1995 And All That

Last year Professor Phil Jones of the UEA Climatic Research Unit was asked by the BBC to respond to a series of viewer’s questions. One of them concerned global temperature increase since 1995;Do you agree that from 1995 to the present there has been no statistically-significant global warming?"Yes, but only just. I also calculated the trend for the period 1995 to 2009. This trend (0.12C per decade) is positive, but not significant at the 95% significance level. The positive trend is quite close to the significance level. Achieving statistical significance in scientific terms is much more likely for longer periods, and much less likely for shorter periods."The answer was the statistical case for warming between 1995 and 2009 could not be made, but it was a marginal statement. There was warming, but not enough to satisfy normal scientific attribution of significance.It was a quote that meant whatever one wanted it to. Some took it literally as “it hasn’t warmed at all since 1995,” which is technically true. Others said look at the statistics, no warming might be true, but it’s only just true, so it must be warming!Now after one more years’ datapoint, 2010, is available (and has been for five months) Professor Jones has looked at his statistics, using the HadCRUT database (but not specifically stated which of those that go under this name) and reached a different conclusion. Recall what he said in 2009, “Achieving statistical significance in scientific terms is much more likely for longer periods, and much less likely for shorter periods.”What a difference a year makes. This time, according to the BBC’s Environment Correspondent Richard Black’s report, warming between 1995 and 2010 is significant.Another year, according to the report, has “pushed the trend passed the threshold usually used to assess whether trends are “real” (their quotemarks).” Data from 2010, it seems, for there are no references to methodology, pushed the significance of warming beyond the 95% confidence level.“The trend over the period 1995-2009 was significant at the 90% level, but wasn't significant at the standard 95% level that people use," Professor Jones told BBC News.“Basically what's changed is one more year [of data]. That period 1995-2009 was just 15 years - and because of the uncertainty in estimating trends over short periods, an extra year has made that trend significant at the 95% level which is the traditional threshold that statisticians have used for many years.“It just shows the difficulty of achieving significance with a short time series, and that's why longer series - 20 or 30 years - would be a much better way of estimating trends and getting significance on a consistent basis.”This means that adding 6 per cent more data (16 as opposed to 15 years) has, paraphrasing Prof Jones’ own words, achieved statistical significance in scientific terms, which is much more likely for longer periods, and much less likely for shorter periods.It is perhaps understandable to revisit the question posed last year, especially given its finding.A significance of 95%, which is a 1 in 20 chance, in the real scientific world is not proof of a finding, but just a starting point. Try submitting a paper to a journal with such a level of statistical confidence and the journal’s reviewer would legitimately complain about having their time wasted until better evidence is presented.Just recall what happened recently with the claim that some particle physicists had detected a new particle, dubbed the DZero. When the results were announced they were given at the "three sigma" level of certainty. This is about a one in a thousand probability the result is due to chance, considerably higher odds than those quoted by Professor Jones. Given the importance of a potential discovery another group of particle physicists later reanalysed the data and concluded that it was actually below 5 sigma, more unlikely to be due to chance, but they still were not impressed. They rejected the claim as not proven. The initial claimants accepted the finding and have gone away to do more work.Professor Jones’ calculations may not be correct. Some have subsequently pointed out that using one version of HadCRUT shows no significant warming since 1995 whilst, depending on how you do the analysis, the other version just about does. Indeed, analysing HadCRUT with the IPCC approved method shows no significant warming between 1995 - 2010. Any prudent scientist would take this as an unsatisfactory position to advocate a positive result. Consequently it would have been nice to see some more details about Professor Jones’  methodology.In fact it is likely, given the La Nina we are experiencing in 2011 that when 2011 data is added to the calculations even by Professor Jones’ own methods the trend will once again fall below significance. Indeed, adding the first 6 months of 2011 and treating the data in half-yearly groupings shows that already.Choosing the year 1995 as a start point is in any case silliness.Looking at the post-1980 warming spell (that occurred after 40 years of not much change) there are several distinct features. Whatever one’s view on the length of time that defines climate, it is obvious and unarguable that there is structure in the temperature dataset of the past 30 years.It got warmer between 1980 – 1997. 1998 was the start of a dramatic El Nino. Subsequently there were two cooler years and post-2001, despite a few El Nino’s and La Nina’s the temperature has show no statistical change. See Warming What Warming for references. It is unequivocal that the temperature since 2001 has not changed. That is a definite feature in the temperature data. It is 2001 – 2010, not 1995 – 2010.1995 -2010 is fifteen years. It seems that some professors and commentators are happy to discuss trends in the temperature data of that duration. But when it comes to the statistically unchallengeable 2001 – 2010 standstill suddenly the time period is too short.1995 – 2010 will show a slight warming (even if not substantial statistically) because although most of the 1980 – 1997 warming had already taken place there was still a little bit of warming to go before the El Nino and the post-2001 plateau.Supporters Not ReportersThe BBC article that declared, Global warming since 1995 'now significant' is, in my view, extremely poor. The article was based on statements made directly to BBC News, (“Professor Jones told BBC News”) though possibly not directly to Mr Black.It states that Prof Jones’ 2010 comment about no warming is “still seen on blogs critical of the idea of man-made climate change.” True. But it also seen on blogs that support the idea that mankind is overwhelmingly responsible for recent climate change. It is also seen on blogs that debate the exact mix of human and natural contributions to climate change, which is an active and scientifically respectable position apparent to anyone who is up to date with the peer-reviewed scientific literature.The BBC article quotes Professor Jones as saying that adding an extra year on a 15-year dataset “shows the importance of using longer records for analysis.” The BBC article does not question that this is only an extra 6% of data. Adding a small amount of data to a dataset and getting a different result is a warning sign to any scientist to be careful. Specialist journalists should know this.The BBC article does however say that it shows the difficulty of achieving significance with a short time series. But it has clearly made its mind up. Fifteen years is not significant, 16 years is enough to achieve significance, is what the article says.The BBC article says that Professor Jones’ 2010 comment was quoted “erroneously” as demonstration that the Earth’s surface temperature is not rising. This is plain wrong. Professor Jones’ answer did include some statistical caveats but his answer, Yes or No, was Yes, it hasn’t warmed. The BBC article does not link to the specific HadCRUT data set used, but only a general website.It is a sloppy, skimpy article in the extreme. It provides little in the way of analysis and that which it does is one-sided. But even if one did not look at the accuracy of the statements it has, not for the first time, an air of triumph, as if those whom it deems skeptic (and it has a strange definition of skeptic) have been overcome. It is not impartial.In its selective coverage of climate change science BBC News has become not a reporter of climate change, but a supporter of it. It has, as this regrettable article shows, veered into advocacy. Science and Environmental journalists are often enthusiasts for the subject but as reporters they must not become cheerleaders and uncritically use shoddy science in a one-sided attempt to trounce those whom, as is obvious from this piece, the reporter thinks are wrong.There is no mention in the article that the statistics for the post-2001 temperature standstill are accepted by the scientific community. This changes the story completely.I look forward to another BBC News item, dated mid January 2012, based on data to 2011, whose headline is, Global Warming since 1995 ‘now not significant (again).’The Carbon Brief website was swift to proclaim some kind of victory by expanding the conclusions in the BBC article. This website says it "fact checks" climate stories in the media. If only it had lived up to that promise. The idea that global warming has stopped “can finally be laid to rest,” it proclaimed a short while after the BBC story was posted. It then took Professor Jones’ conclusion of a warming trend being statistically significant between 1995 -2010 and could not wait to say it discredited statements by the GWPF that there was no increase in the past ten years. In doing this the Carbon Brief demonstrated its inability to tell apples from oranges, to cherry pick data, and displayed its barely contained desire to criticise the GWPF using concocted arguments. In doing so it, in my view, destroyed any reputation is may have wanted to build as being accurate and informed.Feedback: david.whitehouse@netzerowatch.com