climate change adaptation) is not a new phenomenon related to United Nations reports or initiatives of local governments. Rather, it is a continuation of humanity's inexorable impetus to reduce our vulnerability to our environment, driven largely by economic development and technological progress. 

Climate adaptation has a proven track record

The empirical evidence tells a very positive story on humanity’s increasing climate adaptation over time: almost all climate-sensitive aspects of society are trending in generally positive directions. We have observed higher crop yields, translating to more calories available per person, and a reduction in death rates from malnutrition and famines. Access to safe drinking water has increased, while the prevalence of climate-sensitive diseases like malaria has decreased. Furthermore, mortality rates due to extreme temperatures—both cold and heat—have declined, as have deaths from natural disasters.

Economic indicators reinforce this story. General economic measures like GDP per capita and the fraction of people in extreme poverty show major improvement over time as temperatures have risen, and even the economic impact of disasters, relative to GDP, has remained stable or decreasing.

To understand why these indicators are improving over time, we can look at the variation across space today. Notably, the differences between low and high-income countries are stark. Death rates from natural disasters like floods, droughts, and storms are 15 times higher for low-income countries than they are for high-income countries. Moreover, disaster damages as a percentage of exposed GDP are much lower in wealthier nations. Thus, more than anything else, general economic development is the critical factor in explaining the historical success of climate adaptation.

Subscribe now

Why does adaptation have a bad reputation?

If the empirical evidence so strongly supports our increased resilience, why does adaptation have such a poor reputation as being ineffectual, a “lazy cop-out” as Al Gore put it, or even immoral? 

Culturally, there is a great deal of social capital tied to expressing concern about climate change and acknowledging the successes of resilience is associated with a lack of concern. Further impacting nature is seen as inherently morally wrong and inevitably self-destructive under traditional environmentalist frameworks. Thus, there is significant  social and moral motivation to view adaptation as potentially ineffective.

A more technical reason adaptation has a bad reputation comes from academic studies of its effectiveness, which underpin official reports and subsequent news headlines on the subject. These studies often find adaptation to be relatively ineffectual due to a narrow definition that focuses on climate change adaptation and excludes general climate adaptation. Specifically, adaptation tends to be defined to include only technologies or activities that are more beneficial in a changed climate than they are in an unchanged one. 

David B. Lobell’s “Climate Change Adaptation in Crop Production” argues that if technology increases crop yields equally in the current and future climates, it cannot be considered an adaptation to climate change. Instead, only the additional benefits in the future climate (not the overall benefit of the technology) should be counted towards the technology's climate adaptation benefits. 

This is a valid technical point to make in an academic paper or a textbook, but it has limited practical use. As a society, should we really care to distinguish between technologies and behaviors that are “true” climate change adaptations compared to those that generally fortify society against all climates? I don't think so. 

However, this technical distinction has escaped the confines of scholarly academic journals and underlies general public sentiments about the ineffectiveness of climate change adaptation. For example, this narrow definition of adaptation supports the IPCC's assertion that adaptation in crop production will be insufficient to offset the negative effects of climate change (despite the mentioned general increases in crop yields over time as it has gotten warmer). It is also used through a chain of citations in the recent federal government’s increase in the social cost of carbon by 300%, which is our formal all-encompassing number for the impact of climate change. 

This definitional obscurantism elides the true effectiveness of climate adaptation, but we could still be much more adapted to our climate's perils.

Most of the motivation to accelerate climate adaptation is independent of climate change.

Societies around the world are better climate-adapted than ever, but you’d only need to open a news site on any given day to see that we have a long way to go. Consider that in the current climate, approximately 20 million people globally (one in every 400 people on Earth) are forced to move due to weather-related disasters each year, and domestically, we see over 20 “Billion dollar disasters” per year.

These and other climate impacts are captured by a body of literature demonstrating substantial economic impacts from weather variability in the historical record. If we were truly adapted to our climate, the economy would be insulated from the environment, and we would have no forced displacement, no billion-dollar disasters, and no observable relationships between weather and economic outcomes.

Activists would have you believe that all of this sensitivity to the climate is new and a result of climate change, but the data and our physical understanding of the climate system tell a different story. Natural, unforced variability significantly overshadows the climate change signal on planning timescales of about 30 years and local spatial scales. Climate is what you expect, and weather is what you get, but the weather that we get on these spatiotemporal scales is much more determined by the random roll of the atmospheric dice than by how much greenhouse gasses we emit or how much those gasses warm the global climate. Examining various weather and climate hazards individually makes this point more explicit. 

• There is no coherent trend in global floods (quantified with annual maximum streamflow), and projected changes are relatively small. 

• There is no coherent trend in global droughts (quantified with consecutive dry days), and projected changes are relatively small. 

• There is no trend in global hurricane activity (quantified with Accumulated Cyclone Energy), and projected changes are small. 

• There is no documented trend in US tornado and hail activity, and projected changes are relatively small. 

• Global wildfire activity shows a downward trend, and future projections are highly uncertain and depend very much on human land use and land management. 

Do the above facts tell us that we are sufficiently adapted to these hazards and, thus, don't need to concern ourselves with further reducing our vulnerability to them? Of course not! All of these hazards are extremely detrimental to human life and property, and we should ensure that we maintain or accelerate fortification against them.

Moreover, some hazard trends are moving unequivocally in the benign direction, such as the historical and projected decrease in extreme cold. But again, does this mean we need not worry about extreme cold? Cold-related deaths far outnumber heat-related deaths, and events like the 2021 Texas power crisis demonstrate far-reaching societal vulnerabilities. The signal from climate change is simply not large enough to allow society to relax our defenses against extreme cold.

The main exceptions—where explicit consideration of adapting to a new warmer climate is warranted—are coastal flooding (sea level rise) and extreme heat. In these situations, the signal-to-noise ratio is much larger, and thus, fortification against these hazards will be aided by explicitly anticipating their intensification in the future.

How to maintain and accelerate climate adaptation

So, what is the best way to maintain and accelerate global climate adaptation? There are many strategies specific to each weather or climate hazard. 

For extreme temperatures, these adaptations include: widespread well-insulated buildings with heating and air conditioning, a reliable energy grid able to handle increased energy demand during extreme temperatures, energy prices that are low so that heating and air conditioning are not prohibitively expensive, and automation of outdoor labor activities (e.g., mechanization of agriculture) so a larger fraction of the population works inside. 

For floods, these include: flood control systems (dikes, dams, and levees), capable stormwater drainage systems, zoning regulations to reduce building in floodplains, and properly priced insurance so that risk is accurately reflected.

For droughts, these include: building or expanding reservoirs, dams, or other water storage systems, researching and developing drought and heat-resistant crops, using water-efficient irrigation (precision agriculture), and investing in desalination technology. 

For major storms like hurricanes, these include: skillful forecasts and dissemination of information on preparation and evacuation, well-paved roads, a population with easy access to transportation so that evacuation is practical, and implementing building construction codes that ensure a minimum level of resiliency. 

For wildfires, these include: disseminating weather predictions and warnings that discourage activities that ignite fires, large-scale hazardous fuel reduction treatments, constructing firebreaks near communities, promoting fire-safe properties, implementing building construction codes that reduce the flammability of structures, well-funded and equipped firefighting systems with sufficient staff and resources like helicopters and bulldozers, and properly priced insurance that accurately reflects risk.

For all hazards, accurate weather forecasts and the efficient dissemination of warnings is paramount and incredibly valuable. But again, these were not developed explicitly to adapt to climate change but rather to increase our general resilience to our hostile climate. 

So, do we need a top-down global program to dictate the above measures? One way to think about this is to consider historical differential outcomes, like those related to food security. For example, a recent study estimates that the effect of historical climate change on food insecurity is much smaller than the difference in food insecurity across regions. Specifically, they showed that climate change has increased food insecurity by about 3% in all the regions they investigated but that the difference between Africa (50%) and Europe (13%) was 37%. There is also the finding mentioned above that death rates from natural disasters are 15 times higher for low-income countries than they are for high-income countries.

Now ask, is the relatively low amount of food insecurity and natural disaster death rates in higher-income regions due to some grand central program for climate change adaptation initiated in these countries many decades ago, or is success simply a reflection of overall societal wealth? It is primarily the latter. 

Thus, reducing energy poverty and advocating for growth are key to accelerating general climate adaptation on a global scale. More specifically, this means embracing public investments in the foundational aspects of economies: energy, transport, communication infrastructure, health, and education, and encouraging the private sector with business-friendly policies that attract foreign direct investment and facilitate entrepreneurship. Private property rights and fair market competition (i.e., capitalism) promote resource efficiency and innovation. Wealth liberates human mental and physical energy, enabling billions of microdecisions to respond more effectively to local environmental and market signals, thereby accelerating climate adaptation.

A wealthy, technologically advanced society is one in which you are informed via a variety of media sources of an approaching hurricane days in advance. You can then use your SUV to attain and install hurricane shutters (or have Amazon deliver them) on your already sturdily built, up-to-advanced-code home, and can subsequently use that SUV to evacuate to an affordable hotel on safer ground. After the storm, disaster response teams have cleared the roads, and you can return safely to the area with electricity restored and air conditioning running in a matter of days. 

Even projects that are almost certainly in the dominant purview of governments, like the building of major sea wall systems, require a wealthy tax base to fund those projects. For example, high-income countries like the US have the tax base to include climate adaptation measures in laws like the recent 2021 Bipartisan Infrastructure Law and the 2023 Inflation Reduction Act, which assigned $7 billion to FEMA to help communities proactively reduce their vulnerability to floods, hurricanes, drought, wildfires, extreme heat; $7 billion to The Army Corps for projects related to coastal storm risk management, hurricane and storm damage reduction, inland flood risk management, and aquatic ecosystem restoration; and $12.3 billion to the Department of the Interior for drought resilience. 

Overall, the climate being indifferent to our well-being and often hostile to it is the natural state of affairs, and this is why the past and the present are rife with examples of devastating climate impacts on society. What is unnatural—and good—is our relatively recent and continuing fortification against our often hostile climate. 

Embracing established paths of success for general climate adaptation is the headline, and adapting to climate change is the sideshow. Climate-sensitive societal outcomes are predominantly determined by societal wealth, and that wealth is still tied to greenhouse gas emissions. Thus, in order to accelerate climate adaptation, general economic development should be prioritized overrestrictive energy policies that insist that a carbon budget is near-to-be or has already been breached.

If you hear that climate change is causing something to “increase” or “decrease,” what does that mean to you? The words increasing and decreasing are relative terms that have an unspoken baseline, so an important question is increasing or decreasing relative to what? Most people, most of the time, interpret these words to mean increasing or decreasing relative to the past.

This is the straightforward and correct interpretation of the phrase “increasing greenhouse gas concentrations from fossil fuel burning is increasing global temperatures.” We can measure that global temperatures are increasing relative to the past, and physics-based climate models tell us that that increase is being driven entirely (technically, more than entirely because of offsetting cooling by aerosols) by increasing greenhouse gas concentrations.

Subscribe now

The above represents climate science's traditional “detection and attribution” framework. Here, the word detection refers to the idea that a change relative to the past is detected in the data. That is, some longer-term changes have risen above the noise of variability caused by other factors. After that change has been detected, we can move on to attributing the change to some underlying cause, like increasing greenhouse gas concentrations.

However, for many phenomena that we know are affected by increasing greenhouse gas concentrations (or “affected by climate change” in common parlance), we also know that the effect is small relative to natural variability (i.e., the signal-to-noise ratio is small). In such cases, like for meteorological drought at most particular locations, we don't expect any clear trend to be detectable from the amount of warming that has occurred historically, but it may still emerge in the future.

We can still estimate the influence of increasing greenhouse gas concentrations on some phenomena with a low signal-to-noise ratio using climate models to simulate the counterfactual scenario without increasing greenhouse gas concentrations. But now we are moving from the realm of first detecting and then attributing a trend to the notion that we can skip detection of the trend and go straight to attribution.

Here is an example of this applied to the recent southwestern North American “megadrought,” as quantified by how much moisture is in the soil in the summer:

As an aside, the LA Times covered the above difference between the black and green lines with this headline:

There is nothing technically wrong with the headline, but I’ll leave it to readers to assess whether the impression left by the headline is consistent with the impression left by the data.

Regardless, using model-generated hypothetical counterfactuals, we can make claims about whether increasing greenhouse concentrations are “increasing” or “decreasing” that thing, but notice how the meanings of the words “increasing” or “decreasing” in this context now apply to a change relative to a hypothetical situation without increasing greenhouse gas concentrations instead of necessarily referring to a change relative to the past. Technically, the words increasing or decreasing can always be said to be relative to a counterfactual without climate change, and if the signal-to-noise ratio is large enough, then those words can also refer to changes relative to the past (as is the case for global temperature in the first figure).

However, things get a little stranger when it comes to discussing climate change's impacts on societal outcomes. In these cases, the outcomes are often trending in the exact opposite direction that climate change is thought to be pushing them. In other words, things are trending in a good direction, and climate change is calculated to be making things worse than they would be otherwise but, importantly, not worse relative to the past. In this case, the notion of ‘detection and attribution’ is turned on its head because any detected trend would be in the opposite direction of the attribution.

In these cases, using the words “increasing” or “decreasing” can be very misleading if the reader doesn't understand exactly what is being claimed. In fact, I would go so far as to call it climate misinformation.

Take, for example, the statement, “Climate change is decreasing crop yields.” This could mean that we are seeing crop yields decrease over time, and this trend is being driven by climate change (left side), or it could mean that crop yields are increasing over time, but they would have been increasing faster if the climate wasn't changing (right side).

So, which is it? For crop yields, the situation is much more like the right side than the left side. However, the IPCC repeatedly states in its reports that climate change is “decreasing” crop yields, misleadingly conveying the message that the situation is more like the left side.

Climate change impacts on agriculture have been calculated to make up the largest negative societal impact of climate change as quantified in the social cost of carbon, while heat impacts on mortality represent the second largest. So, how does the IPCC report on mortality associated with heat? In its technical summary (B.5.3), the IPCC reports that,

“Increasing temperatures and heatwaves have increased mortality and morbidity (very high confidence).”

Most readers will interpret this to mean that heat deaths are increasing relative to the past (they will infer a direction of trend). But only those who read the technical chapters will understand that only a direction of impact is being reported because there you will read that heat deaths are decreasing over time:

“Heat-attributable mortality fractions have declined over time in most countries owing to general improvements in health care systems, increasing prevalence of residential air conditioning, and behavioural changes. These factors, which determine the susceptibility of the population to heat, have predominated over the influence of temperature change.”

This type of conflation is ubiquitous in climate change impact reporting. In addition to crop yields and heat mortality, I have previously drawn attention to this issue in the context of global wildfire, emissions, hunger, and climate-sensitive diseases like malaria.

It is especially pervasive in studies and reports on the economic impact of climate change. In that subfield, it is conventional to neglect to show the background economic growth completely and only show the calculated change relative to the background growth. This takes the misleading language and turns it into a misleading visual, leaving the visual impression that climate change is a large driver of a negative trend when it is actually a relatively small drag on a positive trend.

This all gives a misleading impression of the magnitude of climate change's negative impact, which is a problem if the goal is accurate scientific communication. More importantly, however, it distorts how we assess the desirability of decarbonization because the framing leaves out the important fact that the background progress in said societal outcomes is often driven by fossil-fueled industrialization.

The importance of fossil fuels in driving the background progress in some of these outcomes is illustrated in the paper Anthropogenic climate change has slowed global agricultural productivity growth. To the authors' credit, they do not state that climate change is “decreasing” global agricultural productivity, and they show the calculated negative impact of climate change in the context of background progress:

Also, to the authors’ credit, they attribute the background progress in agricultural productivity to fossil-fueled industrialization:

“Our estimates should not be interpreted as the effect of a world without fossil fuels on global agricultural production. Agriculture has benefited tremendously from agricultural research and carbon-intensive inputs that would not have been as available without fossil fuels. The counterfactual in our study only removes the effect that fossil fuels and other anthropogenic influences have on the climate system.”

But this brings up a more central question at the heart of so much research on the negative impact of climate change: Is imagining a world without climate change all that useful? It almost goes without saying that if we could keep the benefits of fossil-fueled industrialization and jettison the negative side effects of climate change, we would do so. But what makes decision-making so thorny is that for most climate-sensitive societal outcomes (e.g.,  Crop yields, calories available per person, death rates from malnutrition, the share of the population with access to safe drinking water, the rates of climate-influenced diseases like malaria and diarrheal disease, death rates from natural disasters, death rates from non-optimal temperatures (hot and cold), and the fraction of people in extreme poverty) the net effect of fossil-fueled industrialization and technological change has been good.

Thus, “without climate change” is not always the most relevant hypothetical counterfactual, and often “without fossil-fueled industrialization and technological progress” is more relevant. This framing gives a more honest and holistic picture of the state of the climate change problem, and it does not misleadingly paint the current systems as being less attractive than they actually are. When we assess the best course of action going forward, we must compare alternative systems and weigh the benefits of avoided climate change against the costs of transitioning to alternative energy and agricultural systems over time. This is the only way to be accurate and forthright on the tradeoffs we face.

In the meantime, the climate science and climate communication communities should strive to accurately inform the public and decision-makers on the state of climate change impacts. This means ending the use of the words “increasing” or “decreasing” when they do not refer to trends in time, and it means being forthright about the net effect of fossil-fueled industrialization on various societal outcomes.

Subscribe now

Welcome back to our two-part series on the role of climate change in economic losses from severe storms in the US. In Part 1, I noted that while we see large increases in economic losses from severe thunderstorms, there are not documented increases in tornadoes or hail, which are the hazards responsible for those losses. This discrepancy suggests that changes in assets exposed to those hazards are primarily responsible for rising losses. In this post, I’ll look at the magnitude of the influence of climate change on severe thunderstorms and compare that to the magnitude of changes in the amount of stuff (and its value) exposed to severe thunderstorms.

Fundamental Theory and Mathematical Models

Just because we haven't seen clear historical trends in tornados or hailstorms doesn't mean that warming is having no impact. It just means that so far, the effect is small enough that it can be buried in the noise of natural variability. So, in order to assess the effect of warming on the direction and magnitude of change in these hazards, researchers rely on fundamental theory and mathematical models.

Studying severe thunderstorms using climate models is particularly challenging because they are actually smaller than the resolution of the models and, thus, are not directly simulated. Because of this limitation, studies often rely on more indirect evidence. Specifically, research will often use fundamental theory on the ingredients for severe thunderstorms and look at how these ingredients have changed historically or in climate models driven by increases in greenhouse gasses. 

Fundamental theory tells us that severe thunderstorms require many ingredients, but at least two are particularly important. One is something called Convective Available Potential Energy, or CAPE, and the other is called wind shear. Convective Available Potential Energy is, just as it sounds, a measure of how much energy is available in the atmosphere for a storm to tap into. Wind shear is the change in wind speed or direction as you move up in the atmosphere, and it is an important ingredient for storms to develop rotation. You need a lot of both ingredients to get the “supercells” that are responsible for the longest-lived and most severe tornadoes and hailstorms. 

When we look at the historical changes in these ingredients, we don't see much change, which is consistent with the lack of change in reporting-bias-adjusted tornado and hail reports. 

However, models consistently predict that Convective Available Potential Energy will increase with warming. Wind shear is more uncertain, but the general expectation is that it will stay the same or potentially decrease. 

Some studies have run climate change experiments using mathematical models with high enough resolution to explicitly simulate changes in supercells. The figure below shows the simulated difference in supercell counts per year (represented by a measure called UH) between the historical time period and 2100 (under around 3°C of global warming).

They show decreases over the Great Plains and upper Midwest and increases in the South. Overall, they simulate a 7% mean increase in supercell counts and a 5% mean increase in supercell hours in 2100, but variability is very large, as seen by the large confidence intervals (which represent the range across different years in the simulation).

Given the relatively small impact of long-term warming relative to natural variability (small change in the mean relative to the width of the confidence intervals), it is not surprising that we haven't seen the signal emerge from the noise in historical trends so far. Nevertheless, the balance of evidence seems to suggest that a warming climate will be more conducive to severe thunderstorms.

This wouldn't necessarily translate into an increase in hail, however, because hail melts in warmer temperatures. Essentially, the aforementioned increase in Convective Available Potential Energy should increase the updrafts that are necessary for hail formation, so we should expect to get more hail formed in clouds. However, that hail will fall through a warmer atmosphere and be more likely to melt into rain on the way down. So, which effect wins out? It is uncertain and depends on location and season. There are definitely some locations and seasons where the latter effect should win out. For example, one study shows that hail events mostly become nonexistent in future summers in Colorado. A variety of studies have indicated that drier and cooler regions in North America will experience the largest increases in hail threat, while warmer and more humid regions will experience a reduced threat.

Exposure and vulnerability

So, changes in the hazard of severe thunderstorms and their sub-hazards are small to date and, for most measures, are expected to be on the order of 10% or less this century. What, then, explains the very large changes in economic losses from these events?

In the long term, the expanding bullseye effect that I emphasized in the flood series is a big part of the story.

Stephen Strader provides great examples of this. Below is a recent risk map of severe weather from this spring, in which 775,000 housing units in Oklahoma and Kansas were exposed to a “high risk” of a tornado (right side). But, the exact same risk map would have only included 321,000 housing units (41%) had it been issued in 1970.

That’s a specific example, but it generalizes. Systematic studies that encompass larger regions of the country and longer time periods reveal the same phenomena. Below is the number of structures per square km in tornado hotspots, showing an increase from about 2 structures per square km in 1940 to about 10 structures per square km today—a fourfold increase! These increases in exposure to severe weather hazards are much larger than any change in severe weather hazards themselves and thus are the dominant driver of trends in economic losses. 

The insurance industry confirms as much. In its recent report on the matter, for example, SwissRe says that the main drivers of increased natural catastrophe losses are increases in exposure due to economic and population growth and increases in the cost of construction (independent of general inflation). They say, "So far, the impact of changing climates has been small." 

An interesting finding from the SwissRe report is that hail is responsible for most (50-80%) of the insured losses from severe convective storms, and a key driver of increased losses from hail is increased vulnerability (in terms of the potential to incur monetary damage) in the form of increased solar installations. Again, it is not at all clear that the hail hazard has or is expected to become worse with warming. So, what is in the way of hail and how much it costs to replace it, makes a big difference. 

Similarly, the insurance company Aon had a recent report that breaks down changes in exposure into gross domestic product (GDP), fixed reproducible tangible wealth, property cost inflation, and population distribution in high-hazard zones.

These four components grew at rates of 2.3, 2.1, 2.8, and 1.1% per year, which would absolutely dwarf any type of climate change signal in tornadoes or hail. Overall, they found that changes in value exposed account for 93% of the overall change in economic losses from severe convective storms (8.3% per year compared to the loss trend of 8.9% per year). Aon found that the remaining 7% in loss is unexplained and could be partially attributed to climate change. 

It is worth noting that in terms of monetary losses, a great deal of the effect is due to the value exposed and the cost of reconstruction rather than the population exposed. This is part of the reason why fatalities from tornadoes, for example, have seen a long-term decline.

Essentially, there are many more people in harm’s way than ever before, but we are better protected, warned earlier, and better prepared for severe weather than we ever have been, which has been driving down deaths. 

To conclude, we see large increases in economic losses from severe storms, but economic loss data cannot be used to infer changes in severe storms themselves. This is because the economic loss data is heavily influenced by changes in the value exposed and vulnerability of assets. When examining historical trends in tornadoes and hail over the US, we observe little change (as detailed in Part 1). However, looking to the future, we expect environments conducive to severe weather to become more common, but these changes are relatively modest, on the order of 10%.

In contrast, the changes in exposure—often referred to as the "expanding bullseye effect"—are an order of magnitude higher or more. This significant increase in exposure underscores that the rising economic losses are more a function of where and how we build rather than the frequency or intensity of the storms themselves. Therefore, the tendency of certain media outlets to attribute these increased losses primarily to climate change can be misleading.

The takeaway is that while climate change does play a role, emissions reductions will be an incredibly inefficient lever in the fight against rising economic losses from severe storms. Instead, we should focus on managing exposure and reducing vulnerability in high-risk areas. This involves better planning, stronger building codes, and more effective disaster preparedness.

This is the first part of a two-part series. Part two will be posted June 19, 2024.

Springtime means severe thunderstorm season—tornadoes, derechos, and hail, among other hazards—for much of the eastern two-thirds of the US.

Damage caused by severe thunderstorms is immense and represents the second most frequent type of billion-dollar disaster in NOAA’s disaster database for the US. Severe thunderstorms are also the weather category responsible for the second most insured losses globally, according to SwissRe (in both datasets, tropical cyclones or hurricanes are responsible for the most damage). Both US billion-dollar disasters and global insured disaster losses are increasing, and a large fraction of the overall increase seems to be driven by increases in losses from severe thunderstorms. 

For many, especially in the media, the natural inclination is to see these rising trends in disaster losses from severe thunderstorms and reflexively attribute them to climate change. 

Take, for example, the New York Times’ The Daily podcast (consistently a top 5 podcast) from May 15th, 2024, titled “The Possible Collapse of the U.S. Home Insurance System.” The episode is about how these increases in insured losses are straining the books of insurers. The reporter, Christopher Flavelle, had this to say:

“One of the really striking things about this data was it showed the contagion [climate change-driven disaster losses] had spread to places that I wouldn't have thought of as especially prone to climate shocks. For example, a lot of the Midwest, a lot of the Southeast.”

“Formerly unimportant weather events like hailstorms or windstorms didn't used to be the kind of thing that would scare insurance companies… But those are becoming so frequent and so much more intense that they can cause existential threats for insurance companies.”

“This market is starting to buckle under the cost of climate change, and this is all happening really fast.”

However, as I emphasized in the previous series on floods, disaster impacts result from the combination of natural hazards (e.g., a physical event like hail or a tornado), exposure (the presence of people, resources, or infrastructure in harm’s way), and vulnerability (the propensity of people, resources, or infrastructure to be adversely affected).

Economic loss trends can be driven by changes in exposure and vulnerability even if the hazard is not changing. That’s why Roger Pielke often says, “Climate data should be the basis for claims of detection and attribution of changes in climate variables, not economic loss data.”

So, what does climate science say about historical and expected changes in severe thunderstorms and their associated hazards of tornadoes and hail?

As I have mentioned in previous posts, it is clarifying to divide scientific evidence into 1) Historical trends, 2) Fundamental theory, and 3) Mathematical modeling.

In this post, I’ll examine historical trends and then discuss fundamental theory and mathematical modeling in Part 2. 

Subscribe now

Historical trends

Since thunderstorms are on a continuum of severity, it is not necessarily clear how to track changes in severe thunderstorms over time, and there are many different variables you could look at. An obvious place to start would be to just look at the annual number of tornadoes per year and how they are changing, but it turns out this is not as easy as it sounds. 

You might think we have good data on phenomena as conspicuous as tornadoes, but unfortunately, not all tornadoes are reported, and there is a strong bias towards increased reporting in recent years due to increases in population and infrastructure that can be damaged by tornadoes (since damage itself is often used to report a tornado especially if it occurred at night). Thus, in order to study long-term trends in tornadoes, researchers need to account for these reporting biases. When studies have done this, they get results like those below.

The red line in the top figure shows no long-term trend in the number of US tornadoes, and the study concludes that “long-term climate change has not substantially affected the domain-wide tornado frequency during the 1975–2018 analysis period.” Interestingly, the study found a statistically significant decreasing trend in the strongest tornadoes (bottom right panel).

Other studies have confirmed no long-term trends in tornado occurrence but have noted changes in other statistics, such as a decrease in the number of days per year with tornadoes combined with an increase in the number of tornadoes on those days. Other studies have shown shifts in geography and seasonality of where tornadoes occur, but it is unclear how much of this might be due to natural decadal variability as opposed to being driven by long-term warming.

Hail is another major component of severe convective storms and is responsible for a great deal of economic losses from both property and crop damage. These losses are roughly $10 billion per year in the US, and some single events reach multiple billions, like a 2012 hailstorm in Phoenix, Arizona, which caused $4 billion in damage.

Hail data suffers from the same issues as tornado data in that it is heavily influenced by reporting biases towards areas with higher population density. However, since at least around 2004, we have radar estimates of hail that don't suffer from these reporting bias issues (but come with their own laundry list of uncertainties and caveats). Nevertheless, this data paints a complicated picture of change that depends on location and season. For the spring, we see increases in severe hail that cause the most damage (greater than 2 inches in diameter) over Texas and Oklahoma but decreases over most of the rest of the country, and for the summer, we see increases in severe hail over the northern high plains and Minnesota but decreases over much of the rest of the country.

The authors note that natural variability (associated with phenomena like El Nino) is responsible for the bulk of this pattern and that it is unclear what, if any, effect long-term warming has had thus far. A recent review titled “The effects of climate change on hailstorms” concluded that “Overall, no clear overarching national climatological hail trend has been found for the USA”

We see that there are large increases in the economic losses associated with severe thunderstorms, but there are not well-documented increases in the physical hazards responsible for those losses. When reporting biases are accounted for, we see no increase in total tornado counts and perhaps a decrease in the occurrence of the strongest tornadoes. For hail, there is documented natural variability but no established long-term trend associated with climate change. 

This discrepancy suggests that changes in exposure—such as population growth, and increased value of properties—may play a major role in these rising losses. By simultaneously considering anticipated changes in severe thunderstorm hazards and exposure to those hazards, we gain more direct insight into the causes of and thus potential solutions for increasing damages from these weather events. In Part 2, we will delve deeper into these aspects.

Subscribe now

Note: This post is part of the collaborative series between The Liberal Patriot and The Breakthrough Institute called “The Climate Report” looking at the science and reporting behind extreme weather events and other climate related matters. This is the second part of a series on flooding and climate. Part one can be read here.

In part one of this series, I looked at how the flood natural hazard—the physical phenomenon of flooding itself—has changed and is expected to change as the climate warms. We saw that matters are complicated and that a global increase in floods is not certain. Nevertheless, the IPCC projects that as warming continues, more locations will see increases in the flood hazard than decreases in it.

But as Dr. Diana Liverman has put it, we need to think about risk not as the probability of the hazard but as the probability of harm. For a flood to become a human disaster, things we care about need to be exposed to the flood and be vulnerable to the flood.

Subscribe now

Flood Exposure

Since the mid-1980s, the world population has increased by over 60 percent. We would naturally expect more people to experience flooding because of that fact alone even if floods were not changing at all. This idea can be visualized with the expanding bullseye effect, which shows how the exact same flood (or any hazard) would affect many more people as an urban area grows and the bullseye expands.

In reality, the situation is actually worse than that. The global population has not only expanded but disproportionately expanded into the most flood-prone areas. In fact, the population in flood-prone areas has increased at nearly twice the rate as in less flood-prone areas since the 1980s. To get a good sense of this process, look at these satellite images of changes in Quang Nam, Vietnam, from 2002 to 2021.

The flood plains near the river that were previously unoccupied are now dense urban areas. This is not unique to Quang Nam: the global potential exposure of urban areas to floods increased by more than four timesbetween 1985 and 2018. The figure below breaks down flood exposure by continent, showing that the largest growth in exposure occurred in Asia.

This trend of disproportionate increases in potential exposure is expected to continue into the future. I say “potential exposure” because many of these statistics are calculated by first calculating a flood risk map and then looking at how the population has expanded into flood-prone areas (i.e., how much population is in a predefined 100-year floodplain). In other words, they are looking at potential exposure and not actual or realized exposure to floods.

Regardless, given that changes in the flood hazard itself are not large globally (see part one), fanfare about apparent increases in the reported number of flood disasters will be much more of a reflection of the fact that more people live in the way of floods than it is of climate change causing more or larger floods. But that raises an obvious question: “Have more people actually been exposed to real floods in recent years?” Since the era of good satellite data, it doesn’t appear so.

The plot below from this study shows the satellite-observed flooded area (the flood hazard, in blue) and the population exposed to the flood hazard (red) starting when sufficient satellite coverage begins in the early 2000s.

It appears that over this time period, there is no trend in people experiencing floods despite an increasing number of people living in floodplains. This is a somewhat remarkable finding given the media claims about flooding mentioned in part one:“It’s getting harder and harder to adapt to these changing conditions,” as one climate expert put it. “It’s just everywhere, all the time.”

Over such a short period of time, human exposure to actual floods may show no trend due to pure luck. As time goes on, more people living in flood-prone areas should result in more people being flooded. But there is also the issue of changes in flood vulnerability that affect the risk of flood impacts.

Flood Vulnerability

While our exposure to floods has increased in the long term, our vulnerability to them has decreased. A major component of this is the construction of explicit flood control systems, which include dikes, dams, levees, and urban stormwater drainage systems that allow settlements to deal with inundations of water when they occur.

The San Joaquin River flood management system in California provides a good example.

Extensive flooding is natural and common in California’s Central Valley, and as the region's populations and assets grew, flooding became increasingly intolerable over the 20th century. This led to a system of 12 dams and reservoirs that can substantially absorb excess water when heavy rain and snowmelt occur.

Systems like this are obviously not perfect and are continuously being evaluated and updated, but it is demonstrably the case that they reduce flooding. For example, “regulated” river basins in the United States with substantial reservoir storage have seen marked decreases in annual maximum stream flows over time. This means that when it comes to flooding, we’d rather live in a warmed climate with flood protection than live in a preindustrial climate with preindustrial flood protection.

These systems are common worldwide, and approximately 3 million dams currently regulate about half of the world's river systems. Thanks in part to systems like these, both global flood fatalities as a fraction of those exposed to floods and flood damages as a fraction of GDP exposed to floods have decreased since the 1980s.

This progression is associated with economic growth, and if you look across societies today, you’ll see that wealthier is better: Flood fatalities as a fraction of those exposed to floods and flood damages as a fraction of GDP exposed to floods are smaller for societies with higher GDPs. 

Worldwide, the lethality of river floods decreased 84 percent from 412 deaths per flood in the 1950s to 67 deaths per flood in the 2010s. High-income countries started from a smaller baseline of 86 deaths per event (floods AND storms in this case) in the 1990s but have still progressed to only 16 deaths per event in the decade from 2012-2022—a decrease of 81 percent.

Future Projections

Despite changes in flood exposure (more people being exposed to the flood hazard over time) and flood vulnerability (societies being less sensitive to floods over time) being so important, future projections of flooding often neglect these factors. This is in part due to the fact that it’s simply much more difficult to project changes in technological and socioeconomic factors than it is to project changes in the physical climate system. However, the neglect of these two dominant factors means that a lot of projections that focus exclusively on the physical hazard are of very limited practical use.

Below are the IPCC’s projections of river flooding and the fraction of the population exposed to flooding. Globally you see a 0.7 percent annual exposure to floods historically increase to 1.7 percent under 3°C of warming (bars).

However, IPCC reports calculated exposure for when the magnitude of a simulated flood “exceeds current flood protection,” effectively ignoring the changes in vulnerability that have been so important historically or assuming that no further changes occur. This is somewhat absurd when you consider that 3°C of warming is associated with several-fold increases in global GDP and ~50 percent to ~300 percent increases in total societal energy use. To take these projections seriously, you’d have to imagine that none of that extra wealth and energy use goes towards flood control.

Research that has attempted to take future changes in flood control systems into account paints a more optimistic picture and has suggested that despite increasing population and value in flood-prone areas, the widespread adoption of flood control infrastructure could result in decreases in the absolute number of fatalities and economic losses from floods, even under the worst case climate change scenarios.

Overemphasis of the flood hazard

For a flood to be a natural disaster, it requires not only the flood hazard itself but also, exposure (people living in the way of the flood), and vulnerability (inability to deal with the flood). Overall, there is very solid evidence that one component of the flood hazard—extreme precipitation—has been and will continue to increase in most locations as it warms. However, there are offsetting factors (such as reduced water in the soil and reduced snowpack as it warms) that make it absolutely plausible that the global inland flood hazard could remain roughly steady over time.

Even though there is not a strong coherent global increase in the inland flood hazard, media coverage tends to focus narrowly on the single part of the problem (extreme precipitation) that can be most directly connected to increased greenhouse gasses. This is counterproductive because it leaves the impression that the most effective way to deal with flood disasters is through emissions reductions when in reality, changes in exposure and vulnerability play a much larger role in driving changes in flood disasters, and we can influence exposure and vulnerability much more directly in the near term than we can the physical flood hazard. 

Subscribe now

Originally published by The Liberal Patriot as part of a content partnership with Patrick Brown and the Breakthrough Institute. Subscribe to their Substack here.

Flooding is among the most consequential natural hazards and globally accounts for tens of thousands of deaths per year and hundreds of billions of dollars of damage per year.

When rivers overrun their banks or flash floods occur from extreme rainfall, many media outlets will reflexively report on the flooding as though we are in a fundamentally new situation due to climate change.

Consider the following headlines:

• The Washington Post when major flooding hit the United States last summer: “Floods, fires and deadly heat are the alarm bells of a planet on the brink.”

• The New York Times headline reported major floods in Vermont last year, “Showed the Limits of America’s Efforts to Adapt to Climate Change.”

The Times article began: “This week’s flooding in Vermont, in which heavy rainfall caused destruction even miles from any river, is evidence of an especially dangerous climate threat: Catastrophic flooding can increasingly happen anywhere, with almost no warning.”

A few paragraphs later, we hear: “‘It’s getting harder and harder to adapt to these changing conditions,’ said Rachel Cleetus, policy director for the climate and energy program at the Union of Concerned Scientists. ‘It’s just everywhere, all the time.’”

But is it actually everywhere, all the time, and are we progressing into a fundamentally new world of flooding due to climate change?

When considering the risk of natural disasters like floods, the UN’s Intergovernmental Panel on Climate Change (IPCC) has adopted a useful framework for breaking down the risk of impacts.

Under this framework, the risk of impacts from a natural disaster can be thought of as resulting from the combination of natural hazards, exposure, and vulnerability.

• The natural hazard is the physical event like the flood that has the potential to cause loss of life, economic disruption, or environmental degradation.

• Exposure refers to the presence of people, resources, or infrastructure in harm’s way of the natural hazard.

• Finally, vulnerability is the propensity of people, resources, or infrastructure to be adversely affected when exposed to a hazard.

Essentially, a natural disaster can only occur if things we care about are exposed to a natural hazard and they are vulnerable to it.

This framework is useful for considering the underlying causes of flood disasters because it is clear that, on the many-decade timescale that climate change progresses, there will also be major changes in exposure (e.g., the number of people and what proportion live in floodplains) and vulnerability (e.g., weather prediction, behavior, flood control, storm sewer systems, and other infrastructure).

If we focus exclusively on climate change's influence on the natural hazard, we will miss a major part of the story.

In these two posts, I’ll examine all three components of floods. In Part One, I’ll take a look at changes in the physical flood hazard, and in Part Two, I’ll examine changes in human exposure and vulnerability to floods.

The Flood Hazard

The flood hazard is simply the inundation of normally dry land. Floods are often classified into coastal flooding and inland flooding, both river flooding and rainfall flooding (river flooding is often caused by rainfall but can also be caused by, e.g., rapid snowmelt).

Sea-level rise is an unequivocal result of global warming, and the five inches of global mean sea-level rise we have observed over the past 50 years unquestionably exacerbates global coastal flooding.

When it comes to inland flooding, however, the story is more complicated. The fundamental theory for why warming should increase inland flooding comes from the so-called Clausius Clapeyron relationship, which tells us that warmer air can “hold” more water—about seven percent more water vapor per °C. That means that, all else being equal, a warmer climate is one where there is more water vapor in the atmosphere and thus more water available to condense and rain out of the atmosphere when rain occurs. Studies have shown that, on average, globally, we do indeed see increases in extreme precipitation roughly in line with this seven percent per °C.

All else is not necessarily equal, though, and the degree to which warming affects other parts of rainstorm dynamics can also play a role. There’s also simply a lot of random natural variability in when and where extreme rainfall happens to strike, which obscures the underlying effect of warming.

These factors (and others) mean that the increase in extreme precipitation is not universal, and plenty of locations have seen decreases in extreme precipitation despite warming.

To better understand the influence of warming on extreme precipitation (and therefore how we expect it to change in the future) at the location-by-location level, we run climate models. There are dozens of climate models that differ slightly in their construction and the answers that they give us. The range of output from models serves as a proxy for our uncertainty on the effect of warming on extreme precipitation.

The IPCC shows climate model simulated changes in extreme precipitation and the range across models. The maps below not only show the median model change for 2°C of global warming in the middle panel but also the 5th percentile of the models in the left panel and the 95th percentile of the models in the right panel.

We see in the middle map that over almost all land locations, the median climate model indicates an increase in the magnitude of extreme precipitation. However, if we had very high confidence in warming’s influence on changes in extreme precipitation at any given location then that would be reflected in the models' universally agreeing on the sign of the change. However, we see that in the 5^th percentile map, there is a lot of brown, indicating that some models, at least, actually project a decrease in extreme precipitation over a lot of land locations as it gets warmer.

Let’s bracket that uncertainty for the moment and assume that the median model is correct and that we can expect warming to cause extreme precipitation to increase over almost all land locations. Does this translate directly into an increase in the flood hazard? No.

There are many other factors to consider when thinking about the link between extreme precipitation and flooding but two stand out as being particularly consequential.

One is the effect of the warmer atmosphere on the amount of moisture in the soil before an extreme rainfall event occurs. The same mechanisms that allow more rain to fall out of a warmer atmosphere will also cause more water to evaporate from the land surface prior to the rainfall event. This means that as the atmosphere warms, soil will often have more capacity to absorb the additional rainfall when it does occur.

The other major factor is that many river floods are not driven by extreme rainfall but rather by the rapid melting of snow that has accumulated over the winter in higher latitudes and elevations. As it warms, more precipitation falls as rain rather than snow, which reduces the average snowpack in most locations such that spring melting results in a smaller release of water.

Those two countervailing influences on floods are potentially large enough to offset the effect of more extreme rainfall. With that in mind, let’s look at the overall data on floods.

Floods are difficult to quantify and measure globally, especially before sufficient satellite coverage came online in the early 2000s. Most studies on floods therefore look at changes in extreme streamflow—the rate of water movement in rivers. We have high-quality, long-term data on streamflow in many locations, but it is still lacking in spatial extent.

Nevertheless, we see substantial diversity in the sign of trends where we have good data, and overall most observational studies show no increase in floods globally and, if anything, show decreases. The IPCC essentially says that it is a wash and concludes:

There is low confidence about peak flow trends over past decades on the global scale, but there are regions experiencing increases, including parts of Asia, Southern South America, north-east USA, north-western Europe, and the Amazon, and regions experiencing decreases, including parts of the Mediterranean, Australia, Africa, and south-western USA.

Research has also shown that trends depend greatly on the type of flood. As mentioned above, we have observed (and climate models project) that floods due exclusively to extreme rainfall should increase in magnitude while floods due to snow melting should decrease in magnitude.

This means that both observations and models indicate that globally, extreme rainfall is and will continue to increase, but flooding (as quantified from extreme streamflow) should stay relatively stable.

With no clear coherent change in floods globally, the IPCC cannot “attribute” any change to human-caused increases in greenhouse gas concentrations and thus concludes that:

In summary there is low confidence in the human influence on the changes in high river flows on the global scale. In general, there is low confidence in attributing changes in the probability or magnitude of flood events to human influence because of a limited number of studies, differences in the results of these studies and large modeling uncertainties.

However, the lack of coherent global change does not mean that we don’t expect change at the local level. We just expect decreases in some locations and increases in others.

Concerning projections, the IPCC concludes that:

In summary, global hydrological models project a larger fraction of land areas to be affected by an increase in river floods than by a decrease in river floods (medium confidence).

Though the map above indicates to me that it is pretty close.

I began this part by pointing out that changes in flood disasters will depend on not just the flood hazard but also changes in exposure and vulnerability. Here, we have seen that climate change should enhance extreme rainfall—an important influence on the flood hazard, but far from the only one. More extreme rainfall should, in turn, increase the flood hazard in some locations, but it may not be enough to increase the flood hazard globally. That means that changes in exposure and vulnerability have the potential to be the dominant drivers of changes in flood disaster risk both in the past and going forward. I’ll delve into those two aspects in the next part.