I haven't written anything about global warming for a while (although I have not been idle on that front - watch this space) but there's a great story at Climate Audit this morning.
While a great deal of sceptic attention is focused on the problems with the land surface temperature record (poor station siting, dubious adjustments) it is important to remember that in terms of detecting the alleged global warming at surface level, ocean temperatures are far more important, the seas representing a much larger proportion of the world's surface than land.
So, what's the news?
Well, it seems that a new version of the sea surface temperatures has been released, incorporating a whole bunch of "improvements" to the way they are put together.
Guess what effect these improvements have had on the trend?
Yup, with the improvements in place, the new version 3 shows that the seas are warming much faster than we thought back at version 2.
So what was this improvement? Well, it's too early to be sure but it looks as if the improvement involves a new way of dealing with sparse data. It seems that where there is not much information to work with, the scientists simply insert some numbers generated by a climate model. In other words the new sea surface record is heavily fictionalised.
Even funnier, Professor Ben Santer a man who is probably best known for having been accused of doctoring one of the early IPCC reports, wrote in the International Journal of Climatology that the climate models were splendid and marvellous because they could now accurately predict tropical sea surface temperatures. This is not really very surprising now we know how it seems that the sea surface temperature record is based partly on that same model output.
As someone used to say: hey it's climate science.
Looks like this was a false alarm, at least as far as coercing the data to model output is concerned. It looks as though a woolly explanation of what the methodology was has lead people astray.
That doesn't mean there isn't a problem though. It looks as if what actually happened was that they used models to determine certain parameters in the sea surface temperature algorithm. These changes then lifted the temperature trend as described in my main post.
What this means is that the sea surface temperature calculation looks to be non-robust, and in a big way. If your trend is dependent on some model input you feed into your calculation, then the temperature record is still hypothesis rather than evidence of the truth or otherwise of the hypothesis.
Article originally appeared on (http://www.bishop-hill.net/).
See website for complete article licensing information.
Bishop Hill