The Science That Puts a Name on the Storm: How Attribution Changed Climate Accountability

When a hurricane flattens a coastline or a heat dome kills hundreds of people in a city that never needed air conditioning, the question that follows is almost reflexive: did climate change cause this? For most of the twentieth century, scientists deflected that question. Weather, they would say carefully, is not climate. Individual events could not be pinned on a warming planet. That careful deflection is now obsolete.

In the early 2000s, a new field called extreme event attribution began to take shape, asking not whether climate change caused a specific disaster, but how much more likely or more intense that disaster became because of it. The difference sounds subtle. The implications are enormous.

Attribution science works by running climate models thousands of times, comparing a world with human-caused greenhouse gas emissions against a counterfactual world without them. The gap between those two simulations tells researchers something precise: a heat wave of this magnitude was, say, five times more likely in the world we actually live in than in the world we might have had. Carbon Brief has tracked and mapped hundreds of these studies, and the pattern that emerges is not random. The fingerprint of warming shows up most clearly and most consistently in heat-related events. Of the attribution studies conducted on extreme heat events, the overwhelming majority find that climate change made those events more likely or more severe. The signal is strong enough now that some researchers argue it is no longer scientifically defensible to analyze a major heat event without accounting for the human influence on its probability.

The picture is more complicated for other event types. Rainfall and flooding studies show a mixed record, partly because precipitation is governed by regional dynamics that global models still struggle to resolve cleanly. Tropical cyclones present their own complexity: climate change appears to be making the strongest storms more intense and increasing the rainfall they carry, even as the overall number of storms may not rise dramatically. Drought attribution depends heavily on whether you are measuring soil moisture, streamflow, or atmospheric demand for water, and each metric tells a somewhat different story.

What this means, practically, is that attribution science has matured unevenly. It is most powerful and most legally actionable in the domain of heat, and that unevenness matters for how the findings get used.

The second-order effect that most observers have been slow to appreciate is the legal and financial one. Attribution science was originally a tool for researchers. It is rapidly becoming a tool for lawyers and insurers. Climate litigation has accelerated sharply over the past decade, and plaintiffs in cases against fossil fuel companies and negligent governments are increasingly leaning on attribution studies to establish that specific harms were made measurably worse by specific emissions. The logic is no longer speculative. If a study can show that a flood event was made 40 percent more likely by climate change, and if emissions from a particular set of actors contributed a calculable share of that warming, the chain of liability becomes harder to dismiss in court.

Insurance markets are already responding, sometimes faster than regulators. Several major insurers have pulled back from high-risk coastal and wildfire-prone markets in the United States, not because they doubt that climate change is real, but precisely because attribution science has helped them price the risk more accurately, and the accurate price is one the market will not bear. When insurance retreats, property values follow, and when property values fall in climate-exposed regions, the tax bases of local governments shrink, reducing their capacity to fund the very adaptation infrastructure that might reduce future losses. That is a feedback loop with no clean exit.

The field is also beginning to grapple with a harder question: attribution of slow-onset events. Sea level rise, permafrost thaw, and the gradual bleaching of coral reefs do not arrive as discrete disasters with a date that triggers a news cycle. They accumulate. The methodological challenge of attributing chronic degradation rather than acute catastrophe is one the science has not fully solved, and until it does, the communities most damaged by those slow processes will have the hardest time making their case in any forum, scientific or legal.

What began as an academic exercise in probabilistic modeling has become one of the more consequential scientific developments of the century, not because it answers every question about weather and warming, but because it changed the terms of the conversation from possibility to probability, and from probability to accountability. The storms were always going to come. The question now being asked, with increasing precision, is who made them worse.