By Andy May
On January 22, 2021, John Christy presented a new online talk to the Irish Climate Science Forum. The talk was arranged by Jim O’Brien. A summary of the presentation can be read at clintel.org here. In this post we present two interesting graphs from the presentation. These compare observations to the IPCC Coupled Model Intercomparison project 5 climate models (CMIP5, 2013) and CMIP6 (current set of IPCC models) climate model projections.
The next graph compares the newer CMIP6 models to both the weather balloon data (light green) and the weather reanalysis data (dark green).
The difference between the models and the observations is statistically significant and shows that the models have been invalidated. It is also significant that the CMIP6 spread of model results is worse than the CMIP5 spread. Thus, the newer models show less agreement to one another than the previous set.
It looks like the divergence of the models from measurements occurred during the 1990s. There is excellent agreement over the first 20 years of these datasets.
Why? We can conjecture a lot, which is likely only to point blame at our favorite target, but what actually happened to cause the models to diverge?
It appears the divergence is because the models are too sensitive to CO2. They are missing a natural cause of climate change.