One more point in the graphics for our annual updates of comparisons of model observations. Given how extraordinary ones that have been the last two years, there are some prominent aspects to take into account.
First, we have updated the versions of some of the observational data sets: UAH TLT/TMT are now in version 6.1, and the surface temperature data of the NOAA NCEI surface are now version 6. We use the same Hansen81/Hansen88 collations /Cmip3 /cmip5 /output of the cmip6 model as previously. Comparisons cover surface air temperatures, sea surface temperatures, tropospheric atmospheric temperatures (TLT, TMT), stratospheric temperatures and some variations in these issues that have been of interest in the past. (It would be good to have some unbeated variables in the mixture: feel free to suggest some if you can point (or publish) a file of historical models+projected results).
With respect to the GMSAT, it is surprising how close the real world for Hansen et al. (1988) ‘Scenario B’ (this scenario had a ” ” ‘concentration as usual’ increases in CO2, but too much growth in CFCS and CH4. However, the most skilled projection award is still going to the CMIP3 set ;
The detailed problems that lead to an anguish around the CMIP5 models: the erroneous specifications of the forcing, the importance of the SST/SAT Blend vs. trends. SAT have faded in importance. These are/were real problems, but they are small compared to the ongoing trends. With respect to CMIP6, the observations (through a strip of diagnoses related to temperature) still coincide better with the subsample of selected models (that is, discarding those that ‘were hot’).
Updates with respect to atmospheric temperature profiles (diagnoses derived from MSU/AMSU) have become slightly more favorable for models, although structural variation between RSS data and recoveries of UAH/Noa stars is still clear. For the surface temperature of the sea, the real world seems to heat at the upper end of expectations, but still (just!) Within the projected propagation.
One of the main reasons for maintaining these comparisons is to see where discrepancies arise. To that end, multiple versions of the observation data are obviously useful since they can give an estimate of structural uncertainty (this has been very important in MSU/AMSU comparisons, for example). In other cases, we have less concerns about the observation data, but we are concerned that the models do not receive the right tickets. For example, since internal variability in stratospheric temperatures is much lower than in the lower atmosphere, incorrect forced signals can arise faster. I think we can be seeing some of that in SSU’s comparisons …
The party with the models is very good during the historical period (until 2014), but after 2015, there is some mismatch between the variance of the model and the obs. There are two potential problems: the time of the solar cycle 25 (a solar maximum heated the stratosphere), which occurred before and larger than expected by CMIP6, and the presence of the Hunga Tonga volcano (of 2021) that is having complex impacts on the stratospheric temperatures. However, long -term trends are still well modeled.
As always, if anyone knows the diagnosis of expanded model and the relevant observation data sets to compare, let it know and we can add it to the page.
Thanks as always to the data centers that provide the observation data, the CMIP committees that organized this storage of the outputs, the modeling centers that the executions and the authors that produced the derived data sets that we are using directly here were made ( Complete references on the page listed above).
See you next year!
#REALCLIMATE #Comparison #update
