The GISS homepage formerly said:
The NASA GISS Surface Temperature Analysis (GISTEMP) provides a measure of the changing global surface temperature with monthly resolution for the period since 1880, when a reasonably global distribution of meteorological stations was established. Input data for the analysis, collected by many national meteorological services around the world, is the unadjusted data of the Global Historical Climatology Network (Peterson and Vose, 1997 and 1998) except that the USHCN station records included were replaced by a later corrected version.
In econometrics, you couldn’t use loosy-goosy phrases like “replaced by a later corrected version.” You’d have to identify the version. Actually AGU policies (which apply to JGR) require proper data citation, although climate scientists publishing in AGU journals flout this policy, including Hansen here. Following an email to Hansen and Ruedy of GISS, they changed the introduction at the GISTEMP intro from:to the following:
Input data for the analysis, collected by many national meteorological services around the world, is the unadjusted data of the Global Historical Climatology Network (Peterson and Vose, 1997 and 1998) except that the USHCN station records up to 1999 were replaced by a version of USHCN data with further corrections after an adjustment computed by comparing the common 1990-1999 period of the two data sets. (We wish to thank Stephen McIntyre for bringing to our attention that such an adjustment is necessary to prevent creating an artificial jump in year 2000.)
This doesn’t really clarify the provenance of the data. In his email, Ruedy added:
In 2000, USHCN provided us with a file with corrections not contained in the GHCN data. Unlike the GHCN data, that product is not kept current on a regular basis. Hence we used (as you noticed) the GHCN data to extend those data in our further updates (2000-present)
Well, I hadn’t really “noticed” that they had used GHCN data to extend the USHCN data. I’ve done lots of cross-comparisons with different variations trying to identify exactly where the GHCN raw data came from. I’ve put up many plots at CA showing GISS raw as compared to different USHCN versions and I’ve done many more that I’ve not posted up. At his point, there’s one thing that we can say for sure: I’ve now looked at GISS raw as compared to GHCN raw data in the post-2000 period and in the few sites that I’ve examined since receiving this email (Detroit Lakes, Port Angeles), there is an exact match. So at least we’ve tracked down one aspect of the provenance of GISS raw data. I’ve done a comparison plot below for Detroit Lakes MN, a series that we’ve looked at before. After 2000, the match is exact (the delta is 0.0). However, before 2000, the USHCN TOBS/adjusted series sort of match for a while but the match tails off in the earlier portions, with the GISS raw version being quite different.
So where does the GISS raw version come from? At this stage, I think that we can declare that it doesn’t come from any of the USHCN version 2 (raw, TOBS, adjsted). If it did, then the versions would match as exactly as the GHCN raw and GISS raw after 2000. So where does it come from? Another stupid climate science guessing game, although hopefully it will solved in a shorter time than the unsolved MBH99 confidence intervals.
Ruedy’s letter has opened up the possibility of really obsolete data being used – I hadn’t thought to look at really obsolete data. I browsed through some obsolete data in connection with Swindle – who would have thought that we’d be doing so again. Two possibilities spring to mind: (1) maybe he’s using something from USHCN version 1; this is online, but I don’t think that the dates are right. (2) Maybe there’s an earlier USHCN version 2 that’s been overwritten (3) maybe he’s using an old GHCN version before 2000. Maybe none of the above – hey, it’s climate science.