How “climate change” became the answer to every weather question

Somewhere in the last decade, “climate change” stopped being a hypothesis about the climate and started being a kind of all-purpose explanatory glue. It now gets applied to almost any weather event, almost any environmental observation, and almost any unusual season, often by people whose actual expertise is in the downstream consequence rather than the atmosphere. A vet treating a stranded albatross. A farmer noticing dry pasture. A council engineer signing off on a stormwater design. Each is now expected, almost as a professional reflex, to gesture upward at climate change. And so they do.

The problem is not that climate change is unreal. It is real, it is measurable, and it is consequential. The problem is that the reflex has overtaken the analysis. A single weather event in 2026 is now framed as a climate event in a way that a structurally identical event in 1986 was not, even though the underlying physics has changed only slightly in those forty years. The climate signal is real but small. The framing signal has grown to swamp it.

It is worth being precise about what is and isn’t known, because the looseness of public discourse is starting to corrode trust in the parts of the science that are actually solid.

What the data actually says about New Zealand storms

The strongest, most defensible climate attribution claim about New Zealand weather concerns rainfall intensity. After Cyclone Gabrielle in 2023, the World Weather Attribution team’s rapid analysis found that very heavy rain of the kind associated with the cyclone had become roughly four times more common in the affected region, with extreme downpours dropping about 30 per cent more rain than they would have without anthropogenic warming. That is a quantitative, peer-reviewable, mechanistically grounded claim. It rests on the Clausius–Clapeyron relation (air holds about 7 per cent more water vapour for every degree of warming) and it shows up in the observational record.

The claim about storm frequency in New Zealand is weaker. NIWA, now Earth Sciences New Zealand, and the IPCC’s regional projections are confident that ex-tropical cyclones reaching New Zealand will become more intense, that storm surge combined with rising sea level will produce more frequent coastal flooding, and that extreme rainfall events will become more common. They are markedly less confident about whether the total number of storms in a given year will increase. The mechanisms governing storm frequency in the South Pacific (ENSO, the Southern Annular Mode, the Interdecadal Pacific Oscillation) produce large natural variability that climate change has to be detected against, and the detection is not yet clean.

The number of weather-related states of emergency in New Zealand has clearly risen, from four in 2015 to eight in 2025, and already past that figure in the first months of 2026. But part of that is genuine climate signal, part is policy and reporting change (states of emergency are declared more readily now than they were ten years ago), and part is exposure. There are simply more people, more buildings, and more infrastructure in places that flood. Disentangling those three drivers is the actual scientific work. Collapsing them into “climate change” is the reflex.

What “climate change” has come to mean in casual usage

In casual NZ media usage, “climate change” now functions as shorthand for any of the following, often interchangeably.

• A specific extreme weather event being slightly more intense than it would have been without warming.
• A general impression that “the weather is weirder than it used to be.”
• A policy failure (insufficient adaptation, poor land-use planning, deferred infrastructure maintenance).
• A moral framing that signals the speaker takes the issue seriously.
• An ecological observation about species, seasons, or behaviours that may have nothing to do with the climate at all and a great deal to do with fishing pressure, predation, or land use.

These are not the same claim. The first is testable. The third is a policy argument dressed up as a physical one. The fifth is often a category error. But in everyday speech they all come out wearing the same two-word badge, and the badge does the work of an argument.

The flattening of expertise

Part of how the reflex took hold is that the public sphere stopped distinguishing between kinds of expertise. A climatologist with forty years of peer-reviewed publications, a vet quoted on storm frequency, a policy advocate with a bachelor’s degree in environmental studies, a celebrity who flew in for a documentary, and a council communications officer all appear in the same kind of news sentence, with the same kind of authority signal attached. The reader has no way to weight the inputs differently. They all become “experts say.”

This is a particular problem for a topic like climate, where the actual science is technically demanding and the actual scientists are a small specialist community. There are perhaps a few dozen people in New Zealand whose day job is climate physics or atmospheric modelling at a research level. There are thousands of people who can speak about climate in a public-facing capacity, including environmental managers, sustainability consultants, NGO communicators, local-body planners, well-briefed journalists, vets, and biologists in adjacent fields. The first group does the science. The second group does the speaking. The two groups have different skill sets, and conflating them flatters the second group while quietly burying the first.

The flattening also runs in the other direction. Working scientists who disagree with the consensus framing of their own field, not with the underlying physics, but with how confident a given inference is, or how a particular model output should be interpreted, have learned that public dissent has a career cost. New Zealand’s own history with this is not pretty. A senior Auckland climatologist who spent decades publishing in the field and co-edited a peer-reviewed climate journal found himself the target of a coordinated campaign to remove him from that editorial role after he published a paper whose conclusions departed from the consensus position. The objection was not, in the end, methodological. It was about which conclusions were acceptable. The chilling effect of episodes like that is hard to measure, but every younger researcher watching learned the lesson, and what gets said in public has been narrower since.

The result is a discourse that looks like consensus but is partly the shape of selection pressure. The credentialled voices who agree get amplified. The credentialled voices who disagree get attacked or drift into early retirement. The non-credentialled voices who agree get treated as if they were credentialled. And the public, watching, gets the impression that “the science” speaks with a single voice on questions where the actual peer-reviewed literature is considerably more textured.

This matters because the strong claims in climate science are strong on their merits. They do not need the rhetorical scaffolding of a flattened expertise hierarchy to support them. The Clausius–Clapeyron relation does not become more true because a vet endorses it. The Meltwater Pulse 1A record does not become less true because a sceptical climatologist points it out. Treating the science as a coalition to be defended rather than a body of work to be argued with is a category error, and it has made the public worse at distinguishing the parts of the field that are nailed down from the parts that are still genuinely contested.

The deep time problem

This is where it gets uncomfortable. New Zealand’s coastline has been moving for the entire span of human presence here, and for a very long time before that. The interactive sea level tool on this site lets you drag through 40,000 years of NZ coastline change, and the picture it shows is worth sitting with. At the Last Glacial Maximum, around 21,000 years ago, sea level was 125 metres lower than today. The North and South Islands were joined. Stewart Island was joined to the South. Cook Strait was dry land. The whole continental shelf out to the 130-metre depth contour was above water, and New Zealand had a combined land area about 40 per cent larger than its modern footprint.

Between roughly 14,500 and 14,000 years ago, during an event known as Meltwater Pulse 1A, sea level rose around 20 metres in 500 years, an average rate of 40 millimetres per year, faster than anything projected for the next century under even the most extreme emissions scenarios. The cause appears to have been a non-linear collapse of the Laurentide ice sheet in North America. There were no humans in New Zealand to observe it. There were no humans in New Zealand to be blamed for it.

By the time Polynesian navigators arrived around 1300 CE, the coastline was already essentially modern. The whole 125-metre rise from the glacial low had happened in the 15,000 years before any human eye saw New Zealand. The first wakas landed at Golden Bay and Hawke’s Bay on a coastline that, viewed from a satellite, was within a couple of metres of the one we have today.

Then the climate stabilised. From about 7,000 years ago until the late 19th century, sea level was remarkably flat, never more than a few metres from where it sits now. Three hundred human generations lived inside that stability and built their assumptions around it. Most of the coastal infrastructure, ports, and farmland in New Zealand was constructed on the implicit belief that the sea level you build to today is the sea level you will still be building to in a hundred years. That belief was a fair inference from the previous 7,000 years and a poor inference from the previous 30,000.

The current rate of rise is around 4 to 5 millimetres per year and accelerating, roughly a hundred times the Holocene baseline. That is the genuinely unusual thing about now. Not that the sea is rising. The sea has been rising for most of the time humans have existed. The unusual thing is that we built our institutions during the one period when it wasn’t.

What this means for the framing

Two true statements coexist here, and most public discourse only carries one of them at a time.

The first. The current acceleration of sea level rise, the current intensification of extreme rainfall, and the current trajectory of ocean heat content are anthropogenic, are unambiguous in the data, and have consequences that we are nowhere near prepared for.

The second. The climate has always been changing, sometimes very fast, and many of the things now attributed to recent climate change, including coastlines moving, species shifting range, storms occasionally being severe, and individual seabirds occasionally washing up exhausted, would be happening in some form regardless of human activity. They are part of how the planet works.

Both can be true. The reflex tends to flatten them into one. A weather event happens, and the framing skips directly from the event to the attribution, without the intermediate step of asking how often that kind of event happened in the long pre-industrial baseline. Sometimes the answer is “less often, with high confidence”, as for heavy rainfall, heatwaves, and marine heatwaves. Sometimes the answer is “we don’t yet know”, as for total storm frequency, drought, and thunderstorm activity in the NZ region. Sometimes the answer is “it happened all the time”, as for sea level moving, coastlines reshaping, and individual seabirds being blown ashore by individual storms.

Treating the three as the same claim does a disservice to the first. If everything is climate change, then the strong, clean, testable attribution claims sit in the same conceptual bucket as the weak, gestural, pattern-matching ones, and the public learns to discount the bucket as a whole. That is bad for the science, and bad for the policy that depends on the science being trusted.

What better attribution looks like

The fix is not to deny the climate signal. The fix is to be specific about which signal, in which event, with which confidence level. The peer-reviewed literature has gotten quite good at this. Carbon Brief’s running map of attribution studies now catalogues hundreds of extreme weather events analysed individually. Of the studies it has aggregated, around 80 per cent found the event was made more severe or more likely by anthropogenic warming, around 5 per cent found the opposite, and the rest were inconclusive or showed no detectable signal. That distribution is the actual shape of the science. It is more interesting than the reflex it has been compressed into.

A good attribution claim names the variable (rainfall intensity, marine heatwave duration, fire-weather index), names the comparison (versus a counterfactual world without anthropogenic warming), and names the confidence level. A reflexive claim names “climate change” and stops.

A healthy version of the field also tolerates internal disagreement. Climate science, like any other physical science, advances by working scientists arguing with each other about whether a particular signal has been detected, whether a particular model parameterisation is sound, whether a particular attribution claim has properly accounted for natural variability. That argument is not a sign that the field is unsettled at its foundations. It is the sign that the field is alive. A discipline in which everyone publicly agrees with everything is not a science. It is a press release. The strong consensus that anthropogenic warming is real and consequential coexists, in the actual literature, with vigorous disagreement about magnitudes, rates, regional projections, feedback strengths, and what counts as a clean attribution. The public version of “the science” mostly omits the second half of that sentence, and the omission is itself a distortion.

For New Zealand specifically, the strong claims are about rainfall intensity, sea-level acceleration, marine heatwave frequency, and ex-tropical-cyclone severity. The weaker claims are about total storm count, drought (still uncertain at the regional level despite some site-specific increases), and thunderstorm frequency. Anything about individual species behaviour, including birds blown ashore, fish moving range, and insects appearing earlier, usually involves three or four causal variables tangled together, of which climate is one and not always the largest.

The drunk driver test

A useful discipline, when reading or writing a sentence that attributes something to climate change, is to ask whether the same sentence would survive substitution. If “climate change” can be swapped for “the weather” without losing any information content, then the sentence is doing rhetorical work, not causal work. If the sentence makes a specific quantitative claim, like “this kind of rainfall is now four times more common”, then it is doing causal work and can be checked.

A drunk driver who crashes on a rainy day did not crash because of climate change. He crashed because he was drunk. The rain was real, and it is even possible that the rain was slightly heavier than it would have been in a pre-industrial atmosphere. That is the honest version of the attribution claim. But the drunkenness is doing all the explanatory work for that particular crash. Reaching past it to the climate is a category error dressed up as concern, and it is the kind of move that makes the genuine attribution science harder to take seriously.

A storm-battered albatross arrives on a Taranaki beach because a storm hit it. Storms hit albatrosses. They have been hitting albatrosses for as long as there have been albatrosses, which is around 60 million years longer than there has been a Taranaki. Whether the storm that hit this particular albatross was made more intense by anthropogenic warming is a real, answerable, scientific question, one that requires a counterfactual climate model run, not an interview quote. The question is worth asking carefully. It is not worth answering reflexively.

The signal is real. The reflex is in the way of the signal. New Zealand, of all places, sitting on a coastline that has moved 125 metres in living geological memory, on islands that were one island when the first humans were leaving Africa, should be able to hold both ideas at once.

Where do you draw the line between attribution and reflex? Have you noticed the same flattening in NZ coverage, or is the framing pulling its weight? Share your thoughts in the comments below.