UC observed a couple of days ago that Hadley Center, authors of the pre-eminent temperature series, have suddenly identified an “error” in how they presented temperature data. For presentation of their smoothed temperature series in a part-year situation, their methodology calculated the average of months then available and used that to estimate the current year’s temperature for presentation purposes. For their influential graphic showing smoothed temperature series, they used a 21-point binomial filter (this is reported) extrapolating the latest number for 10 years. This obviously places a lot of leverage on January and February temperatures. (UC has replicated their smoothing method; he sent me code and I’ve confirmed that we can exactly replicate their smoothing methods.)
As has been widely reported, January and February 2008 temperatures are noticeably lower than last years. Here is a plot of Hadley Center GLB monthly temperatures, showing the two 2008 months in bold points. This ties into the monthly plot at the HadCRU site here.
These cold January and February 2008 temperatures have led to a noticeable downturn in the smoothed annual series. This has not escaped the notice of the Hadley Center, who were extremely quick off the mark to notice an “error” which resulted in graphical emphasis of a downturn (here)
We have recently corrected an error in the way that the smoothed time series of data were calculated. Data for 2008 were being used in the smoothing process as if they represented an accurate estimate of the year as a whole. This is not the case and owing to the unusually cool global average temperature in January 2008, the error made it look as though smoothed global average temperatures had dropped markedly in recent years, which is misleading.
Heading into the IPCC WG1 conference in Paris in February 2007, January 2007 was a very warm month. I thought that it would be interesting to plot the HAdCRU style result as of January 2007 and compare it to the January 2008 style (now excised from the website). The blue dots below show the effect of the CRU smoothing method used in 2007 incorporating Jan and Feb 2008 – showing the downturn, which caused the Hadley Center to notice the “smoothing error”. The black shows the present annual series – not using 2008 data – which is what is currently displayed on the Hadley Center website (prettied up and with pseudo-“error” bars.) The red dots show what their 2007 method would have yielded in February 2007, at the time of the IPCC WG1 conference.
They noticed the “smoothing error” like a dog on a bone when temperatures went down, but didn’t notice precisely the same “error” last year, when it yielded record high results. Looks like there are some pit bulls in England as well.
One minor curiosity which some reader may be able to explain. I compared the HadCRU GLB annual series – column 2 in http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/annual with the HadCRU GLB monthly series – column 2 in http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/monthly. I calculated annual averages from the monthly version and compared them to the archived annual version with the following result. I then manually compared values for 1861 (monthly: -0.811 -0.477 -0.491 -0.375 -0.765 -0.172 -0.308 -0.173 -0.379 -0.397 -0.410 -0.191 with an average of -0.4124167; as compared to the annual of -0.568. Perhaps this is explained somewhere. I didn’t see any explanations in the website explanations – if any one sees an explanation, I’d be interested.
Merely from looking at the monthly temperature histories, I urge readers not to draw any particular conclusions from a couple of cold months. The monthly history has many such cold downspikes and recoveries tend to be quite rapid. If the HadCRU results are an accurate history, one could just as easily look at this graph and argue that the most recent downspike was not as cold as corresponding downspikes in the 1980s. (Evaluating the HadCRU results is very difficult because their data as used is not disclosed.)
From my immediate view in Toronto, we still have banks of snow, which will still be here at the beginning of April in two days. I certainly don’t recall such a situation during my adult life – so the present downspike seems a little unusual from a Toronto perspective but I recognize that this is only one perspective.