I keep coming back to a finding I can't quite metabolize. A study published in February in PLOS Climate tested every U.S. state for warming across the full range of its temperature distribution. The average, yes, but also the hottest days, the coldest, everything in between. Eighty-four percent of states showed statistically significant warming somewhere in that range.

Texas showed none. At no tested level. Extremes, middle range, cold snaps. Zero signal.

Texas, where more than 300 people died from heat in 2023, the most since the state began tracking. Where Hurricane Harvey dumped so much rain on Houston that attribution studies found the rainfall was 38 percent higher because of global warming, displacing over a million people, destroying or damaging more than 200,000 homes and businesses. Where, in February 2021, 4.5 million homes lost power during Winter Storm Uri and the grid came within four minutes and thirty-seven seconds of a complete statewide blackout that could have taken weeks to reverse. Where the people who froze to death froze in their own houses, in a state whose temperature record, according to this study, shows no warming signal at all.

No warming signal. I keep reading that and then reading it again.

What the study actually found

The researchers, Gadea Rivas and Gonzalo, did something that sounds obvious once you hear it. Traditional climate analysis tests whether a state's average temperature is rising. They tested the whole shape. Co-author Gonzalo compared it to understanding a country's inequality by looking only at GDP per capita:

By conventional mean-temperature analysis, 44 percent of states show no significant warming trend. Test the full distribution, and that number drops to 16 percent.

Then there's the cluster where the study detected no warming at any tested level: Texas, Oklahoma, Kansas, Arkansas, Mississippi, Alabama, and others. Scientists have studied this "warming hole" for years. It's a real meteorological phenomenon. It is also, almost certainly, temporary, built on atmospheric and oceanic patterns being overtaken by the forcing they offset.

Where the data and the body agree

Arizona is the inverse case. In Maricopa County, 645 people died from heat in 2023 and 608 in 2024. Phoenix endured 113 consecutive days at or above 100°F in 2024, shattering the previous record of 76. Grid operators set new peak demand records on a 118-degree day in 2025. Among indoor heat deaths, 88 percent had an air conditioning unit present. Seventy percent of those units weren't working.

Arizona is warming, and the data says so, and planners and public health officials are at least working from a shared reality, however inadequately they respond to it.

That shared reality matters more than it sounds like it should. And in Texas, it's missing.

Living inside the gap

Temperature trends are what planning offices use, what risk models are calibrated to. They shape public perception of whether urgency is warranted. In Texas, the trend line and the lived experience of climate disruption have diverged so completely they describe different places.

Insurance markets have noticed. Between 2020 and 2024, Texas experienced 68 billion-dollar weather disasters, more than in the entire preceding decade. Premiums have risen roughly 57 percent since 2015. Carriers are pulling back from the state. One in six Texas homeowners reportedly went without coverage, citing affordability. Actuaries read the claims.

But actuaries are paid to see what's coming. What about a city planner in Houston deciding whether to fund cooling infrastructure? A county official allocating emergency preparedness money? A school superintendent weighing whether the budget can absorb upgraded HVAC? The absence of a warming signal in the data is, for these people, the thing someone points to in a meeting to argue the problem is elsewhere. The budget line that doesn't get funded. A reason, sitting right there in the official record, to wait.

The study's authors are careful. They're reporting what temperature trends show, and they're clear that a stable trend line says nothing about whether climate change is reshaping a state's weather. But temperature trends are the instrument we've pointed at the problem, and when the instrument reads stable while the reality underneath it destabilizes, the gap between them is where preparation doesn't happen.

The warming hole will close. The conditions producing it are eroding. And when it does, the states inside it will catch up all at once to a trend that's been accumulating underneath the signal's absence, in a landscape of infrastructure and planning decisions made during the years the thermometer said everything was fine. A decade or more of under-preparation meeting the full weight of what was always coming.

What do you do when the way you measure a problem can't see the problem? The insurance industry built its own instruments. Everyone else is still reading the thermometer.