by Judith Curry
Skeptics doing what skeptics do best . . . attack skeptics. – Suyts
Last week, the mainstream media was abuzz with claims by skeptical blogger Steve Goddard that NOAA and NASA have dramatically altered the US temperature record. For examples of MSM coverage, see:
- Telegraph: The Scandal of Fiddled Global Warming Data
- Washington Times: Rigged ‘science’
- RealClearPolitics: Climate Change: Who are the real deniers?
Further, this story was carried as the lead story on Drudge for a day.
First off the block to challenge Goddard came Ronald Bailey at reason.com in an article Did NASA/NOAA Dramatically Alter U.S. Temperatures After 2000? that cites communication with Anthony Watts, who is critical of Goddard’s analysis, as well as being critical of NASA/NOAA.
Politifact chimed in with an article that assessed Goddard’s claims, based on Watt’s statements and also an analysis by Zeke Hausfather. Politifact summarized with this statement: We rate the claim Pants on Fire.
I didn’t pay much attention to this, until Politifact asked me for my opinion. I said that I hadn’t looked at it myself, but referred them to Zeke and Watts. I did tweet their Pants on Fire conclusion.
Skepticism in the technical climate blogosphere
Over at the Blackboard, Zeke Hausfather has a three-part series about Goddard’s analysis – How not to calculate temperatures (Part I, Part II, Part III). Without getting into the technical details here, the critiques relate to the topics of data dropout, data infilling/gridding, time of day adjustments, and the use of physical temperatures versus anomalies. The comments thread on Part II is very good, well worth reading.
Anthony Watts has a two-part series On denying hockey sticks, USHCN data and all that (Part 1, Part 2). The posts document Watts’ communications with Goddard, and make mostly the same technical points as Zeke. There are some good technical comments in Part 2, and Watts makes a proposal regarding the use of US reference stations.
While I haven’t dug into all this myself, the above analyses seem robust, and it seems that Goddard has made some analysis errors.
OK, acknowledging that Goddard made some analysis errors, I am still left with some uneasiness about the actual data, and why it keeps changing. For example, Jennifer Marohasy has been writing about Corrupting Australian’s temperature record.
In the midst of preparing this blog post, I received an email from Anthony Watts, suggesting that I hold off on my post since there is some breaking news. Watts pointed me to a post by Paul Homewood entitled Massive Temperature Adjustments At Luling, Texas. Excerpt:
So, I thought it might be worth looking in more detail at a few stations, to see what is going on. In Steve’s post, mentioned above, he links to the USHCN Final dataset for monthly temperatures, making the point that approx 40% of these monthly readings are “estimated”, as there is no raw data.
From this dataset, I picked the one at the top of the list, (which appears to be totally random), Station number 415429, which is Luling, Texas.
Taking last year as an example, we can see that ten of the twelve months are tagged as “E”, i.e estimated. It is understandable that a station might be a month, or even two, late in reporting, but it is not conceivable that readings from last year are late. (The other two months, Jan/Feb are marked “a”, indicating missing days).
But, the mystery thickens. Each state produces a monthly and annual State Climatological Report, which among other things includes a list of monthly mean temperatures by station. If we look at the 2013 annual report for Texas, we can see these monthly temperatures for Luling.
Where an “M” appears after the temperature, this indicates some days are missing, i.e Jan, Feb, Oct and Nov. (Detailed daily data shows just one missing day’s minimum temperature for each of these months).
Yet, according to the USHCN dataset, all ten months from March to December are “Estimated”. Why, when there is full data available?
But it gets worse. The table below compares the actual station data with what USHCN describe as “the bias-adjusted temperature”. The results are shocking.
In other words, the adjustments have added an astonishing 1.35C to the annual temperature for 2013. Note also that I have included the same figures for 1934, which show that the adjustment has reduced temperatures that year by 0.91C. So, the net effect of the adjustments between 1934 and 2013 has been to add 2.26C of warming.
Note as well, that the largest adjustments are for the estimated months of March – December. This is something that Steve Goddard has been emphasising.
It is plain that these adjustments made are not justifiable in any way. It is also clear that the number of “Estimated” measurements made are not justified either, as the real data is there, present and correct.
Watts appears in the comments, stating that he has contacted John Nielsen-Gammon (Texas State Climatologist) about this issue. Nick Stokes also appears in the comments, and one commenter finds a similar problem for another Texas station.
Homewood’s post sheds light on Goddard’s original claim regarding the data drop out (not just stations that are no longer reporting, but reporting stations that are ‘estimated’). I infer from this that there seems to be a real problem with the USHCN data set, or at least with some of the stations. Maybe it is a tempest in a teacup, but it looks like something that requires NOAA’s attention. As far as I can tell, NOAA has not responded to Goddard’s allegations. Now, with Homewood’s explanation/clarification, NOAA really needs to respond.
Sociology of the technical skeptical blogosphere
Apart from the astonishing scientific and political implications of what could be a major bug in the USHCN dataset, there are some interesting insights and lessons from this regarding the technical skeptical blogosphere.
Who do I include in the technical skeptical blogosphere? Tamino, Moyhu, Blackboard, Watts, Goddard, ClimateAudit, Jeff Id, Roman M. There are others, but the main discriminating factor is that they do data analysis, and audit the data analysis of others. Are all of these ‘skeptics’ in the political sense? No – Tamino and Moyhu definitely run warm, with Blackboard and a few others running lukewarm. Of these, Goddard is the most skeptical of AGW. There is most definitely no tribalism among this group.
In responding to Goddard’s post, Zeke, Nick Stokes (Moyhu) and Watts may have missed the real story. They focused on their previous criticism of Goddard and missed his main point. Further, I think there was an element of ‘boy who cried wolf’ – Goddard has been wrong before, and the comments at Goddard’s blog can be pretty crackpotty. However, the main point is that this group is rapidly self-correcting – the self-correcting function in the skeptical technical blogosphere seems to be more effective (and certainly faster) than for establishment climate science.
There’s another issue here and that is one of communication. Why was Goddard’s original post unconvincing to this group, whereas Homewood’s post seems to be convincing? Apart from ‘crying wolf’ issue, Goddard focused on the message that the real warming was much less than portrayed by the NOAA data set (caught the attention of the mainstream media), whereas Homewood more carefully documented the actual problem with the data set.
I’ve been in email communications with Watts through much of Friday, and he’s been pursuing the issue along with Zeke and help from Neilsen-Gammon to NCDC directly, who is reportedly taking it seriously. Not only does Watts plan to issue a statement on how he missed Goddard’s original issue, he says that additional problems have been discovered and that NOAA/NCDC will be issuing some sort of statement, possibly also a correction, next week. (Watts has approved me making this statement).
“I mean, I’m not a scientist either, but I’ve got this guy, John Holdren, he’s a scientist,” Obama added to laughter. “I’ve got a bunch of scientists at NASA and I’ve got a bunch of scientists at EPA.”
Who all rely on the data prepared by his bunch of scientists at NOAA.
How to analyze the imperfect and heterogeneous surface temperature data is not straightforward – there are numerous ways to skin this cat, and the cat still seems to have some skin left. I like the Berkeley Earth methods, but I am not convinced that their confidence interval/uncertainty estimates are adequate.
Stay tuned, I think this one bears watching.
Update: Watts has a new post Scientific method is at work on the USHCN temperature data set