Reconciled climate response estimates from climate models and the energy budget of Earth

Mark Richardson, Kevin Cowtan, Ed Hawkins, Martin Stolpe


Why did you use the HadCRUT4 temperature record rather than the NASA or NOAA temperature records in your analysis?
We used HadCRUT4 because the papers which we were re-evaluating also used HadCRUT4. In order to determine why they obtained the results they did we had to use the same data.

Different versions of the temperature record are suitable for answering different questions. When comparing observations to climate models, HadCRUT4 is a good choice because the other records 'infill' the unobserved regions in the early record, which can distort the comparison. A like-with-like comparison using HadCRUT4 avoids this problem, as long as the same coverage is used for both models and observations.

Infilling can reduce the uncertainty in the temperature record when the unobserved regions are small (e.g. the Arctic in the recent record). However when the gaps are large (such as in the early record) infilling generally assumes that the unobserved regions behave similarly to the observed regions. Both models and observations reject this assumption. When using an infilled record it is impossible to determine whether deviations between the observations and the models arise from problems in the infilling or problems in the models.