Claims that the United States experienced its hottest summer on record in 2026 are being challenged over the way national temperature figures are calculated and the influence of urban development on long-running weather stations.
Anthony Watts, a senior fellow for environment and climate at the Heartland Institute, argues that headlines describing 2026 as the country’s “hottest summer ever” fail to convey important qualifications surrounding the National Oceanic and Atmospheric Administration’s temperature analysis.
NOAA reported that the average temperature across the contiguous United States between June and August 2026 was the warmest in its 132-year record, exceeding the summers of 1936 and 2021 by approximately 0.4°F.
Watts notes, however, that NOAA itself uses the term “warmest summer on record,” rather than “hottest summer ever.” He argues the distinction is significant because the record concerns an average incorporating both daytime maximum and overnight minimum temperatures.
Meteorologist Chris Martz, whose analysis is cited by Watts, says the 2026 result was driven substantially by unusually warm overnight minimum temperatures rather than unprecedented daytime heat.
When summers are compared using average daytime maximum temperatures, Martz says 2026 falls behind several earlier years, including the severe Dust Bowl summer of 1936.
The geographical distribution of the heat also raises questions about how the national record should be interpreted.
According to Martz’s examination of NOAA data, only seven of the country’s 344 climate districts — about 2 percent — recorded their warmest summer in 2026, with most located in the Desert Southwest. By comparison, he says 51 climate districts established summer records during 1936.
Watts argues this provides important context missing from reports that could leave readers with the impression that unprecedented summer temperatures occurred throughout much of the country.
He also raises questions about NOAA’s methodology.
The agency’s long-term temperature record does not rely exclusively on unmodified observations. Historical station measurements undergo homogenisation procedures intended to compensate for factors including changes in instruments, observation times and station locations.
Such adjustments are a standard part of NOAA’s national temperature analysis, but Watts argues their significance warrants greater attention when the difference between competing record years is only a few tenths of a degree.
He cites Martz’s contention that some adjustments can exceed the approximately 0.4°F separating 2026 and 1936. Martz has also questioned whether NOAA’s homogenisation methods fully eliminate gradual warming associated with urban development.
Watts points to the United States Climate Reference Network as another source of temperature information that he believes complicates the picture.
The USCRN was specifically designed as a high-quality climate-monitoring network, with stations selected and maintained to reduce problems associated with urban development and other changes around older observation sites.
According to Watts, USCRN measurements recorded an exceptional temperature anomaly during March 2026, reaching nearly 7.7°F above average nationally. By contrast, he says temperatures during June, July and August were less exceptional when viewed against the approximately two decades covered by the network.
Watts argues the difference between the USCRN observations and NOAA’s longer-term national ranking deserves closer investigation.
He suggests urban heat island effects could be one contributing factor. Older weather stations are sometimes situated at airports or in locations that have become increasingly urbanised over decades. Roads, concrete, buildings and other development can retain heat, particularly overnight.
That is potentially significant because elevated minimum temperatures were an important contributor to the 2026 national average.
Watts cites weather stations in Reno, Nevada, and Phoenix, Arizona, as examples where researchers have questioned whether local development has introduced warming biases into official measurements.
He also criticises NOAA’s treatment of stations that have closed or ceased reporting. Missing observations can be estimated using measurements from surrounding stations in order to maintain spatial and historical continuity in climate datasets. Watts describes these as “ghost stations” and argues estimated values should receive greater scrutiny when they contribute to national temperature assessments.
His broader criticism is directed at media reporting rather than simply NOAA’s finding.
Watts contends that reporting of the 2026 record has frequently presented the national ranking without adequately explaining the role of adjusted datasets, estimated observations, urbanisation and differences between average temperatures and daytime maximums.
He argues that a record separated from previous summers by only 0.4°F should be reported alongside the methodological uncertainties and alternative measurements relevant to interpreting it.
The dispute does not necessarily turn on whether the United States experienced an unusually warm summer. Rather, it concerns what can reasonably be concluded from different temperature datasets and how confidently a narrowly established national record should be portrayed.
For Watts, describing 2026 simply as America’s “hottest summer ever” removes much of that complexity. He maintains that climate reporting should distinguish between measured and adjusted observations, explain the uncertainties surrounding close temperature records and give readers sufficient information to understand how such records are produced.
