Yesterday I surfed to the Datasets and Images page on GISS, something I ordinarily do only around the fifth of each month as I wait for the latest monthly temperature updates from GISS. The information on that page rarely changes, so I was surprised to see the following addition:
June 9, 2008: Effective June 9, 2008, our analysis moved from a 15-year-old machine (soon to be decommissioned) to a newer machine; this will affect some results, though insignificantly. Some sorting routines were modified to minimize such machine dependence in the future. In addition, a typo was discovered and corrected in the program that dealt with a potential discontinuity in the Lihue station record. Finally, some errors were noticed on http://www.antarctica.ac.uk/met/READER/temperature.html (set of stations not included in Met READER) that were not present before 8/2007. We replaced those outliers with the originally reported values. Those two changes had about the same impact on the results than switching machines (in each case the 1880-2007 change was affected by 0.002°C).
Now, the thing that caught my eye was the “June 9, 2008″ in bold. Perhaps I am reading more into the use of a specific date and boldface type than I should, but I interpret it to mean the website entry was made on June 9. However, I am 99% certain that entry was not on the GISS website on July 5, which is the date I started looking for the July update (after that date I just did a reload of the tab holding the GLB.Ts+dSST.txt file, so I would not have seen updates to the parent page after July 5). I say 99% because I am almost certain I would notice the additional entry as I scrolled down to the GLB.Ts+dSST.txt link, but I cannot say I am 100% certain. We will need to wait a number of months before the internet archives updates the history for that page.
[Update] I found a cached version of http://data.giss.nasa.gov/gistemp/updates/200806.html dated June 20 off of Live search. This page is linked directly from and further explains the June 9 update. Thus, I would conclude that the update was made at least as early as June 20. The remaining search engines all have this page and the parent page cached on July 11 or later. [End Update]
Why is this important? For those of you who followed the GISS Step 2 thread on this site, you will recall that on June 19 Steve and I were commenting on a mysterious seventh version of the Chatham Island record. I noticed that the seventh version came from a UK site: http://www.antarctica.ac.uk/met/READER/temperature.html (these days, Google will find almost anything). Furthermore I noted:
Here is where it gets interesting. Looking at the Chatham record on the UK site, it still contains the errors described by Steve in comment #61. The version on the GISS website contains the corrections / changes Steve noted.
Steve dryly responded:
So Hansen and his boys are manually editing data. My, my. This GISTEMP thing really is a pig.
We bumped into this by accident a full ten days after the changes were made, and it seems another 19 or so days were required to post the update notice on the GISS website. Right now, if one looks at the bottom of the GISS page, one will find:
I’d love to look at earlier versions of http://www.antarctica.ac.uk/met/READER/temperature.html. Unfortunately, they are not archived by the wayback machine. I am surprised changes were made after 8/2007, because the file ends with 2006 and the actual Chatham Islands data ends with one entry in 1998. The next latest entry is 1992. Why would the British Antarctic survey change that one dataset during the summer of 2007?