Eric Rosten and Blacki Migliozzi over at Bloomberg have put together an excellent set of infographics over at Bloomberg which show global land and ocean temperature observations from 1880 onwards, and compares the changes therein to simulations of the net effect of natural and anthropogenic causes conducted by NASA GISS using ModelE2.
For those interested enough to scroll beyond the interactive plots, the article does a pretty good job of including some technical details in an approachable way:
Research groups were asked to see how well they could reproduce what’s known about the climate from 1850-2005. They were also asked to estimate how the various climate factors—or “forcings”—contribute to those temperatures.
Researchers do not expect their models to reproduce weather events or El Niño phases exactly when they happened in real life. They do expect the models to capture how the whole system behaves over long periods of time. For example, in 1998 there was a powerful El Niño [… a] simulation wouldn’t necessarily reproduce an El Niño in 1998, but it should produce a realistic number of them over the course of many years.
The temperature lines represent the average of the model’s estimates. The uncertainty bands illustrate the outer range of reasonable estimates.
In short, the temperature lines in the modeled results might not line up exactly with observations. For any year, 95% of the simulations with that forcing will lie inside the band.
Interesting read and a fine reminder that delivering research products like this to a broader audience is an important component of scientific work today.