Wood Romances

A paper was published this week that argues that low-sensitivity models do a poor job of reproducing CERES-derived EEI trends. In the words of the paper, the authors used CMIP6 models "to illustrate that low climate sensitivity models have an EEI trend behavior that is inconsistent with the satellite-derived EEI trend." Even though models with an ECS near 3°C do a good job of reproducing current warming, CMIP6 models often differ in the in EEI trends. For instance, The CERES data show a stronger trend in EEI than the multi-model CMIP6 mean and higher EEI in 2023 than any of the CMIP6 models. However, for individual CMIP6 models and ensembles, EEI is comparable to or higher at other periods than the CERES value in 2023. The difference in trends can be seen by comparing the red CERES line to the black CMIP6 model mean. Even though the interannual variability in the CMIP6 models is consistent CERES observations, the observed trend in EEI, especially since about 2010, is higher t...

The graph below from John Shewchuk is intended to show that models predict too much warming compared to satellite data. Shewchuk claims that the red line is the average of 102 IPCC CMIP5 model runs for the surface through 50K feet. In all likelihood, this is just lifted from graphs of model-observation comparisons that John Christy has plotted in the past. Shewchuk claims the blue line is UAH satellite observations for the total troposphere layer (TTT). The problems with this graph are numerous, and many of the problems are inherited from Christy's graph (problems with his comparison are well-documented ). The two time series are separated from each other even in 1980 to exaggerate the differences between the two (even if the trendlines do intersect at 1980). The 102 model runs are not shown, only the model mean, and the 95% confidence interval is also not shown, so we have no idea what the spread in the model runs might be. But Shewchuk has added his own dishonest twist to this g...