Robust comparison of climate models with observations using blended land air and ocean sea surface temperatures

Kevin Cowtan, Zeke Hausfather, Ed Hawkins, Peter Jacobs, Michael E. Mann, Sonya K. Miller, Byron Steinman, Martin B. Stolpe, Robert G. Way

FAQ

Why don't the marine air temperature observations show faster warming than sea surface temperatures?
The difference between the marine air temperature sea surface temperature observations is most likely dominated by observational biases.
more/less

The difference between marine air and sea surface temperatures in models is around 0.05C. By contrast, the known uncertainties in the marine air and sea surface temperature observations are both around 0.1C (Kent et. al. 2013 Figure 18). Therefore we do not expect to be able to detect the blending bias in the observations.

The difference between the marine air and sea surface temperature observations is much greater than, and of a different character to the difference between marine air and sea surface temperatures in models (Figure S7). How can this be explained? There are four possibilities:

  1. Both the marine air temperatures and the sea surface temperatures contain large problems. If this is the case, then we don't know what the observed temperatures are doing well enough to use them to critique the models.
  2. The sea surface temperatures are more realistic than the marine air temperatures. This is the assumption that has driven the use of sea surface temperatures in the observational record. If this is the case, then we should also be using sea surface temperatures from the models, with the consequences described in our paper.
  3. The marine air temperatures are more realistic than the sea surface temperatures. If this is the case, then the observational record should be changed to use marine air temperatures instead of sea surface temperatures. Against this there is evidence of problems in the marine air temperature record around WW2 - however that does not necessarily invalidate the record for other periods.
  4. Both the marine air temperatures and the sea surface temperatures are realistic. If this is the case, then new physics is required to explain how the world can maintain multidecadal shifts in the difference between marine air temperature and sea surface temperature. No mechanism which could support such changes is present in current climate models.

The relative importance of errors and biases between the marine air and sea surface temperatures is best determined by the observational data providers, however the uncertainties suggest that the question is currently unresolved. As users we have to rely on the best data that they can provide (including the uncertainties) and do the most appropriate comparison that we can using those data.