Somewhere in College Park, Maryland, a bank of computer servers hums silently through the night, processing satellite data, atmospheric pressure readings, and ocean temperatures spanning decades. NOAA’s operational branch for long-range forecasting, the Climate Prediction Center, never sleeps. It appears that the warming doesn’t either.
Even by today’s standards, the 2024 numbers were startling. Last year was the warmest since global record-keeping started in 1850, according to NOAA’s National Centers for Environmental Information. Not by a slim margin, but by 0.18 degrees Fahrenheit over the previous record, which had only been set the year before. Two years in a row that broke records. It’s the kind of pattern that causes even cautious scientists to stop in their tracks.
The average surface temperature of Earth has increased by about 0.11 degrees Fahrenheit every ten years since 1850. That sounds insignificant, almost unmemorable. However, if you look at the years after 1982, the number more than triples—0.36 degrees Fahrenheit every ten years. Something picked up speed. How climate scientists discuss their own predictions has been subtly altered by the controversy surrounding what, why, and whether early models captured it.
The movement of energy through the atmosphere, oceans, land, and ice is recreated mathematically in climate models. Built on thousands of data points, calibrated against historical records, and tested backward against previous climate events before being trusted to look forward, they are, in many respects, remarkable accomplishments. By the end of the century, the IPCC has long predicted warming of between 1.8 and 5.8 degrees Celsius. It’s a broad range, and the range itself conveys a sense of ambiguity.
In contrast, NOAA’s prediction centers monitor the signal in real time. They measure the heat content of the upper ocean, track sea surface temperatures, and keep an eye on El Niño patterns developing in the equatorial Pacific. That heat content, which may be more significant than air temperatures but doesn’t make headlines as frequently, reached its highest level ever recorded in 2024. Most of the extra heat is absorbed by the oceans. The atmosphere eventually follows when they are operating at a high temperature.
The peculiarity of that 2-degree Fahrenheit increase since the pre-industrial era is worth pondering. It appears to be tiny. Thirty-degree temperature swings before lunch are familiar to anyone who has experienced a cold spring. However, climate scientists quickly note that even a 1-degree Celsius change in the global average amounts to a massive redistribution of energy throughout the entire planetary system, including the oceans, ice sheets, jet streams, and monsoon cycles. The entire apparatus reacts.
Speaking with researchers who use these models gives the impression that they are becoming less shocked by the warming trend itself and more concerned about its speed. Regarding the direction, early models were correct. They might have been too cautious with the speed. Not all early projections accurately or urgently predicted that the ten warmest years in recorded history would occur within a single decade, from 2015 to 2024.
To be honest, no model is comprehensive. Technically speaking, climate systems are chaotic; minor fluctuations cascade in ways that are genuinely hard to predict. With the addition of improved ocean circulation data and multimodel ensemble approaches that average across multiple competing simulations, NOAA’s prediction frameworks have significantly improved. However, equations do not negotiate with the atmosphere.
The planet appears to be reacting more quickly than the middle estimates indicated, which is becoming more difficult to ignore. Not the worst-case scenarios, which are still ahead of us and theoretically preventable. However, the comfortable middle ground appears to be getting smaller. That’s the subdued conclusion that keeps building up in the data from Maryland servers that never stop counting, year after year.
