Re-visiting Dawson, Canada

In May 2007, we took a look at station data for Dawson, Canada, data that is important to dendros because it goes back to the gold rush days of the late 1890s and because there are important tree ring chronologies in the area. In this case, the dendros decided that the GHCN adjusted data was unreliable and used the Canadian adjustments. At the time, Rob Wilson gave encouragement to the analysis of station data as it was obviously frustrating to dendros to have to sort through station versions and to be put in a position where they had to pick one temperature history version rather than another – not something that a dendro should really have to do. Looking back on this post, I can see definite progress in our ability to sift through the various adjusted versions.

In the analysis of the important site of Twisted Tree Heartrot Hill, Yukon. Davi (D’Arrigo) et al had noted a major Divergence Problem. Wilson, re-visiting this analysis (see prior post for refs) suggested that part of the divergence might occur because of inaccuracies in the GHCN temperature history. He showed the figure shown below comparing the GHCN and Canadian versions, hypothesizing that part of the divergence that bothered D’Arrigo might be due to inaccuracies in the GHCN history.

Original Caption: Fig. 2. Time series plots of mean July-“August temperatures for the Dawson meteorological station, Yukon Territory. The data were taken from the homogeneity corrected Global Historical Climate Network (GHCN, Peterson and Vose, 1997) and Historical Canadian Climate Network (HCCN, Vincent and Gullett, 1999) data-sets. The lines denote the linear trend of each record over the 1900-2000 period.

The next figure replicates the Canadian and GHCN versions and adds in information on 3 other versions, adding some interesting perspective to several issues.

The GHCN version in the Wilson diagram is the GHCN adjusted version, which goes to 1989 at GHCN. I’ve downloaded Canadian data from the Canadian HCN website, which goes at present to 2004. (You can download today’s data from Dawson on the internet BTW.) I’ve shown data up to 2000 (as in Wilson) in a thick red line and the extension in a thin line.

I’ve also added in the GHCN raw version which coincides with the GISS Raw and GISS Adjusted version, all shown in green. All 3 versions end in 1989 and overlap the other versions towards the end.

Figure 2. Dawson July-August temperatures.

A few comments.

First, it’s disgusting that GHCN and GISS have not found it within their power to locate temperature data for Dawson more recent than 1989. These organizations have professional staffs. Updated information is readily available on the internet. How hard can it be to locate Canadian data? Maybe it’s time to ask the people who do the Consumer Price Index to compile temperature statistics. It’s all just data – maybe professional data people would do a better job than the present people who seem to have trouble getting off their LaZBoys. I’m not sure that academics who want to write little papers are logical people to be doing mundane tasks like keeping temperature data current. Obviously Hansen’s more interested in “big picture” issues than in collecting temperature data from Dawson, Canada. That’s OK. But if he (and his GHCN equivalent) aren’t going to collect the data from Dawson and other places, get out of the way and let someone do the job who will do it properly.

Second, we can learn something here about GISS adjustments. GISS doesn’t adjust the Dawson data, but GHCN does. In the GISS file, we see the following for Dawson:

719660000 DAWSON,Y.T. lat,lon (.1deg) 641 -1391 R A cc=403 0

So while it is in Canada, it has no lights code attached to it and is treated as a ROW R-site. My current understanding is that sites without a lights code are governed by their R/S/U code and Dawson therefore gets included in the Unlit pool. But that’s just a surmise right now. In any even, it isn’t adjusted in the GISS adjustment stage.

Third, we learn something about the Canadian HCN network. It looks like the GHCN has preserved a fossil record of Dawson data and that the Canadian HCN has adjusted the Dawson record. The Canadian adjustment has the effect of reducing Dawson temperatures in the early part of the 20th century.


The Canadian adjustment is far less extreme than the GHCN adjustment, which reduces summer temperatures in the early 20th century more than 3 deg C in some cases, as shown below.


Fourth, CRU refuses to archive or provide the data sets that they use. As we’ve see, there are a bewildering number of different versions.

At present, I take no view on which of these adjusted versions provides the more correct understanding of temperature in Dawson. However, we do have 3 “adjusted” versions from 3 different sets of experts, which have a range of 3 deg C in their estimates of summer temperature in Dawson, Yukon, a town that has been continuously settled since the late 1890s.


  1. David Jay
    Posted Feb 23, 2008 at 10:32 AM | Permalink

    As a Non-Climate-Scientist, it appears that current temperatures aren’t warming but rather historic temperatures are cooling!

    That’s one way to create trend lines…

  2. Posted Feb 23, 2008 at 11:21 AM | Permalink

    so temping, why not?

    Will Steve find even more?

    To find out, join us for the next episode of Muppet lab, with doctor hansen hunnydew, where the past is being made,…today!

  3. Michael
    Posted Feb 23, 2008 at 11:34 AM | Permalink

    Let’s outsource all GISS and GHCN functions to the Weather Underground at . Unlike the House of Hanson, they appear to have no difficulty in locating Dawson Canada and providing almost real-time weather updates .

    The Weather Underground has Dawson histories back to June 30, 1996 . The Weather Underground is always very careful to add disclaimers such as “Averages and records for this station are not official NWS values”.

    I have two questions for Steve:

    1. How accurate is the Weather Underground data?
    2. Why are we giving taxpayer monies to bozos when the Underground is making a profit with the data and provides a high quality free service?


  4. Lars Kamél
    Posted Feb 23, 2008 at 11:49 AM | Permalink

    Vincent and Gullett have published papers about their correction methods and I think their ideas look good. If I had to make a choice, I would prefer their corrected version of any weather station over versions by GISS, GHCN or CRU.

  5. bender
    Posted Feb 23, 2008 at 10:58 PM | Permalink

    I believe that widespread availability of Canadian historical climate data is a relatively recent development. That would mean they’ve been sitting in their lazyboys for only ~8 years, not 18.

  6. Kenneth Fritsch
    Posted Feb 24, 2008 at 12:37 PM | Permalink

    At present, I take no view on which of these adjusted versions provides the more correct understanding of temperature in Dawson. However, we do have 3 “adjusted” versions from 3 different sets of experts, which have a range of 3 deg C in their estimates of summer temperature in Dawson, Yukon, a town that has been continuously settled since the late 1890s.

    More variations for data mining cannot be a good thing. Such a large difference in data sets has to make tracking down cause less difficult.

    This instance also points to why accurate historical local temperatures can be just, if not more, important than the illusive national and global averages on which the data set keepers seem to concentrate.

One Trackback

  1. By The End of CRUTEM? « Climate Audit on Jan 30, 2010 at 4:24 PM

    […] again in a couple of 2008 posts here here noting (in this case GISS’) inability to locate Canadian station data: How hard can it […]

%d bloggers like this: