The Economist has a summary of what yesterday’s IPCC report says. Here is a graphic that Deutsche Bank research is circulating.
Source: Jim Reid, Chart of the Day, Deutsche Bank, 9 August 2021.
I had hoped to have something to contribute on how to assess the forecasting accuracy of the IPCC reports, but this is harder to do (in an easily understandable language) than I had thought. So I pass on some informal thoughts circulating amongst some people in the know, related to me by my colleague Greg Nemet.
- Global average temperature is the most basic prediction, so a good one to focus on
- Each IPCC assessment (2000, 2007, 2014, 2018, 2021/2…) updates temperature projections based on recent temperature observations.
- The way any changes get most clearly revealed is in the “carbon budget” ie how many tons of CO2 available to put in the atmosphere before reaching +2 degree temperature change.
- There was an expansion of the carbon budget in the 2014 report because models had been “running hot” compared to observations
And he also points me toward this twitter thread from Zeke Hausfather (I don’t usually link to tweets, but I’m making an exception here). That thread describes how they narrowed the “band of sensitivity”, i.e., the likely ranges, using outcomes vs. predictions. Using my language of forecasting (which is probably not very accurate of the proper way of describing the statistics), the point forecast is relatively unchanged, but the forecast interval is smaller.