Hot City or Global Warming?

An analysis of the historical temperature data for the state of Victoria in Australia, including the city of Melbourne, suggests an Urban Heat Island (UHI) effect but no general warming trend.

Urban Heat Island versus Global Warming – A Study of One Region
By Michael Hammer

CITIES represent concentrations of commerce and energy use.  This energy release raises the temperatures in the immediate vicinity.  Cities are also areas where there is intense development with extensive masonry constructions, skyscrapers, paved surfaces and little vegetation.

Large masses of masonry and paving store heat during the day and release it at night keeping the night time minimum temperatures significantly higher than they would otherwise be.   As a result, cities are usually significantly warmer than nearby rural areas, especially at night.  This is termed the urban heat island effect or UHI and it can be very large.  For large cities such as New York or Tokyo, the UHI has been reported as raising minimum temperatures by up to 6-8C.  Even more modest cities like Melbourne show very significant UHI temperature increases.

More significantly, UHI increases as the size of the city increases and as the level of development rises.  Both typically increase with time which means the UHI increases with time.  This is exactly similar to the claimed global warming signature.  

People comment that they have experienced global warming for themselves.  That it is now warmer than it used to be and cite examples such as ice covered puddles in the past which they no longer see today.  If you live in a city (as the majority of people do) that is quite probably true.  However, what you are experiencing is not necessarily global warming but rather the impact of UHI in your immediate environment.  As already stated, the impact can be extremely large – several degrees in large cities.  Nor is it necessarily limited to just the city area.  If there is a prevailing wind and you are living down wind of the city centre then you will be enveloped in the spreading plume of warmth.   So, if we see signs of warming in cities is it UHI or is it evidence of global warming?

UHI has minimal impact on global temperatures because the cities represent such a small fraction of the total area of the total planet.  However, it can have a large impact on the estimation of global temperatures because so many of the measuring stations are in cities.  These stations will show a temperature rise with time which is the sum of any global heating plus the local UHI heating.  If the impact of UHI is not allowed for, the result will be an inflated estimate of global warming.  Even worse, many measurement stations which were originally sited in very reasonable locations are, through later developments, now severely impacted by nearby heat source such as an air conditioner waste heat vent or on top of bitumen paving.  This is yet another factor in addition to UHI adding to the warming bias. 

It is not reasonable to assume that this is allowed for by amalgamating data from a very large number of stations worldwide because many of these sites will be tainted by the same problem.  Even worse, tainted city sites will probably have the most comprehensive records.  Rural sites are more likely to have incomplete records due to the poorer infrastructure.  It is natural to place the greatest reliance on those sites with the most nearly complete data.

The IPCC claimed in the past that one of their corrections to the raw temperature data allows for UHI by applying a linear correction with time amounting to 0.06C per century.  However, in the latest revision to the historical global temperature record even this minimal correction has apparently been eliminated. 

At the same time, measuring stations that have moved from the city to the airport show lower temperatures at their new location and it is claimed that this needs to be compensated for by elevating the airport readings.  This would seem to be clear evidence of bias.  The airport readings are lower because the station has moved away from the city UHI.  Raising the airport readings, while not adding downwards compensation for UHI, results in an overstatement of the amount of warming. It would be more accurate to lower the city readings to match the airport readings rather than vice versa.
It is interesting to explore this issue by looking at the temperature record for the state I live in – Victoria, Australia.  The data presented below comes from the Australian Bureau of Meteorology data base published on their website .  The annual average maximum and minimum temperatures have been used and it is worth noting that all these numbers are averages over 30 years so they reasonably reflect climate rather than weather.
Starting with Melbourne city centre the data is as shown in Table 1.

This data shows clear signs of warming – 1.7C since 1861 in the minimum temperatures.  Furthermore, the warming is far from linear.  Over the first 40 years the minimums rose 0.5C while over the last 40 years they rose 1.5C.  The maximiums were steady over the first 40 years yet rose 0.4 C over the last 40.  This pattern is extremely similar to the pattern of global warming claimed by AGW proponents.   However are we seeing global warming or increasing UHI?

We could differentiate by looking at temperature records from locations which are nearby by not in the city centre.  Since these would still be in metropolitan Melbourne, some UHI could be expected but it should be less.  Unfortunately the Bureau does not list any complete temperature records for nearby locations (exactly the incomplete record problem I mentioned earlier).  However the Bureau does list some data for several Melbourne suburbs and these are useful comparisons. 

Table 2 shows the minimum and maximum temperatures for various locations along with distance and direction from the city centre.

The thing that immediately stands out is that the latest temperatures from all of these sites is very close to the temperatures for Melbourne city centre in the late 19th century: 9.5 minimum and 19.7 maximum.  These sites are all reasonably close to Melbourne central and cover essentially all directions.  It is hard to believe that they were all cooler than Melbourne central in the 19th century.  Nothing in Melbourne’s topography would suggest that.

Only 3 sites have enough data to form any opinion about temperature trends and these all show some increase but far less than the city centre which is what one might expect from UHI.

We could get a stronger confirmation by looking at some site in regional Victoria.  These are likely to show much less UHI since the population centres are far smaller.  Again one is hampered by the lack of comprehensive data for most locations but there are a few sites for which the Bureau gives reasonably comprehensive data.  These are shown in table 3.

The data for Ballarat, Alexandra and Mildura suggest no temperature rise at all over the measurement period and especially (Ballarat and Mildura) over the last 3 decades.  Horsham data suggests some temperature rise early in the century but if anything cooling over the last 3 decades.  Cape Otway is questionable, there is an 0.7C rise in minimums but a 2.1C fall in the maximums most of which occurred in the early part of the century before the rapid rise in carbon dioxide. The last 3 decades have seen a return to the temperatures of the 1920’s.

Overall one would have to say the Bureau data suggests no significant warming over the last century and in particular the last 3 decades in Victoria.  On the other hand it does show significant UHI.  Consider that an increase of 1.5C in the minimums for Melbourne over the last 3 decades corresponds to 5C per century.  Averaging the minimum and maximum readings yields a rise from 14.7C to 15.7C over 30 years equating to 3.3C per century and Melbourne is far from the worst city in the world for UHI.  Look also at how non linear the UHI rise really is.  Compare that with a linear allowance of 0.06C per century.  For that to be appropriate it would mean that only 1 in 50 sites is showing the same degree of UHI as does Melbourne and the other 49 show no UHI at all, even then the linear allowance is not correct.

The data analysed here is real, high quality, data from Australia’s premier weather/climate organisation. It is completely at odds with the claims of about 0.6C of warming over the 20th century predominantly in the last 3- 4 decades.  Is Victoria really so anomalous or are the AGW claims questionable?


Michael Hammer graduated with a Bachelor of Engineering Science and Master of Engineering Science from Melbourne University. Since 1976 he has been working in the field of spectroscopy with the last 25 years devoted to full time research for a large multinational spectroscopy company.

Other articles by Michael Hammer can be found here: 

The picture was taken in Melbourne in 2008 by Jennifer Marohasy.

47 Responses to Hot City or Global Warming?

  1. spangled drongo June 11, 2009 at 9:37 pm #

    Some wise person once said that if the world were really warming that the best way to measure this would be by collecting temperatures from lighthouses and rural areas world wide and thus UHI would be eliminated.
    There are enough of these data points scattered around the globe that have been in operation for over a century to get this information.
    If NASA GISS were honest, that is what they should do instead of “adjusting” on the basis of fairy lights.

  2. Luke June 11, 2009 at 9:41 pm #

    If MH wants to do an analysis to see some real trends in temperatures – try frost frequency and date of last frost in the northern cropping zone. The usual denialist cherry picks.

    Next !

    BTW – Victoria – where’s that?

  3. MAGB June 11, 2009 at 10:02 pm #

    “It is completely at odds with the claims of about 0.6C of warming over the 20th century predominantly in the last 3- 4 decades. ”

    Well done Michael. Even if you’re wrong, 0.6C over a century is hardly anything to get excited about when we have a temperature range of about 100 degrees at any moment on Earth. As Bjorn Lomborg has shown in his expert Copenhagen Consensus exercise, the whole climate change issue is peripheral in the scheme of important things.

  4. bill-tb June 11, 2009 at 11:23 pm #

    Same thing we found in the USA … Where the USHCN measured the growth of the postal service postage increases very accurately. The state of the USHCN is dismal.

  5. cohenite June 11, 2009 at 11:28 pm #

    This paper has an interesting variation on the UHI effect;

    The author ‘simply’ looks at the heat produced by the gross consumption of all the energy sources and concludes that adds ~ 1.12C per century; this conclusion ignores any radiative effect from GHGs. However the paper does not consider the effect of latent heat transfer from the surface and therein the proven temperature moderating ewffect on consequent cloud formation. The solution offered by the paper is drastic reductions in per capita power consumption; that is drastic reductions in SOL; even the green orthodoxy does not advocate this being content to lie about the ability of green energy to maintain the power supply of the fossil fuels.

    Elsewhere Jone’s paper gives an UHI effect of 0.1C per decade;

  6. cohenite June 11, 2009 at 11:43 pm #

    UHI measurement has been one of the areas of dubious science by AGW supporters; the same Jones and Wang are currently under investigation for artificially deflating the UHI effect;

    And BoM has been fudging the books in Australia as well;

  7. Les Johnson June 12, 2009 at 3:35 am #


    If MH wants to do an analysis to see some real trends in temperatures – try frost frequency and date of last frost in the northern cropping zone. The usual denialist cherry picks.

    I have no data for Oz, but the Arctic has shown no significant change in the length of the melt season. While melting is starting slightly earlier, freezing is also starting earlier. In fact, there is an insignificant shortening of the arctic melt season. All of this, in spite of a reduction in arctic ice areal extent.

    A study of the onset of spring, in North America, shows no trend. From the conclusion:

    We found no evidence for time trends in spring arrival from ground- or model-based data; using an ensemble estimate from two methods that were more closely related to ground observations than other methods, SOS trends could be detected for only 12% of North America and were divided between trends towards both earlier and later spring.

    My emphasis.

    White, M.A., K.M. de Beurs, K. Didan, D.W. Inouye, A.D. Richardson, O.P. Jensen, J. O’Keefe, G. Zhang, R.R. Nemani, W.J.D. van Leeuwen, J.F. Brown, A. de Wit, M. Schaepman, X. Lin, M. Dettinger, A. Bailey, J. Kimball, M.D. Schwartz, D.D. Baldocchi, J.T. Lee, W.K. Lauenroth. Intercomparison, interpretation, and assessment of spring phenology in North America estimated from remote sensing for 1982 to 2006. Global Change Biology (in press).

  8. Eyrie June 12, 2009 at 7:17 am #


    How did you get to read the paper? The link only goes to the abstract.

    Last I heard the total human race’s consumption of energy from all sources was about 1/10000 of incoming solar radiation. This leads to a much lower estimate of temperature rise due to this – something under 0.01 deg C.

  9. janama June 12, 2009 at 7:29 am #

    Ole Humlum at http://www.climate4you has also carried out the same tests of UHI that Anthony Watts did in the US. Drive from the country, through the city and out the other side with a thermometer measuring the temp change – he got the same result as Anthony and has documented it at his site.

  10. michael hammer June 12, 2009 at 8:21 am #

    Luke you amaze me. Cherry picking means selecting only that data from a data set which supports a position and ignoring data that contradicts it. In this case the subject is the temperature record in a particular region – the state of Victoria Australia. I have reproduced all the data I could find. I have not rejected or omitted any data I found nor have I in any way changed any numbers. Yet you say this is cherry picking and suggest that a more defendable approach would be to look instead at one narrow proxy whch presumably better supports your point of view. (to forstall your obvious comment – frost frequency in parts of the northern hemisphere is not equivalent to a temperature record, it says nothing about maximum temperatures, nothing about minimum temperatures in areas where frosts do not occur, says nothing about average temperatures, most importantly says nothing about urban heat island effects and anyway does not relate to the region under discussion).

    It seems to me that when comments become too extreme the most likely consequence is that they simply rebound on the person making them

  11. cohenite June 12, 2009 at 8:59 am #

    Eyrie, I was sent the paper; if you provide your e-mail I can forward a copy to you; but really the paper is a bit strange.

  12. Ian Thomson June 12, 2009 at 9:17 am #

    I travel widely in western New South Wales and Victoria ( Australia ) and in the course of my work I talk to many people who have lived out here for generations.
    I can tell you that what you theorise is accepted fact in the country. You mention a posible warming early last century and this corresponds with a trend picked up in collated Arctic temperatures – (University of Alaska, Fairbanks ). Temperatures then trended cooler since about the early 1930s.
    The study is called -Variability and trends of air temperature and pressure in the maritime Arctic, 1875 – 2000

  13. Luke June 12, 2009 at 10:42 am #

    That’s northern cropping zone – Emerald to Dubbo (part of Russia?)

    Why make a single statement about Melbourne when you have an entire continent.

    I agree with your rebound statement.

  14. spangled drongo June 12, 2009 at 11:26 am #

    “If MH wants to do an analysis to see some real trends in temperatures – try frost frequency and date of last frost in the northern cropping zone.”

    Luke, do you mean this cropping zone?

  15. Green Davey June 12, 2009 at 1:10 pm #

    Good one on Canadian frost, Spanglo,

    I have a cousin in Saskatchewan who says he is still wearing his long johns in June.

  16. Luke June 12, 2009 at 3:32 pm #

    SD – No that’s not of relevance.

    BTW any serious student of the record would note an individual year is irrelevant.

  17. kuhnkat June 12, 2009 at 3:57 pm #


    “BTW any serious student of the record would note an individual year is irrelevant.”

    So, computing a trend from 1998 till present is as valid as any other??

    How is that frost measurement in the Sahara working out for you???


    “…but really the paper is a bit strange.”

    Friends of Luke, SJT and Brer Rabbit??

  18. Rick Beikoff June 12, 2009 at 4:28 pm #


    Are you starting to get that horrible, sinking feeling that you’ve backed the wrong horse and are starting to look, well… foolish?

  19. spangled drongo June 12, 2009 at 8:35 pm #

    Do you think the words “For the first time in history Perisher is looking to open Mt Perisher for the second weekend of the season”, might indicate some lack of warming?,28318,25619481-5012995,00.html

  20. Luke June 12, 2009 at 9:13 pm #

    It’s simply irrelevant. If you think that uniformitarianism is how AGW works well you’d be a drongo.

    Oh that’s right – you are.

    The warm autumn suddenly forgotten too as if it mattered.

  21. bobn June 13, 2009 at 7:40 am #

    Yea it’s as stupid as using the example of a couple without kids as evidence the national birth rate is falling. But people will cherrypick and those people tend to be AGW skeptics.

  22. ecosceptic_II June 13, 2009 at 8:49 am #

    Anyone know which model Luke follows?


    Peter Plail (11:46:39) :
    “I’m relatively new to all this scepticism stuff, and I actually started out on Realclimate looking for an intro to climate change science. I then started to read the blogs and was particularly struck by Dr Schmidt’s comments. Now I can understand (to an extent) his attitude to those who expressed sceptical views, but I found unforgivable his impatience with contributors who had a simplistic view of climate science or who simply misunderstood topcs.

    I have a question for any meteorological professionals who may actually know Dr Schmidt – is he really as unpleasant and arrogant as he comes across on his blog, or is he just being aggressively defensive?”

  23. Ian George June 14, 2009 at 2:47 pm #

    ‘The warm autumn suddenly forgotten too as if it mattered.’

    Luke, nationally this autumn was above average maximum-wise (19th warmest) and below average minimum-wise(27th warmest).
    ‘As if it mattered’ is quite right.

  24. kasphar June 14, 2009 at 3:08 pm #

    Michael Hammer’s post above got me checking the records for the town I live in. The trend is similar to Melbourne but the maximum averages have dropped from 27.0C between 1911-1940 to 26.6C between 1971-2000 (average is 26.8C for the entire period). However minimums have risen about 0.1C from 13.1C (1911-1940) to 13.2C (1971-2000). 13.2 is long-term average.

    It appears that from 2001 – 2008 (incl), maximum temps still average 26.6C. Again, the minimum average has risen about 0.7C in the past 8 years but when included in the overall period 1971-2008 shows an increase of around 0.05C . Therefore our mean temps are below the historical average over the past 100 years. I wonder how many smaller centres have experienced the same trend.

  25. michael hammer June 14, 2009 at 6:25 pm #

    Kasphar, your comment is very interesting and pertinent. If anyone is still reading this thread why not look up the data for the town you live and in post the results. We can have an on line survey of how other towns stack up. It would be most interesting.

  26. Jan Pompe June 14, 2009 at 6:42 pm #

    “If anyone is still reading this thread ”

    I am.

  27. louis Hissink June 14, 2009 at 8:29 pm #


    “Yea it’s as stupid as using the example of a couple without kids as evidence the national birth rate is falling. But people will cherrypick and those people tend to be AGW skeptics.”

    Supported by what evidence?

  28. Ian George June 15, 2009 at 9:01 am #

    As a quick follow-up to my above comments, I went to the BOM site ‘Climate statistics for Australian sites’ and randomly selected country sites throughout NSW. I only checked sites where the weather station had a history from early 1900 to 2000+. I found that the following stations had higher average maximum temps during 1911-1940 than during 1971-2000 (these are reasonably comparable periods as there was a significant world-wide warming increase from 1910 -1940 and again from 1976 – 1998). The sites are;
    Murrurundi, Gunnedah, Bourke, Dubbo, Parkes, Balranald, Deniliquin, Broken Hill, Condobolin, Corowa, Parkes, Casino and Scone.

    Maybe some of these sites have changed location (ie from PO to an airport) but they are a consistent record. Despite this, with a rise of approximately 60ppm+ in CO2 from 1911-2000, there seems to be little effect on av max temps over 30 year periods in country NSW.

  29. michael hammer June 15, 2009 at 9:15 am #

    Hi Ian;

    Thanks for that. To be fair one should document all sites found not just those showing higher temepratures in the 1930’s than the 1990’s – maybe that’s what you did although your wording suggests that you are only reporting those sites fitting the above criterion. One should also look at minimum temperatures since they are probbaly even more indicative.

    Maybe I will have a look at BOM data for other Australian states starting with the sites you mention – its just a lot of work. If I find the time to do it I will report back on this thread in the next couple of days.


  30. Ian George June 15, 2009 at 9:57 am #

    I have checked many other sites but they don’t meet my criteria of single site measurements 1910-2000 and smaller country areas where UHI should be insignificant. I will look at coastal areas and check minimums of the above but, as you say, it’s a lot of work.

    I have a feeling that CO2 may be only a passive insulator and may have more effect on nighttime temps rather than daytime due to emission of stored heat from buildings and tarred roads being trapped by CO2, much the same way as clouds effect temps at night.
    Thus, larger centres should show a greater increase in average minimum temps than maximum temps as shown by your Melbourne example. However, the positioning of stations is probably important in the analysis.
    One example is Sydney (Obs Hill) which shows an 0.6C increase in av max temps but an 0.9C increase in av min temps. Yet Newcastle’s Nobby Beach shows an increase of 0.2C for both max and min averages.

  31. wattle June 15, 2009 at 10:38 am #


    Your post was well presented and researched but would appear to be contradicted by the BOM site.

    Perhaps their data is corrupted by UHI!

  32. Ian George June 15, 2009 at 2:48 pm #

    I have just run a check on minimum temps as well and on some more small country sites and am finding that average max temps are down from 1911 (except Griffith – 0.1 increase) but minimum temps are up except in a few cases (eg Murrurundi).
    Remember I am only checking smaller towns with a 1911-2000 historical record (I shouldn’t have included Dubbo or Condobolin as they both are a little short of the 2000 year end).
    Looking at your link, much of the data is taken from 1955 – 2007. The 1950 – 1975 period was a comparative cool period (temps began dropping in the 40s and started rising again in the late seventies). A longer data period may have been more interesting.
    Also the climate temperature trend maps at don’t show this trend I am seeing as they may be based on mean temps rather than straight max temps.
    But I have checked enough NSW towns to see that there is a trend towards max temps dropping and the min temps rising over the last 90 years. What this really means is anyone’s guess.
    Maybe other states are different.

  33. spangled drongo June 15, 2009 at 6:42 pm #

    Michael Hammer,
    Following my first post re lighthouses I wasn’t aware this info was available but I have just checked a few in my area:

    Lady Elliot Is. Earliest 1940 1940 25.2c 1998 27.0c 2008 25.4c Highest 1998 27.0c

    Sandy Cape 1907 26.4c 1940 27.0c 1998 26.6c 2008 25.7c 1942 27.5c

    Cape Moreton 1913 23.6c 194023.2c 1998 23.9c 2008 23.6c 2005 24.5c

    Yamba 1878 23.7c 1940 23.4c 1998 23.9c 2008 23.9c 1885 24.5c

    Smokey Cape 1939 22.9c 1940 22.8c 1998 24.0c 2008 23.3c 2002 24.9c

    Coffs Harbour 1952 23.7c 1998 23.7c 2008 23.3c 2004 24.2c

    Nobbys Newcastle 1862 24.3c 1940 22.1c 1998 22.2c 2008 21.5c 1877 25.2c

    It is obvious from these data that GW is not happening as in every case where the data went back to the 19th century, it was hotter then than the 20th or 21st.

  34. michael hammer June 15, 2009 at 7:55 pm #

    This is really interesting but somewhat perplexing. Both the BOM site cited by Wattle and the BOM site cited by Ian George claim consistent and significant warming over the 20th century in all states including Victoria – at least 0.5C to 1C and in many cases more if I read the data correctly. I presume this is based on the BOM data record although that is not specifically stated. They also do not sate what “corrections” if any were applied to the data. Maybe no correections were applied but after seeing the NOAA site I am cautious.

    Yet when I look up the same? data record what I find is not warming in Victoria. Spangled Drongo has apparently done the same for several NSW sites and the results he cites show no consistent warming. One question S.D. your values look like they are individual year results which can be expected to have great variability. Did you find any longer term averages or was there enough data for you to calculate such averages. Citing a sparse record of individual years leaves on open to a charge of cherry picking. As an example, Ian George cited Nobbys Newcastle as showing an O.2C increase in min and max yet your data shows a significant decrease.

    Ian you comment that the trend you found is for dropping maxima and rising minima. How much fall and how much rise?

    Kasphar also found no significant change in the town he lives in. It seems to me that several of us have gone looking and have not found the warming trend in the raw data.

    I guess I should fill in the gaps and collate data for other states and present that in table form as well.

    Of interest, I also looked at the NOAA global temperature data both raw and after “corrections”. The corrected data shows a distinct warming trend along the lines of the graph we have all seen from NASA GISS. However I plotted the raw data and it shows a large hump in the 1930’s a smaller hump in the 1980s and a return to baseline by 2000 (there was no data post 2000 on that site), in short, variability bu no net warming. All the warming was due to the correctons that were applied. I have written this up as an article and its with Jennifer for posting when she has the time and space.Its beginning to sound as though the same result is starting to emerge here.

  35. michael hammer June 15, 2009 at 8:20 pm #

    I just downloaded the data for Nobby’s Newcastle. This is from the climate data on line portion of the BOM website – climate statistics for australian locations. The first number is the minimum the second the maximum


  36. michael hammer June 15, 2009 at 8:20 pm #

    I just downloaded the data for Nobby’s Newcastle. This is from the climate data on line portion of the BOM website – climate statistics for australian locations. The first number is the minimum the second the maximum


  37. michael hammer June 15, 2009 at 8:33 pm #

    Opps this website does strange things. Apparently the tab key submits? To try again. The third number is the average of max and min

    1871-1900 13.7 22.3 18
    81 – 1910 14.1 22.1 18.3
    91 – 1920 14.2 21.8 18
    1901-1930 14.2 21.8 18
    11-40 14.3 21.5 17.9
    21-50 14.3 21.4 17.9
    31-60 14.4 21.3 17.9
    41-70 14.3 21.1 17.7
    51-80 14.3 21.2 17.8
    61-90 14.3 21.4 17.9
    71-2000 14.5 21.7 18.1

    so yes the last number shows about 0.2C rise in both max min and average as Ian suggested but then 1900 to 1910 showed an 0.3C rise and then an 0.3C drop after a further 10 years. To me this loks like a very stable temperature record. The changes are far smaller than the 0.5 to 1C being talked about and look pretty random. I don’t think this site corroborates the BOM presented analysis.

    I also looked at Coffs harbour but the data was far too sparse to draw any sort of conclusion at all. It only really covered 1930 to 1966 although ove that time the min temps were stable 13.3 while the maximums fell from 23.6 in 31-60 to 23.2 in 51-66. As I said, far ot sparse data to draw any conclusion. This is one of the problems – there are not many good data records to use.

  38. spangled drongo June 15, 2009 at 9:03 pm #

    I used the mean maximum and selected the earliest and the latest data plus the two warm periods of 1940 and 1998.
    It is interesting that of the sites with no 19th century data, Sandy Cape had its warmest in 1942 with 27.5 and a median for all years of 25.8, Lady Elliot warmest 1998 27.0 and median of 25.5, Cape Moreton warmest 2005 24.5 and median 23.1, Smokey Cape warmest 2002 22.9 and median 23.0, Coffs Harbour warmest 2004 24.2 and median 23.2.
    The two that had 19th cent data Yamba warmest 1885 27.3 median 23.1, Nobbys Newcastle warmest 1877 25.2 median 21.6.

    I can’t understand [or maybe I can] why a lot of the 19th century data was thrown out on the basis that the old thermometers were on verandahs and not housed in stevenson screens and therefore hot-biased.
    Often those verandahs were where the old open topped canvas water bags and various other water wicking devices designed to cool everything possible, were kept. Some verandahs even had two layers of chicken wire with spinifex needles in between wicking up water from a trough around the bottom. The air passing through would cause evaporation and create a delightful microclimate. It was common to hear someone say, “it’s 112 in the bottom of the waterbag”.
    I doubt that those old verandah readings would have been biased towards hot.

  39. michael hammer June 15, 2009 at 9:56 pm #

    Now have data for some other NSW sites


    1881-1910 15.5—24.3—19.9
    91-20 15.5—23.5—19.5
    01-30 15.3—22.9—19.1
    11-40 15.3—22.8—19.1
    21-50 15.0—22.8—18.9
    31-60 15.0—23.0—19
    41-70 15.0—23.0—19
    51-80 15.2—23.0—19.1
    61-90 15.4—23.0—19.2
    71-2000 15.5—23.2—19.4

    certainly a dip mid century and a rise back up towards the temps at the start of the 20th century

    Gabo island

    1861-1890 11.6—19.4—15.5
    71-00 11.5—19.2—15.4
    81-10 11.4—118.4—14.9
    91-20 11.8—17.7—14.8
    01-30 12.2—17.4—14.8
    11-40 12.3—17.5—14.9
    21-50 12.2—17.4—14.8
    31-60 12.2—17.2—14.7
    41-70 12.3—17.2—14.8
    51-80 12.5—17.4—15.0
    61-90 12.5—17.7—15.1
    71-2000 12.4—18.0—15.2

    again a dip in the middle part of the century followed by a rise not quite back to the late 19th century

    Broken Hill

    1891-1920 11.7—24.6—18.2
    01-30 11.8—24.3—18.1
    11-40 11.7—24.3—18
    21-50 11.7—24.3—18
    31-60 11.5—24.0—17.8
    41-70 11.8—23.8—17.8
    51-80 11.9—23.7—17.8
    61-90 12.2—23.8—18
    7-2000 12.3—24.0—18.2

    looks like a substantial rise late in the 20th century but in fact its the same pattern witha dip in the middle of the 20th century and a rise back up to in the case the saem as the start of the 20th century

    I also have data for Parkes – shwos exactly the same thing and for Golburn which shows a dip the a rise toa bout 1960 and then another dip from 13.7 av in 1890 to a trough of 13.3 in the 20’s and 30’s then a peak of 14.3 in the 50’s and a drop to 13.9 in the 70’s.

    I am getting tired of typing tables.
    However what all these sites seen to me to show is a very stable temperature record moving around by a few tenths of a degree but with late 20th century temps very close to late 19th century temps. Hardly a monotonic trend and certainly not anywhere near 0.1C per decade (1C per century).

    I still can’t reconcile this with the BOM digested results.

  40. Jan Pompe June 15, 2009 at 11:55 pm #

    Mike “so yes the last number shows about 0.2C rise in both max min and average as Ian suggested but then 1900 to 1910 showed an 0.3C rise and then an 0.3C drop after a further 10 years. To me this loks like a very stable temperature record.”

    Indeed it has actually been more stable than we aim to get our industrial temperature controllers, which we generally do to the nearest degree c. Though I’m sure we could do better if we needed to.

    You guys are finding much the same as I did 4 years ago. In a way it’s discouraging to have to go over the same ground over and over again.

  41. Jennifer Marohasy June 16, 2009 at 1:49 am #


    Yes, people keep rediscovering the facts. I think it is wonderful. And everyone comes from a slightly different perspective.


    I’ve emailed you with some suggested changes to the next piece. I am a bit short on pieces with some data at the moment, so am keen to post ASAP.

  42. Ian George June 16, 2009 at 7:13 am #

    ‘How much fall and how much rise?’

    Your results mirror my own jottings and so I presume you have used the BOM’s ‘Climate statistics for Australian locations’ data (30 year statistics). I just compared 1911-1940 temps to 1971-2000 temps for smaller country towns (to try to eliminate UHI).
    To answer your question, max temps range from no change to 1.0C below over the past 90 years and minimums range from no change to 1.2 above. There are exceptions such as Griffith (0.1C max increase) with Balranald, Deniliquin and Murrurundi showing decreases in minimum temps but I am finding the majority of the 20 country towns I checked fit this profile (I haven’t check coastal towns as yet).
    What were your conclusions from your work you did 4 years ago?
    I only began to take an interest recently when I read about all these records supposedly being broken, only to check back and find that far warmer periods existed in the past (BraveNewClimate is good at that).
    The 1981-2010 averages will be interesting when they become available.

  43. Jan Pompe June 16, 2009 at 4:15 pm #

    Ian: “What were your conclusions from your work you did 4 years ago?”

    From memory much the same as you are finding I can’t dig it up since it was all on a hard disk on a computer that found it’s way to the nature strip two cleanups ago. I’ve kept the disk so I may see if I can retrieve the data. My focus however was more on which stations were being kept for the statistics (GISS) and which dropped out over the years. IIRC more stations from aerodromes in rural areas were kept and others dropped out.

  44. Ian George June 17, 2009 at 9:36 am #

    I have just checked 30 NSW cities and towns including coastal, coastal hinterland, tablelands, slopes and western plains using BOM data found at Climate statistics for Australian locations.
    I found that maximum temps during 1971-2000 are on average 0.27C below the 1911-1940 temperatures but minimum temperatures are 0.33C above the earlier period. This gives an approximate mean temp rise of just 0.06C rise over the past 90 years.
    For some of the locations I had to take data from a neighbouring site (ie, in the same town/city) as the record had stopped at the original weather station.
    This is rather confusing as it does not tally with official BOM anomaly temp maps and current thinking. However, if you look at the long-term temperature history for NSW at there is some correlation between these two periods but the national anomalies don’t show the same. So I guess that other states do not show the same patterns we have seen in NSW.

  45. jennifer July 3, 2009 at 9:47 pm #

    Hi Jennifer … problems again with the message submission form on your website.

    I tried twice yesterday to submit a message to the page

    I tried again today but the site told me I’m trying to submit duplicate content. I’m not sure if the message has appeared on some other page or not at all.

    Whichever, below is the message I’m trying to submit …

    It’s interesting to see the min and max temps from different Victorian locations over the past 150 years but it should be noted that data before the Stevenson Screen (~1910) isn’t considered reliable for comparison with years hence. That’s the BOM’s position and, although it’s debatable whether the screen created a negative or positive min and/or max temp bias against the earlier records, I think the valid comparison is since the 1911-1940 bracket. If anything, that suggests a slight fall in temps across the five regional locations in Victoria.

    For WA figures extracted from the BOM databases, see

    WA in summary…

    1876-1899 vs 1979-2008 from 13 locations – average min up .41 degrees and max up .34 degrees

    1876-1899 vs 12 months to May 2009 from 13 locations – average min down
    .51 degrees and max up .45 degrees

    There are 20 WA locations that have both pre and post 1910 records.

    ~1900 vs 1979-2008 from 20 locations – average min up .44 degrees and max up .48 degrees

    1910-1939 vs 1979-2008 from 20 locations – average min up .7 degrees and max up .5 degrees

    There are 32 WA locations with consistent records from ~1900 to the present day

    ~1900 vs ~2000 from 32 locations – average min up .56 degrees and max up .52 degrees

    ~ 1900 vs 12 months to May 2009 – average min down .12 degrees and max up .75 degrees

    The WA data suggests a rise in maxima and a drop in minima, particularly over the past decade and possibly reflecting the reduced rainfall cloud cover in the south of the state – albeit at odds with a huge increase in rainfall in the north.

    Perth, which is by far the state’s biggest heat sink, had the highest averaged increase in mean maximum temperature among all 32 locations, including pre-1910 data, over the hundred years (1.7 degrees C).

    A comparison of the five WA locations with a population above 20,000 and the remaining 27 locations below 20,000 suggests the heat sink adds between .5 and 1 degree C.

    Chris Gillham

    Average temperature trends across Western Australia



  1. Jennifer Marohasy » How the US Temperature Record is Adjusted - June 27, 2009

    […]  […]

  2. An Australian look at USHCN: 20th century trend is largely if not entirely an artefact arising from the “corrections” « Watts Up With That? - June 29, 2009

    […] removed in the latest round of revisions to the historical record.  To put this in perspective, in my previous article on this site I presented bureau of meteorology data which shows that the UHI impact for Melbourne […]

Website by 46digital