The Canada Y2K Bug

I have been analyzing the Canadian temperature record, and it is a farce. There are less than 25 stations in Canada with a relatively complete long term daily temperature record .

The average maximum temperature for all Canadian GHCN stations has plummeted since the 19th century, but the trend is meaningless because the average latitude of the stations has moved four degrees further from the equator during that time period.


If we look at only the stations which were active in both 1918 and 2017, they show a big jump in temperature at the year 2000. But there are some problems which make the trend meaningless.

First (huge) problem is that there are only about 20 GHCN stations in Canada which have reported consistently for the past century, compared with more than 700 in the US.

But here is the real killer.  The jump in temperatures after 2000 is largely due to something going on at one station – Banff. Something went seriously wrong with that station after the year 2000, where almost all of the low maximum temperatures disappeared.

Without Banff, 2010 is barely warmer than 1940.

The yellow pin is at the latitude and longitude where GHCN lists the thermometer. It is almost like the moved the thermometer indoors.

You really can’t make this stuff up. Like just about every other country outside the US, the Canadian temperature record is a farce. I’ve developed some pretty sophisticated tools for detecting these problems, and I doubt that the people at Berkeley Earth, NASA and NOAA are even aware of how awful the global temperature record is.

The number of stations with long term records in Canada and their geographical location is much too small to draw any meaningful conclusions about Canadian temperature trends. You can get whatever trend you want by picking the right set of stations.

This entry was posted in Uncategorized. Bookmark the permalink.

28 Responses to The Canada Y2K Bug

  1. Alan Tomalty says:

    Tony That is why the only important study on Canada’s temperatures split the study into 2. One part for the period 1900-1998 for southern Canada below 60th parallel and the 2nd part for the years 1950- 1998 for the entire country. I have critiqued this study as per below.

    The most important climate study ever done for Canada. It finally got published online in 2010 even though the study was completed and submitted in September 1999 and published in the Journal Atmosphere and Ocean so was obviously peer reviewed. I doubt whether it would get past a proper peer review (well it still would get past pal review) today. Countless other climate reports (whenever talking about Canadian climate) afterwards are all citing this report.

    What follows are specific criticisms of the above report. Pay particular attention to the last criticism no. 13.

    “From 1900–1998, the annual mean temperature has increased between 0.5C and 1.5C in the south.”

    1) This is the 1st conclusion in the abstract and it is shown to be an alarmist statement because in the actual report on page 11, they say
    “There is a statistically significant positive trend, which accounts for an increase of 0.9 C, for the region during the period.”

    Anybody that just reads the abstract will get a misleading picture.

    2) The authors measure 6 variables maximum, minimum and mean temperatures; diurnal temperature range; precipitation totals and the ratio
    of snowfall to total precipitation; with 3 types; temperature,rainfall, and snowfall. They discuss the ratio of snowfall to total precipitation and in the study the authors state that the issue is complex. In reality they don’t understand the science of snowfall to total precipitation and its importance. Therefore the authors should not have included it in the report.

    3) The authors were not satisfied with measuring only trends of the above variables but attempted to measure extreme climate conditions. They were following in the footsteps of many others before whereby “extreme” means precipitation or temperature below the 10th or above the 90th percentile (approximately 2 sigma) of the relevant time series data. However they still werent satisfied with that so they invented the term “abnormal climate”. This is precipitation or temperature below the 34th or above the 66th percentile (approximately 1 sigma) of the relevant time series data.

    I conjecture that this has led to some of the claims we see on the news today of abnormal weather events. It certainly categorizes much more weather than before as abnormal and when applied to climate studies it makes it seem the majority of weather is abnormal, if the population curve is Gaussian.

    4) On page 19 of the study they say “The annual mean temperature anomalies relative to the 1961–1990 mean for the whole of Canada (1950–1998) are also shown in Fig. 3. No significant linear trend is detected for the entire country during the period.”

    Why wasn’t this statement in the abstract? By omitting this statement the authors have shown their bias toward the AGW crowd.

    5) It is when we get to page 3 that we find out the real reason for doing this study.

    “As well, the anthropogenic climate change signal is projected to be stronger in the high-latitudes (Nicholls et al., 1996). This suggests that it might be easier to detect climate change in a country like Canada.”

    However the next paragraph puts a minimum condition that has to be met to achieve the above.
    “The detection of climatic trends, including those predicted to occur from rising
    concentrations of atmospheric greenhouse gases (Wigley and Barnett, 1990), may
    be sought in historical climate records providing that such records are representative
    and cover a long enough period of time (typically more than 100 years).”

    They had to break up the report into 2 time periods 1900- 1998 and 1950-1998. That is because they didnt have enough data for the above 60th parallel in the far north until 1950. They did not state how many weather sations are part of the far north since 1950. They only state the total number for Canada. 210 temperature stations and 489 precipitation stations. They found a 0.9 C difference( annual mean temperature) for the whole century 1900-1998 which covers Southern Canada because of the reason explained above. Since some of that was before 1950 it may explain why, the authors don’t break down the differences in temperatures between the 2 periods 1900-1949 and 1950-1998 except in Table 1 which gives gridded area % based on the abnormal data and in Figure 3 which graphs the Departures from the 1961–1990 mean of area average mean temperature(C).

    So, for the one period 1900-1998 (Southern Canada only) that meets the above minimum condition for attempting to prove global warming (at least from a Canadian perspective) ; they failed to even look at the differences except by the graph in Figure 3. What they do say is the following:

    “The linear trend is not exactly monotonic. The rises of temperature prior to the 1940s and after the 1970s account for the significant trend. There is a modest decrease during 1940–1970. ”

    So it seems they gave up on their non-stated null hypothesis process. They actually produced this report and then ignored the real reason for writing the report!!!!!!!!!!!!!!!!!!

    6) As for precipitation data I quote;

    “The station data were gridded using a procedure developed by Hogg et al. (1997).
    The procedure uses statistical optimal interpolation and employs climatology as a
    “first guess” field and interpolating only the departures from the climatology of the
    relevant field. ”

    When you have to torture your data to this extent, it calls into question the relevance and accuracy.

    7) The authors state on page 4 “Karl et al. (1996) developed and analyzed a climate extremes index
    (CEI) and a greenhouse climate response index (GCRI) for the United States. They
    found a positive trend in the U.S. GCRI during the 20th century that is consistent
    with projections resulting from increased emissions of greenhouse gases.”

    This is very curious because it was around that time that James Hansen had said there was no warming in the 20th century for the US.

    8) It seems that Canadian data is also subject to adjustment upwards as witnessed by the above quote on page 5.

    “A significant improvement over previous versions of the temperature database is the adjustment
    in the minimum temperature to account for changes in observing procedure at principal stations in 1961 (Vincent and Gullet, 1999). Adjustment for this bias results in a less pronounced cooling in minimum temperature in terms of magnitude (by as much as 0.5 C) ”

    9) On page 9 they attempt to parameterize an equation that takes the serial correlation into account. However they assume the trend is linear.
    Since climate has seen many sine curves in the data over the years, this assumption is not valid.

    10) On page 12 we find the source for the alarmist claim of 1.5C increase in the abstract.

    “The greatest warming, which is in the Prairies, is about 1.5 C over the 99-yr period.”

    11) No account is taken for Urban island heat effects.

    12) On page 14 of the report the authors state:
    “Most of the decrease in the diurnal temperature range (DTR) occurred prior to the 1950s, especially late in the first half of the 20th century,
    coinciding with an increase in total cloud amount in Canadian mid-latitudes during
    the first half of the 20th century (Henderson-Sellers, 1989; McGuffie and Hender-
    son-Sellers, 1988). Henderson-Sellers (1989) did not propose any specific reasons
    for the cloud increase. Significant decrease in the DTR did not occur in the second
    half of the century when the greatest increase in greenhouse gases took place. This
    suggests that trends in the DTR are closely related to changes in total cloud amount.
    The trends in both DTR and total cloud cover differ from one season to the other.
    Future investigation into the relationships between the changes in DTR and cloud cover is needed”

    Since this is only conjecture , the authors should have left out any discussion of DTR. They admit it has nothing to do with global warming.

    13) On page 25 of the report the authors state :

    “The annual mean temperature for Canada has increased by 0.3C over the last 49 years(1950-1998), but this increasing trend was not statistically significant.”

    However there are at least 4 parts of the report where the authors profess their belief in global warming.


  2. Alan Tomalty says:

    Tony Are there any other station problems with Canada? Can you throw the Banff data out and redo the analysis only for years 1950 to 2018?

    • Alan Tomalty says:

      The reason for starting in 1950 as the report makes clear is that Canada added stations in the North around that time.

  3. JCalvertN(UK) says:

    It seems a very unlikely location. I found another place in the GHCN website which gave the position as “51.1833°, -115.5667° ” – which is about 20m NE in the back garden of a somewhat ramshackle wooden house with tall coniferous trees around it. (434 Cougar St.)

    If that really is the location (and I not convinced about that – does NOAA employ people with dyslexia?) then it could have been profoundly affected by three large nearby buildings of apparently recent construction: a) Best Western Siding 29 Lodge; b) a large condo with underground parking immediately west of the Best Western; c) large apartment building called “Caribou Creek 440 Cougar Street” which has a large asphalt carpark at the rear.

    • R Shearer says:

      Could it possibly be the towel rack near the Best Western’s hot tub?

      • JCalvertN(UK) says:

        The weather station moved out of there round about 1997, I think.
        Based in metadata from Env Canada:
        BANFF lat = 51.18 lon = -115.57 (from 1887-01 to 1995-03)
        BANFF (AUT) lat = 51.18 lon = -115.57 (from 01/01/1997 to 31/12/1997)
        BANFF CS lat = 51.19 lon = -115.55 (from 1995-04 to 2007-11)

  4. JCalvertN(UK) says:

    Found it! [ -) ],-115.5717199,2a,44.2y,286.25h,87.96t/data=!3m6!1e1!3m4!1sKWUy6FCwSHVrDf0tDhNSHg!2e0!7i13312!8i6656

    It is literally indoors at the Banff Park Museum National Historic Site
    91 Banff Ave, Banff, AB T1L 1K2, Canada 51.173741, -115.571675

  5. JCalvertN(UK) says:

    There is some kind of fenced compound with huts and instrumentation at
    51.193429, -115.552562

    • R Shearer says:

      Man, you’re good at finding those things!

      From the street view, it does appear that there is a weather screen there. The one in the museum is more recognizable though.,-115.5529817,3a,15y,28.9h,88.69t/data=!3m6!1e1!3m4!1sHl50jMLZabWQSE_FTOAtUg!2e0!7i13312!8i6656

      • JCalvertN(UK) says:

        Yep. That’s the new station all right.
        It’s called “Banff CS”
        Lat = 51 deg, 11 min, 36.090 sec.
        Long = -115 deg, 33 min, 08.05 sec.
        Elev = 1396.9 m (compared with 1383.7m for the old station)
        The difference in minimum temperatures might be down to the change to an automatic station – with the all-to-common issue of proximity to airconditioned huts.
        It might be down to microclimates. A move of less than a mile could mean the new station is in a completely different microclimate. Banff is overshadowed by some mighty impressive mountains and I would have thought this must exacerbate huge differences in microclimate The old site could have lain within a ‘cold pool’ in the shadow of Tunnel Mountain – as evidenced by the nearby forested slopes to the south. The new location is grassy – suggesting a sunlit meadow on a south-facing slope.

        • Steven Fraser says:

          If the station was relocated, there is another reason. The new location has access to some heat sources.

          Here is an aerial view:

          On my pic below, the weather station is circled.

          Diagonally across the intersection is a developed area. I did not capture all the buildings, but there are paved parking lots. There are four roads which bracket the area: Trans Canada highway (just off the bottom of my pic), Banff Avenue (off the upper left) and Compound Road that connects them. There is also an extension of Hawk Avenue which serves as the parking lot for the weather station.

          Not too far away are the resorts.

          So, on a Winter night, with still air, it would be very possible to have an inversion form, trapping some of the heat in the valley, and raising the low temp reading.

          • Steven Fraser says:

            Sorry, but I forgot to mention the RR tracks just to the left of the circled weather station.

          • Gator says:

            Do we know when it was changed to MMTS?

          • JCalvertN(UK) says:

            Yes. I saw that. I was wondering how much of an effect it would have. A breeze from the west would affect it.
            I bet that whole town regularly gets covered in a foggy stream of cold air flowing down the valley towards Calgary. But it would need a local to know whether it strictly follows the river round the SE of Tunnel Mountain, or if it overflows along the line of ‘Banff Avenue’ around the NW side.

  6. GW Smith says:

    But, more importantly, has it been adjusted?

  7. Josh says:

    Do Russia and China please. I want to post the graphs so the leftists claim its denying AGW is a conspiracy by the Russians and Chinese :)

  8. JCalvertN(UK) says:

    Environment Canada website has a nice set of weather data up to 1995.
    Here . . .
    (See graph attached)
    After 1995 however, I don’t know what happened. They seem to have converted to Automatic weather station. But most of the data is lost in cyberspace.

  9. JCalvertN(UK) says:

    FWIW I have managed to extract full sets of Tmax and Tmin data. See Tmin below.

  10. JCalvertN(UK) says:

    Further to . . See Tmax below.

  11. JCalvertN(UK) says:

    And an aerial photo post 1984 / circa 1990s. Below is marked-up cropped view of the Cougar St /Marten St area round the purported old weather station.

  12. JCalvertN(UK) says:

    I have now digitised the GISS graph for 2016 and compared it with their current graph – see below. In the few months since May 2016, GISS have adjusted most of Banff’s historical temperatures downwards by 1 degree Celsius.

    “Now there’s your problem . . .”

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.