Beneath the seemingly placid world of U.S. weather co-operatives, recent analysts have found a turbulent world of changing observation times, with regime change after regime change. Hansen and Karl have been forced – unwillingly, I’m sure – to adjust past temperatures downwards. In contrast to the seemingly almost Iraqian chaos of the U.S. weather observation network, Jones et al 1990 reported that islands of homogeneous measurement could be found in Russia and China, where dutiful comrades seemingly set aside minor worldly concerns, like revolutions and famines, and homogeneously attended to their max-min thermometers creating, in the process, what Peterson 2003 described as two of the few studies that used “homogeneous data”.
When David Lean made the movie Dr Zhivago – a blockbuster in my teenage years with Julie Christie then a famous beauty – he must have forgotten the scenes where Omar Sharif and Julie Christie, just before fleeing, instructed arriving soldiers on how to transcribe max-min thermometer readings, with time suspended as in a Roadrunner cartoon so that the soldiers could take on these duties before Julie and Omar fled. In the tranquil world of Jones et al 1990, such scenes must have been repeated time and time again throughout the gulags.
Jones et al 1990 did not do anything as mundane as identifying the sites in their Russian “rural” network (and have thus far resisted efforts to identify the sites), but Jones et al 1990 does have a chicken-scratch map. Warwick Hughes has enlarged the chicken-scratch map and begun the process of trying to identify the 38 sites with some partial success. His analysis is here – see, for example the map here. Some sites can be plausibly identified. Malye Karmakuly on Novaya Zemlya is one such site, about which I’ll make some comments today.
Let’s re-visit the quality control for the Russian network that Jones described as follows:
All the site records were assessed for artefacts due to factors such as site moves or changing methods used to calculate monthly mean temperatures. At twelve stations the observing station was moved slightly. Comparisons with neighboring sites were made before and after each change and, where necessary, corrections were made to ensure homogeneity of the rural-station record. No corrections were deemed necessary for the remaining 26 stations where no station moves were reported.
So Jones has specifically verified the integrity of each of the 38 stations in the Russian network. Let’s look at the first site that Warwick identified – the lat-long’s of the most northerly station in the Jones Figure are approximately 72.37N, 52.7E, which matches WMO station 207440 (Malye Karmakuly) on Novaya Zemlya. There are no other nearby candidates in any listing of stations. This station is listed in both GHCN v1 in 1992 (NDP041) and in the current GHCN v2 (updated through Feb 2007). So I think that we can safely proceed on the assumption that this temperature series is one of 38 stations in Jones’ very homogeneous Russian network and use this to spot-check the degree of homogeneity.
I extracted data for WMO station 20744 from both GHCN v1 (NDP041) and GHCN v2. The first figure is a plot of the information in GHCN v1 – the version contemporary with Jones et al 1990.
Petty minds may think that they see gaps in this graph, suggesting that the record may not be totally homogeneous. A report on NDP041 here states that over 40% of the data from this site is missing. Sure enough when one examines the raw data, although the data goes to 1990, one observes gaps as follows:
from Oct 1900 to Aug 1901;
from Aug 1910 to Dec 1912;
only odds and ends from Nov 1916 to Sep 1921
none from July 1940 to Jul 1942
none from Jan 1951 to Dec 1960;
none from Apr 1963 to Dec 1980
The corresponding information from GHCN v2 is shown in the next figure, with more continuity. Two separate identifiers are noted and are colored black and red respectively. The values from GHCN v1 are marked with blue circles and match closely to GHCN v2 values where they overlap. While there are additional values in this data set, like GHCN v1, the values do not extend further to the present than 1990.
Let’s think about the homogeneity of this record and whether this provides evidence that the Russian network achieved a homogeneity that eluded observers in the turbulent United States. In both GHCN v1 and GHCN v2, there are notable interruptions from Nov 1916 to Sep 1921 and from July 1940 to July 1942. One can surmise that the comrades responsible for the max-min thermometers in these periods may have given priority to other things. Astonishing as this possibility may seem, the shameful truth is that the responsible comrade seems to abandoned his thermometer and failed to preserve the homogeneity of the record. So how did Jones establish the homogeneity of the record before and after Nov 1916-Sep 1921 and July 1940-July 1942? At this remove, it seems hard to understand Jones’ confidence. But the ways of the Team are deep and mysterious.
The period from 1951-1980 even has its own interest. In GHCN v1 (1992), 27 of 30 years were missing in GHCN v1 (to which Jones has sent information on 1872 stations.) GHCN v2 has second record which fills in most of the missing years. However, note that there are a couple of years in which the two versions overlap and the red record is higher than the black record in both years. What adjustments did Jones make, if any? Another Caramilk secret.
Was the information in GHCN v2 available to Jones et al 1990 but not to GHCN? The mind boggles at the thought. Novaya Zemlya was home to the Russian nuclear program. Maybe local temperatures in Novaya Zemlya were classified. Now we know that the Team has access to internationally sensitive high-security data – such as the password-protected tree ring data at the CRU website – long sought after by international agents. Did Jones have access to high-clearance Malye Karmakuly max-min thermometer readings? Ah, intrigue everywhere.
But leaving aside the Team’s access to classified information, does this record show evidence that dutiful comrades were making homogeneous temperature records that observers in the turbulent United States Co-operative Network were unable to achieve? Not to me. And when you think about it, the idea of looking to Russian and Chinese records as havens of homogeneous temperature measurement, unavailable elsewhere, is completely preposterous. Novaya Zemlya certainly doesn’t support this idea.