Picture the sun at left streaming shortwave energy onto the Northern Hemisphere. A gray-brown sulfate haze shell brightens and reflects sunlight as industrial emissions grow; a teal greenhouse shell thickens with CO2 and traps outgoing infrared; volcanic veils flash after Krakatoa 1883, Santa María 1902, Novarupta 1912, Agung 1963, El Chichón 1982, and Pinatubo 1991.
Everything below still works: the model, the chart, verdicts, persistence and the exported report all run without 3D.
Forcing attribution lab · 1880–2020
The quoted passage is Hansen et al. 1981 stating the puzzle. Solve it the way they did: find the forcing mix that fits the record.
Stage 1 of 3. The record really does dip between 1940 and 1970. Commit to a prediction, then run the model with CO₂ only and see whether it can produce that dip.
With CO₂ rising every year, can a CO₂-only model reproduce the 1940–1970 cooling?
3.0 °C per CO₂ doubling
“Global temperature barely rose 1998–2012 despite record CO₂ emissions — so CO₂ can’t be driving warming.” Your strongest reply:
Progress saves locally in your browser. The report records your prediction, every run’s forcing mix, sensitivity, error scores, verdicts, and the reasoning.
three not loaded
scene starting…
“The major difficulty in accepting this theory has been the absence of observed warming coincident with the historic CO₂ increase. In fact, the temperature in the Northern Hemisphere decreased by about 0.5 °C between 1940 and 1970 (9), a time of rapid CO₂ [buildup].”
That is Hansen, Johnson, Lacis, Lebedeff, Lee, Rind & Russell, Climate Impact of Increasing Atmospheric Carbon Dioxide, Science 213 (1981) — climate scientists stating the difficulty on the way to resolving it. The same paper fit the record with greenhouse, aerosol, volcanic, and solar forcing together, and projected the warming that the following four decades then delivered. Quoting the difficulty without the resolution inverts what the paper showed.
The model is a real zero-dimensional energy balance: T′ = T + (Fnet − λT)/C each year, λ = 3.7/ECS W·m⁻²·K⁻¹, C = 14 W·yr·m⁻²·K⁻¹, re-baselined to 1880–1920. Forcing and temperature series are digitized decadal approximations of published GISS/IPCC-style estimates (aerosols NH-weighted); they are not live data and carry real uncertainty. Hansen’s 1981 dataset showed about −0.5 °C for 1940–1970; modern smoothed reconstructions show a smaller but real −0.2 to −0.3 °C dip, which is what the chalk line here encodes. There is no ocean/land split and no internal variability — which is why fitting one window is necessary but never sufficient.