The Science of Climate Change: What the Physics Actually Says

"Climate has always changed" is the most common sophisticated skeptic argument about global warming. It's true. The climate has always changed. The question is what's causing the current changes — and attribution science is the field that answers that question with quantified confidence rather than handwaving.

Attribution science applies the same basic methods that epidemiologists use to determine whether a specific factor causes a disease. You can't prove that a specific person's lung cancer was caused by smoking. But you can show that smokers develop lung cancer at a much higher rate than non-smokers, controlling for other variables, and you can identify plausible biological mechanisms. The same logic applies to climate attribution: you can't prove that a specific storm was "caused" by climate change. But you can show that certain types of extreme weather are occurring more frequently than pre-industrial baselines, quantify the difference, and identify the physical mechanisms.

For global temperature trends, the attribution case is extremely strong. The global average surface temperature has risen about 1.2°C since the pre-industrial baseline (defined as 1850-1900). The rate of warming since 1970 is about 0.18°C per decade — faster than any period in the instrumental record, and faster than can be explained by natural forcing factors alone.

How do we know it's human-caused and not solar variation? Several lines of evidence. First, solar output actually declined slightly from roughly 1980 to 2010 while temperatures rose. If solar variability were the driver, you'd expect a correlation between solar output and temperature; the data don't show it. Second, the stratosphere has cooled while the troposphere has warmed — this is a fingerprint of greenhouse gas forcing. If warming were solar-driven, both layers would warm together. Third, the warming pattern (stronger at night than day, stronger at high latitudes) matches greenhouse gas forcing predictions but not solar forcing patterns.

For individual extreme events, attribution science has become much more sophisticated since the early 2000s. The basic approach is counterfactual: run climate models with and without human greenhouse gas emissions, compare the probability distribution of a specific event type in each scenario, and report how much more (or less) likely the event is in the observed climate versus the climate that would exist without human forcing.

Results from this literature: the 2021 Pacific Northwest heat dome (temperatures 5°C above the previous all-time record in Portland) was assessed as "virtually impossible" without human-caused climate change. The 2019-2020 Australian bushfire season was found to be 30% more likely due to human-caused warming. European heat waves have become dramatically more likely, with the 2003 heat wave (which killed ~70,000 people) now estimated to be about twice as likely due to climate change.

This isn't claiming that climate change "caused" these events in the sense of being the sole factor. Weather events have multiple contributing factors. What attribution science shows is the probability change — how much more (or less) likely events of a given magnitude or type become under actual climate conditions versus the counterfactual without human forcing. The answer for extreme heat events, in particular, is "substantially more likely," and the quantified increase keeps getting larger as warming accumulates.

Attribution Science — How We Know Which Changes Are Human-Caused

// COMMENTS

ON THIS PAGE