It doesn’t matter how beautiful your theory is, it doesn’t matter how smart you are. If it doesn’t agree with experiment, it’s wrong.
The above principle, stated by Nobel Prize-winning physicist Richard Feynman, is one of the foundations of science, yet this seems to have eluded many in the world of climate science.
If rising atmospheric carbon dioxide is the primary driver of recent warming, we should expect relatively uniform warming across the months of the year. Carbon dioxide is a well-mixed gas; its radiative effect does not change strongly from one month to the next. The world’s longest continuous instrumental temperature record tells a different story.
The data and the method
The Central England Temperature (CET) series, maintained by the UK Met Office, runs from 1659 to the present and is the longest continuous instrumental temperature record available. For each calendar month I calculated the linear trend over two periods — 1659-2025 and 1850-2025 — and expressed the result as degrees Celsius per century. This produces 12 independent rates of change that can be compared directly. Because the same month is compared with itself across centuries, many of the practical difficulties that affect long temperature series (instrument changes, site moves, time-of-observation bias) are held constant.
Full record: 1659-2025
When the monthly rates of change are plotted for the entire 365-year period, a clear pattern appears. March is one of the strongest warming, roughly 0.45-0.5 °C per century. June shows the weakest, around 0.05°C per century — more than eight times slower. Winter months (December-February) and October also warm relatively quickly; late spring and early summer warm far more slowly.
The industrial period: 1850-2025
Restricting the analysis to the modern industrial era produces higher overall rates, as expected. Yet the seasonal contrast remains. March still warms roughly three times faster than June. October and November now show the highest rates, both exceeding 1°C per century, while June remains the slowest of the 12 months.
What the pattern implies
The data confirm that central England has warmed. They also show a persistent, large difference between the warming rates of individual months. A forcing that is essentially constant through the year does not, by itself, predict that March should warm three to nine times faster than June. Seasonal differences can of course arise from other processes — changes in snow cover, cloud, atmospheric circulation or land-surface feedbacks. The magnitude and longevity of the March-June contrast therefore raise a legitimate question: is carbon dioxide the dominant control, or are other factors playing a larger role than the prevailing narrative assumes?
Limitations
CET is a regional series, not a global one. Linear trends calculated over centuries can still be influenced by earlier natural variability. Nevertheless, it remains the longest continuous instrumental record we possess, and the monthly pattern is robust across both the full series and the industrial subset.
The observations do not ‘disprove’ a role for carbon dioxide. They do show that the simple hypothesis — CO₂ as the primary driver producing broadly uniform warming — fails a straightforward test against the longest temperature series available. Something else is clearly at work.
For a longer discussion of possible physical explanations, see my previous article for the Daily Sceptic and my conversation with Tom Nelson:
First published on Stephen Andrews’s Substack page. Subscribe here.
Since the 1970's the Clean Air Acts have removed Sulphur Dioxide from the atmosphere, hence reduced cloudiness, let through more heat from the sun and hence contributed to 0.5 dec C.
The counter argument is that this is just uncovering earlier warming but it also means that the warming since the1970's attributed to CO2 is 0.5 deg less than claimed.