A newly released peer-reviewed study analysing more than a century of United States temperature records has reported a decline in both extreme heat and cold events, raising questions about widely cited claims of accelerating temperature extremes.
The research draws on the United States Historical Climate Network (USHCN), extending its dataset from 1899 through to 2025. Originally comprising 1,218 monitoring stations selected by National Oceanic and Atmospheric Administration in the 1980s, the network has since seen significant attrition, with nearly half of the stations closing after 2000. To maintain continuity, researchers supplemented incomplete records with nearby stations showing strong correlation, producing usable time series for 1,211 locations.
The study focused on extreme temperature metrics, including peak summer highs and winter lows. Its central finding was that “metrics for extreme summer heat… show modest negative trends since 1899,” while extreme cold events have also declined, particularly since the 1990s. Overall, the authors concluded that “instances of both hot and cold extreme metrics have declined since 1899.”
The findings contrast with conclusions presented in the Fifth National Climate Assessment, which states that extreme heat events are increasing in frequency and intensity.
Reacting to the study, Earth science professor Matthew M. Wielicki highlighted the apparent discrepancy. In a post on X, he wrote: “A study released today using actual U.S. station data (1899–2025) finds no significant trend in summer daytime highs and even declines in some temperature extremes.” He added: “Because the Fifth National Climate Assessment confidently told us extreme heat is accelerating. The data, however, begs to differ.”
The study’s authors noted that their dataset provides a long-term, station-based perspective on temperature extremes, offering a different lens from “model-based assessments” often used in national climate reporting. They also emphasised the importance of maintaining high-quality observational networks to better understand long-term climate variability and trends.
New peer-reviewed research just dropped… and it’s awkward.
A study released today using actual U.S. station data (1899–2025) finds no significant trend in summer daytime highs and even declines in some temperature extremes.
Let that sink in.
Because the Fifth National Climate…
— Dr. Matthew M. Wielicki (@MatthewWielicki) April 21, 2026
Image credit: Karsten Wurth
Not surprising. There is no evidence e of sea level rise either.
Exactly. Look at the big “ beef ups” the national news media are giving recent events. They try to link all these weather events to climate change.
Weather is weather! It varies. No sea rise. More ice on the caps etc.