Global SST, WWII, global cooling, and forgetting

One of the things that made global cooling plausible was the observed temperature record although you have to be a bit wary, when talking about what people thought in the 1970s, because of course the record they had available then wasn't as good as the one we have now. But anyway, as you can see from that pic, the "cooling" from the 1940s to the 1970s was more of a plateau, coupled with a peak during WWII (insofar as it makes any sense to talk about a record in that way).

This was always a bit problematic, but in 2008 quite a bit of it disappeared, when A large discontinuity in the mid-twentieth century in observed global-mean surface temperature (Thompson et al., Nature 453, 646–649 (29 May 2008) doi:10.1038/nature06982) appeared; also known as "Post-World War II cooling a mirage". I noticed this at the time; and JA was characteristically caustic; but other than making the modelling easier it didn't really have much consequence and I don't know quite how it got folded into the records. Not very much, I suppose, since I've just pointed you to a recent record with the same problem in it.

And now, Kevin Cowtan, Robert Rohde, Zeke Hausfather have discovered the same thing. At least, it looks very much like the same thing. Sou notices the new stuff, but doesn't refer to the old. Cowtan's blog explainer doesn't mention "Thompson". The paper itself does: "(Thompson et al. 2008) detected an inhomogeneity in the sea surface temperature record arising from a change in the shipping fleet at the end of World War 2 by comparison of sea surface temperatures to temperatures from coastal weather stations and from climate models".


wereatheist said...

The photo looks un-European to me.

PaulS said...

The cooling addressed in the Thompson paper was more of the sudden dip variety, immediately after the hump, rather than the multi-decadal trend. The adjustment was implemented in the HadSST3 release. It actually, if anything, enhanced the appearance of a cooling trend compared to HadSST2.

The ERSST peeps abandoned metadata-based adjustments entirely for ship-to-ship biases, instead relying on comparison to HadNMAT. That doesn't appear to work very well around the WWII area.

Cowtan et al. are trying a different non-metadata approach for the whole record. They do seem to reproduce the bias identified by Thompson quite nicely, but also suggest the whole period from early 1940s through mid-1970s was almost uniformly about 0.15ºC cooler than the other records.

William Connolley said...

It's less the approach to fixing it that I was thinking of; more the method for detecting the problem. Both are using ships-near-islands.

Paul Kelly said...

Is this anything? https://www.nature.com/articles/nature25152

William Connolley said...

Might be interesting. Is paywalled. He has a blog: http://climatescientistblog.info/.

Hank Roberts said...

Evolving Understanding of Antarctic Ice‐Sheet Physics and Ambiguity in Probabilistic Sea‐Level Projections (pages 1217–1233)
Robert E. Kopp, Robert M. DeConto, Daniel A. Bader, Carling C. Hay, Radley M. Horton, Scott Kulp, Michael Oppenheimer, David Pollard and Benjamin H. Strauss
Version of Record online: 14 DEC 2017 | DOI: 10.1002/2017EF000663

Key Points

Incorporating ice‐shelf hydrofracturing and ice‐cliff collapse mechanisms highlights ambiguity in post‐2050 sea‐level projectionsThese mechanisms make post‐2050 sea level more heavily emissions dependent and can significantly revise high‐emissions projections upwards
Current Antarctic retreat is dominated by by different processes than and exhibits little correlation with late‐century changes


William Connolley said...

There's lots of stuff with ice shelves at the moment. In the old days, I'd have cared about the details. Now I'm content to wait a little while to see how it settles down.