There is some recent discussion of Marysville and Orland at Eli Rabett and Tamino. The gist of their arguments is that the failure of NOAA to implement their promised quality control doesn’t “matter” since the problems get adjusted out in the adjustment software so that trends at Orland end up being similar to trends at Marysville. Tamino says:
Not only are the graphs strikingly similar, the trend rates are nearly identical.
One thing that you have to watch (as I’ve mentioned before) – GISS adjustments are not used in GHCN (and thus not used in CRU or NOAA NH calculations.) The relevant series for CRU is probably the GHCN adjusted series (which is usually similar to GISS raw), but since CRU won’t disclose their data, this is just a guess. Eli observes that GISS results come out in the wash about the same as CRU; why they do so is presently an unexplained Caramilk secret although the accounting would be trivial if the methods were properly disclosed.
While I looked at adjustments for these stations at the time, I’m a bit more familiar with the adjustment sequences now and thought that it would be interesting to re-visit and compare the Marysville and Orland adjustments in a similar format to the Arizona adjustments.
First here are plots of two Marysville and Orland versions – USHCN TOBS on the left and GISS adjusted on the right. In the TOBS version, we see an almost characteristic difference between two classes of site – the Orland site has a rather high 1930s, comparable in some way to Grand Canyon, while the Marysville site has higher recent values than the 1930s, similar to Tucson though not quite so dramatic. In the GISS adjusted version, the two versions are much more similar – a point observed by both Tamino and Eli. However neither investigated the adjustment sequence.
The figures below illustrate the adjustment sequences for Marysville (left) and Orland (right). Again, the patterns are somewhat similar to the adjustment sequences for Tucson relative to Grand Canyon, but with some interesting differences. GISS made a substantial UHI adjustment for Marysville, the result of which past values were increased relative to recent values. But look at Orland: here GISS has made a negative adjustment (something that I observed on a previous occasion). The Orland station seems to have been in the same location since at least 1909 – why should there be a negative UHI adjustment to Orland? This is goofy. The trends will come out similarly, but Hansen’s wrecked what seems to be good Orland data to get there.
In both cases, GISS raw versions match USHCN adjusted moderately well – so the Hansen Y2K problem observed in the Grand Canyon and some others series doesn’t come into play here.
In the bottom panel, there are USHCN adjustments to the TOBS data, which have the effect of lowering values in the 1930s relative to modern values by about half a degree C. (There is no material change in elevation of the Marysville station).
The station history for Orland does not show any moves since 1909 (with an identical elevation since 1883). Aside from looking sensible now, it has an excellent history and should be somewhat of a gold standard. So the question for me is: shouldn’t this be the sort of station that is used to benchmark nearby stations with flawed histories, rather than trying to do a whole lot of adjustments to all the stations. Here’s the Orland TOBS version (showing pre-1902 values in dotted lines since it looks like there may well be some sort of unreported discontinuity there).
Interestingly, Orland, despite its seemingly adequate rural appearance is not itself a lights=0 GISS site. California USHCN sites with GISS lights=0 are: BRAWLEY 2SW, CEDARVILLE, CUYAMACA, DEATH VALLEY, ELECTRA PH, FAIRMONT, FORT BRAGG 5N, HAPPY CAMP RS, INDEPENDENCE, LAKE SPAULDING, LEMON COVE, NEEDLES FAA AP, ORLEANS, SUSANVILLE AP, TEJON RANCHO, WILLOWS 6W and YOSEMITE PARK HEADQUARTERS. Some of these sites have been discussed previously and many are now surveyed at surfacestations.org. Lake Spaulding for example has many inappropriate aspects to its history. Why anyone would intentionally use Lake Spaulding to “fix” Orland is not at all clear to me? To the extent that this is what the GISS clergy are doing, then this sort of behavior should definitely be discouraged.