BOM vs GISS- Who’s Right?

Ken Stewart,  December 2010

Thanks to Chris Gillham who first noticed this oddity and has posted at

The Goddard Institute of Space Studies (GISS) produce global temperature records with raw data sourced from raw GHCN data with USHCN corrections, which in turn comes from meteorological agencies all over the world. In Australia, the Bureau of Meteorology (BOM) supplies the data from selected sites. As it is important that the data is as accurate as possible, good quality sites are preferred. In Australia, there are two sets of preferred sites for climate analysis: the High Quality (HQ) series and the Reference Climate Station (RCS) sites. These two sets overlap – there are many sites in both.

As BOM explain: “Operational monitoring of Australia’s changing climate is based on (the) high-quality data series.”


… and …

“The Australian Reference Climate Station (RCS) network has been established for high quality, long-term climate monitoring, particularly with regard to climate change analysis. The establishment of the network followed a request by the World Meteorological Organisation to all of its member nations in 1990.”        See


The HQ sites are used to produce BOM’s trend maps and time series graphs, and the CSIRO uses RCS data to produce the State of the Climate reports. The primary purpose of RCS is to provide a set of timeseries from which data can be interpolated/ extrapolated to check and/or fill in missing data when gridding. Therefore data from these sites are critical.

Chris Gillham has analysed the records from the HQ sites which are still being used by GISS in 2010. A large number of sites, mainly rural, were dropped after 1992. Most of these have very long temperature records. It seems that all of the sites he has looked at except Launceston Airport, Perth Airport and Brisbane Airport are RCS sites. It appears that GISS get their data from a selection of RCS sites, major airports, and one other site- Eucla.

In this post I will look at the remaining RCS sites.

I have found exactly the same phenomenon as Chris found in his analysis. The GISS annual data matches the raw BOM data almost exactly in most years before 1995, and monthly data is exactly the same in 2010 (there are slight differences as GISS calculates annual data from December to November, not calendar years). Therefore we can say with confidence that the GHCN raw data which GISS uses is drawn from BOM raw data.

When I examined the data, I found that between 1995 and 2009 for the majority of sites the GISS raw data, which should be the same as the BOM data, is significantly less (29 sites significantly less than BOM,  2 higher and 14 similar). It is up to 1 degree Celsius lower, but typically about 0.3 to 0.6 lower.

This is very odd. The result is that trends calculated from BOM’s raw data for the CSIRO climate analyses are much greater than those from GISS. However, for many sites, the RCS data is far too short to be of any use!

It’s an embarrassment. They can’t both be right.

There must be a logical explanation for this. Perhaps there has been a system wide problem with download/ upload of data, or with interpretation of site information (latitude/ longitude).

Whatever the reason, the 1990s / 2000s divergence between BOM and GISS means must be investigated and corrected.

In summary:

Sites with 1993- 2006 data missing:

Cape Nelson Lighthouse, Nowra RAN, Winton, Maree, Forrest (92- 2000), Christmas Island, Oodnadatta, Cape Bruny Lighthouse, Tarcoola, Robe, Wilsons Promontory, Cape Otway Lighthouse, Yamba Pilot Station, Richmond RAAF, Moruya Heads Pilot Station, Richmond PO, Amberley RAAF, Cape Leeuwin, Melbourne Air (1993-99).

Following are plots of the RCS sites. For the remainder, see

First, the only site in all of Australia recently used by GISS that is neither HQ nor RCS: Eucla in Western Australia.

It seems likely that Eucla was once considered as a possible HQ/RCS site but was subsequently not included. However, its feed to GISS continued from 1997, albeit with gaps in the yearly record.

Note: two sites with GISS data above BOM:


Lord Howe Island:

Note that Antarctic sites do not show this pattern – only mainland and offshore islands.

Macquarie Island:




Concerning Antarctica: there are certainly few indicators of global warming (rising temperatures, mainly minima, mainly in winter and greater in polar regions) at Australia’s Antarctic bases.

• Rising temperatures? While the temperatures at Davis and Casey are rising at about 1.2C and 1.0C respectively per 100 years, this is no worse than many Australian sites; Macquarie Island’s trend is about 0.85C, peaking in the 1980s and falling since, and Mawson is cooling at -0.2C.

• Mainly minima, mainly in winter? The fastest warming site, Davis, is warming less in winter (0.3C) than the rest of the year and minima trend (0.5C) is less than maxima trend.

Davis minima:

Davis winter:

• Greater temperature rise in polar regions? The average trend at these sites is 0.71C compared with 1 degree Celsius for the Australian mainland (according to BOM).

Offshore islands-

Willis Island:

Cocos Island:

Norfolk Island:

Christmas Island:


Learmonth Air:


Gove Air:


Mt Isa Air:

Townsville Air:


Coffs Harbour:

Williamtown RAAF (1998 above!):

Nowra RAN:

Canberra Air:

Cape Nelson Lighthouse:

Eddystone Point:

Hobart Air:

Sydney Air:

Melbourne Air:

Adelaide Air:





11 Responses to “BOM vs GISS- Who’s Right?”

  1. Geoff Sherrington Says:

    Ken, One answer to Chris Gilham’s question “What happened in 1992?” could be that in 1990, agencies agreed to procedures coordinated by the World Meteorological Organisation. This might have led to a review and tidy up. One other observation is that the differences between Tmax and Tmin at a number of Australian stations showw a closeness greater than sorrounding years 1990-2 roughly.

    If you are interested, I have yet another version of several of your stations above, taken from the BOM CD ending on March 2007. On a preliminary look, there are again differences between BOM and your GISS figures. Some of this is no doubt due to infill of missing data.

    Let me know what you would like to share.

    I’m interested in a previos observation that some inland BOM stations 1968-2008 show a greater T increase than some coastal stations. Later, I’ll be looking at your Giss figures to see if they overcome this problem.

    Good work Geoff.

  2. Geoff Sherrington Says:

    Please beware of abundant typos above. Ken, if you email me your email address I will give you a web site where I have alternate plots for a number of these stations, plus the daily data on Excel. sherro1 at optusnet dot com dot au

  3. PeterB Says:

    Slightly off topic. Buderim Qld just has had the wettest December since records began in 1893. Not long ago BOM and all sorts of alarmists were predicting dry weather for South East Qld. The state government built a $1.2B desalination plant at the Gold Coast (now to be shut down) with plans for another at Bribie Island or tripling the size of the Gold Coast plant (Courier Mail 4 Apr 2010)
    The record Deember rainfall at Buderim was 609.4 mm set in 1926. This December we already have had 617.1mm. I live about 200m from a closed weather station (040032). This operated in parallel from 1945 to 1981 with 040031 (about 1.5km away) which open in 1893 and closed 1995. I did statistics on the monthly values over the overlap period. The differences of averages for each month and the standard dev. of the difference of each month of the overlap were only a few mm so by the “t” and “F” test one can say they came from the same population.
    One can see cycles in the rain pattern. At other times BOM have linked rain to SOI which has its cycles. The recent drought (below average rainfall) lasted from 2000 to 2008 which is considerably shorter than the Federation drought from 1899 to 1915 (every year below average with the lowest on year on record at 388.3mm in 1902- the highest on record was 1898 at 3996.9mm).
    BOM appear not to be looking at their own records. Further, compounding the forcast situation is the dropping of longterm stations. Landsborough PO (04117) which has been operating since 1892 was closed in October this year 2010. Sunshine Coast Airport (04861 open in 1994) which is right at the coast is one of the few left in this area.

    Happy New Year

  4. Len van Burgel Says:

    Ken, you wrote:

    “When I examined the data, I found that between 1995 and 2009 for the majority of sites the GISS raw data, which should be the same as the BOM data, is significantly less (29 sites significantly less than BOM, 2 higher and 14 similar). It is up to 1 degree Celsius lower, but typically about 0.3 to 0.6 lower.”

    The answer to your question is in the paper by Blair Trewin “Effects of changes in algorithms used for the calculation of Australian mean temperature”

    The mean temperature can be determined in one of three ways:

    (Tmax +Tmin)/2 or
    The mean of hourly or three hourly observations or
    Weighted means of temperatures at irregular hours

    Trewin wrote:
    “In November 1994, a change was made to the method of calculating mean temperatures reported in CLIMAT messages. CLIMAT messages are the means by which monthly climate data are transmitted internationally, and form the basis for international data sets such as those maintained by CRU and NCDC. From that date onwards the mean daily temperatures have been calculated using the mean of all available fixed-hour observations.”

    Trewin concluded “the results of this study indicate that a change from the use of (max+min)/2 to a mean based on regular fixed hour observations (as recommended in the WMO Guide to Climatological Practices) would introduce an inhomogeneity of between -0.18ºC and -0.43ºC into the Australian mean temperature record. This would be a substantial inhomogeneity in the context of the overall Australian temperature trend “

    The Bureau continues to use (Tmax +Tmin)/2 for its calculation of mean temperature and the long term trend. However since 1994, GISS will be using the new means. It will need to adjust for the inhomogeneity. It can go back before 1994 to recalculate the means for Australian stations where there is three hourly data. Or it can simply remove data before 1994 as you report.

    As Trewin notes in his paper, the new method has its problems. If the station is on reduced hours, usually the night time and coldest temperatures are omitted, resulting in an upward bias.


    • kenskingdom Says:

      Thank you Len.
      I’m in holiday mode for the next couple of weeks so I will look at Trewin’s paper and the effects of this algorithm change when I can. My first reaction is that it caused a remarkable difference, and only in the past couple of years have BOM and GISS data converged.

  5. Geoff Sherrington Says:

    For Len van Burgel

    You have brought me a Christmas present. I read Trewin a while back but the significance of your observation did not pop out at me with my then state of understanding.

    I have just taken the liberty of posting your post and reference to Ken’s blog on Climate Audit under “NASA GISS – Adjusting the Adjustments.”

    Maybe some of the puzzles of my post at December 9, 2010 at 12:16 pm above have answers in your response.

    Thank you, thank you.

  6. Geoff Sherrington Says:

    Further to the 1990s transition from mercury-in-glass to electronic sensors with increased reading frequency each day, I was wondering about this two years ago. With help from David Stockwell, I composed a short (and inconclusive) essay.

    It seems that the BOM are in a bit of a bind. When you change a fundamental measurement method, you tend to do so because you expect a better result. That can also be a DIFFERENT result. I’m not a Luddite, when it is time to change to better gear, you do it.

    However, as events have panned out, the BOM has chosen not to be terribly open about changes. One can look up rudimentary meta data sheets for many weather stations to find dates of method changes, but there is no concise official description that I can find to look at the magnite or treatment of the change.

    The BOM had two broad choices: to make the change in one step, or to taper it out over a number of years. The latter approach looks more likely from the data available to me and the essay above, though it is not 100% diagnostic.

    If the new equipment was overall reading lower than the decades before, then a tapered adjustment could be made in two ways, one being by lowering the temperature record preceding the change and the other by raising the temperature after the change. However, sooner or later the reports would have to match the actual record for the day, or the taper adjustment would distort the record too much.

    Strangely, there is some evidence that the BOM are taking another approach, by having a public record and a private record that is different. Australian taxpayers can buy one version for a fairly hefty sum, while the other seems to be going to NOAA in USA free of charge and further responsibility, for overseas adjusters to work further magic.

    Which leads one to wonder if the statement made on 5 Jan 2011 by Dr David Jones – that the decade just ended was Australia’s warmest since 1900 – is correct in an absolute sense, or an adjusted sense, or both.

    Remember that many early records about 20 years each side of 1900 were also adjusted cooler because of the presumed absence of Stevenson screens (even though documentation is sparse).

    Scientic courtesy might indicate that an explanation was in order, given the announced change in methodology. There is still a plausible case that there has been no overall Australian temperature change in the last 110 years, within the limits of error and method change.

  7. Ian Beale Says:


    Not particular to this thread

    Are you aware that Charleville Airport started in 1942. It ran in parallel with Charleville PO 1942 – 1959. The latter record goes back to about the 1870’s

  8. val majkus Says:

    Ken you’ll be interested in this new post by (I think) Chris Gillam
    (a bit of cut and paste)
    The charts below are sourced to data from the GISTEMP recordings by the Goddard Institute of Space Studies and the Australian Bureau of Meteorology High Quality database.

    The GISS data is sourced from raw GHCN data with USHCN corrections. The original source is raw temperature data from the Australian Bureau of Meteorology. Temperatures in the BoM HQ database are adjusted by the bureau.

    Note that in most but not all charts the BoM HQ temperatures trend above the GISS data as of 1993, and temperatures in many of the locations are the same or similar in 2008 and 2009. Various locations in the GISS database, particularly those with long records preceding 1940, have missing data beginning in 1993.

    Why do temperatures in the BoM HQ data, upon which Australia’s historic warming trends are based, consistently rise above the actual raw temperatures in the GISS database, usually from 1993 to 2008?

    Similar trends have been observed by researcher Ken Stewart within the Reference Climate Station sites maintained by the BoM

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: