Perth's summer- hottest or not?

Perth temperature history

Guest post by Chris Gillham

It’s worth digging a bit deeper into claims by the Bureau of Meteorology that Perth endured its equal hottest summer in 2010/11 and its hottest first four months of the year on record.

On May 2, 2011, The West Australian newspaper published the following story, claiming that “Perth has experienced its hottest start to the year on record, with a mean maximum temperature of 31.5C for the first four months, breaking the previous record of 31.3C set in 1978”.

record perth summer

Oh dear, Tim Flannery was right and Perth really is at the forefront of global warming.

But can we trust the BoM calculations? Was the mean maximum in the first four months of 2011 really 31.5C?

Well, we can do it the BoM way and calculate the average of the rounded monthly figures for January (32.5C), February (34.1C), March (31.9C) and April (27.3C). Yep, that’s an average 31.45C, which further rounds out to 31.5C. Too easy.

But what if we don’t average the rounded monthly figures and instead average the actual maximum temperatures recorded on each day during the first four months? It’s pretty easy to do. Just check the monthly figures for Perth Metro (station 9225), which is the official “Perth” site for the BoM, for January, February, March and April 2011.

The actual maximum temperatures in all 120 days of January, February, March and April 2011 can be seen here. Get your spreadsheets busy and figure out the average maximum. You’ll find it’s 31.42, which rounds to 31.4C.

So the actual mean maximum for Perth in the first four months of 2011 was 31.4C, not 31.5C as claimed by the BoM.

On March 3, 2011, the BoM claimed that Perth’s 2010/11 summer was the equal hottest on record with a mean maximum of 32C. But if you use the daily temperatures instead of the rounded monthly figures, the average summer maximum was actually 31.89C, or 31.9C when rounded. Again, their calculations were based on an average of the rounded monthly maxima rather than the days themselves (see Was Perth’s summer of 2010/2011 a record?).

2010/11 was the third hottest summer recorded in Perth, but not the equal record.

Temperature station location

Ah, but this is being a bit pedantic, is it not, since the first four months of 2011 were still the hottest on record?

That depends upon whether you also want to research a bit deeper into the record itself. The BoM references its official “Perth” temperatures from station 9225 in Mt Lawley, which opened and recorded its first annual temperature means in 1994 – picture below in May 2011 courtesy Stephen Harper.

perth metro bom station

However, the BoM quotes the previous hottest first four months from 1978. So where was this location?

Before 2004, the official Perth temperatures were recorded at Perth Regional Office (station 9034). Check the first four months of 1978 and the mean maximum was 31.3C, the previous record as stated by the BoM.

Stations 9034 and 9225 are about four kilometres apart and the evidence suggests the Mt Lawley site is slightly hotter during the day than the Perth Regional Office site. Site 9034, which opened in 1876 and closed in 1992, has a 19 metre elevation.

Site 9225, which has had a stable mean temperature since opening in 1993, has an elevation of 25 metres. Proximity to broad river waters and the UHI influence of inner city and university development are probably relevant.

map of perth regional and perth metro stations

Below are the average annual maxima of station 9034 from 1975 until 1992, bleeding into station 9225 from 1994:


















1992/93 n/a as switch made from 9034 to 9225 below


















The average annual maximum at Perth Metro over 17 years is about .4C warmer than it was earlier at Perth Regional Office. If you suspect the increase is because of global warming, not the four kilometre change in location, consider the annual minima in “Perth” from 1975 to 2010:


















1992/93 n/a as switch made from 9034 to 9225


















The average annual minimum at 9225 over the past 17 years was about 1.6C cooler than it had been at 9034, the drop happening immediately the station change was made. Maybe a rapid onset of global cooling, but a more robust explanation is that the four kilometre change in station locations significantly affected temperature recordings and made historic pre-1994 comparisons dubious at best and irrelevant at worst.

The temperature record at Perth Regional Office

Perth Regional Office (9034) started recording from 1897 and finished in 1992. This location was originally known as the Observatory and started recording temperatures in 1897 at the Mt Eliza site now mostly occupied by Dumas House, adjacent to Kings Park. Although the BoM gives no clues that Perth Regional Office has a history of more than one location, station 9034 is the benchmark for historic temperature comparisons in the capital of Perth.

perth observatory

View of Observatory grounds soon after completion with Fraser Avenue and Kings Park in the background. The meteorological instruments including a Stevenson Screen can be seen in the right of the photograph.

Most of the Observatory was knocked down in 1967 to make way for Dumas House, and at that time the Commonwealth transferred its meteorological instruments to Wellington St in East Perth – about two kilometres away. (Source: Perth Observatory History)

bureau of meteorology in wellington st perth

bom in wellington st

The Bureau of Meteorology Perth Regional Office at 127 Wellington St, East Perth, facing north with the instrument enclosure in the foreground

The elevation changed from about 61 metres at the Observatory to 19 metres in Wellington St …

bom observatory and perth regional office

… and look what happened to the temperatures recorded for Perth Regional Office (9034) …

perth regional office temperature records

The trend is a bit more obvious just looking at the mean temp for Perth Regional Office …

perth regional office mean temperatures

It should be noted that in 1963, Perth Regional Office was moved about 300 metres to the east of the Old Hale School building, also atop Mt Eliza (Source: Perth’s Observing Sites)

And for Perth’s temperature record from 1897 to 2010 …

perth historic temperature record

So Perth Regional Office changed its physical location in 1963 and again in 1967, when it moved about two kilometres east into the central city, closer to the river both in distance and elevation but separated by blocks of large buildings, and Perth’s temperatures increased thereafter.

To further illustrate the distortion caused by moving the Stevenson screen temperature recording instrument a few kilometres from its original location, the Mt Lawley 9225 temperatures four kilometres inland show higher maxima and significantly lower minima.

Based on the temperatures charted above, the last two decades in Perth have had the hottest days and coldest nights since records began in 1897. Why are records claimed for average maxima but minima and mean are ignored, and why are any of them relevant for comparison with other locations before 1994?

It’s worth looking at Results of Meteorological Observations Made in Western Australia During 1908 (PDF 29mb) published in 1912 by Commonwealth Meteorologist Henry Hunt:

history of the observatory in perth

Well, until 1967 anyway.

Perth Gardens temperature overlap

Perth Gardens was Perth’s first temperature recording location as of 1876, situated about 1.5 kilometres from the Observatory site on Mt Eliza in an area since renamed as the Supreme Court Gardens (see map above).

The 1912 Hunt document also contains a table with monthly and annual mean temperature readings at Perth Observatory from 1897 to 1907, and at Perth Gardens from 1876 to 1907, which provides a useful 11 year overlap.

perth temperature comparison

The 11 year mean from 1897 to 1907 at Perth Observatory atop Mt Eliza is 63.9F (17.72C) and the same 11 year mean at Perth Gardens is 64.9F (18.27C). That’s almost 0.6C difference in the same city in the same weather in the same years. The Perth Gardens did not record temperatures from a Stevenson Screen and this may have artificially increased the temperatures, but Commonwealth Meteorologist Henry Hunt suggests the thermometer was well shaded:

perth gardens temperature

The mean temps for Perth Regional Office from 1897 to 1907 (go to BoM Climate Data and look up min/max for station 9034) are exactly the same as the mean temps for the Observatory in the 1912 Hunt document. These identical temperatures confirm that Perth Observatory on Mt Eliza was the original location for Perth Regional Office temps in its early years starting 1897.


When informing the public that record high temperatures have been experienced, the Bureau of Meteorology should ensure the accuracy of its figures by averaging daily temperatures instead of rounded monthly calculations.

Rounded averages of rounded averages are not accurate and a .1C error is significant if claiming record temperatures, particularly when Australia’s economic and political future is likely to be determined by a public understanding of the need or otherwise for a carbon dioxide tax.

The BoM should also acknowledge that temperatures recorded at its Mt Lawley 9225 station are being compared with stations in different locations where local environmental factors have a significant influence on Stevenson screen thermometers.

Perth’s temperature history is too disjointed to make meaningful comparisons and present-day recordings at Mt Lawley should only be compared to recordings from the same site beginning in 1994.

It might mean Perth has a very short history of temperature recordings, but at least the data will be relevant.


Chris Gillahm is a 50 year old print/electronic journalist based in Perth who also does graphics, writes html, takes photographs, etc. His interest in the climate change issue stems from the fact that he was university-trained as a journalist to do old-fashioned research and not rely on the accuracy of government media statements. Digging such as this almost always reveals the public only gets half the facts about issues such as climate change. Visit his website here:

0 0 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
Scottish Sceptic
May 19, 2011 12:38 am

When I was born in Australia it was something like 110F? in the shade and my mother put me in a bath of water, which being rapidly cooled was extremely cold which apparently made me rather unhappy. And it just so happened that as I was being dumped in freezing water,the same time the UK experienced one of its coldest winters.
And what do we find now: the UK has just experienced one of its coldest winters, and Australians are being dumped in the cold water of carbon taxes because of the “unprecedented” heat!

Mike Bromley
May 19, 2011 12:41 am

Hottest summer, by a fraction of a degree. What does that really mean? If you read the newspaper story, That implies that everything is wilting and dropping dead, when it really means that a person literally wouldn’t be able to tell the difference…especially if they walked between the two temperature stations. Either case is basically picking fly sh*t out of black pepper, creating alarm over a tiny bit of difference.
As much as it is good to pay attention to siting as regards absolute temperature values, if you plotted this in kelvin, the trend is flat. The difference between warmists and skepticas is that warmists like to place significance on nearly immeasurable differences, wheras skeptics just shrug.
We are still within ‘normal’ ranges.

May 19, 2011 12:46 am

Isn’t averaging averages one of the things you learn NOT to do at an early stage? It can lead to all sorts of errors, mistakes and misconceptions.
Isn’t that what they did with the numbers lead them selves up the garden path by averaging the averages?

Al Gored
May 19, 2011 12:51 am

Outstanding research. Your head must have been exploding when you discovered the details of that historical record.
You’ve put together into an overwhelming case. How will they dodge this?
Rather good timing too.

May 19, 2011 12:54 am

Very detailed investigation Chris.
Mt Lawley station resembles a football field size square of dry brown dirt. Is that what all of Perth is like? If not, no wonder it is now reading higher maximums and also lower minimums. If I remember right dry plowed fields will do the same, great absorbers in the day (heating) are also great emitters at night (cooling) leading to higher diurnal range.

May 19, 2011 1:08 am

I am impressed that Chris has done the basic research for this article that all those who call themselves journalists should but few ever do. It’s not exactly a surprise that the changes in Perth during the last century as the city developed render the ‘historical record’ a nonsense, equally unsurprising that the BOM uses the muddled history and silly maths to claim the most recent summer in Perth “Hottest evah, Sport!” The BOM is another organisation that resists embracing openness and honesty in it’s ‘scientific’ practices and resists FOI requests with the thinnest of excuses.

Philip Thomas
May 19, 2011 1:14 am

Great work. Now, can we make BoM acknowledge their flaws in print?

Hector Pascal
May 19, 2011 1:51 am

Nice work. Thanks.
The Mt Eliza site is at the top of a big hill (by Perth standards), the press cutting shows 200ft. Wellington St is close to sea level. In the summer, when the afternoon SW sea breeze comes in, the wind fairly whips off the river into the city area. Mt Eliza is well sheltered from that by Kings Park.
Mt Lawley would be the most “typical” site for suburban Perth, but all three are geographically distinct.

John Marshall
May 19, 2011 2:04 am

Which only goes to reinforce my assertion that global temperature measurement is not only difficult to measure correctly but really not a significant measure of human input into the climate mix of drivers. The only real human input seems to be changing the station positions and fudging the data to fit the output of the models upon which reliance is everything to some.

Alan the Brit
May 19, 2011 2:04 am

W.S. Churchill is reputed to have said,”there are lies, damned lies, & statistics!”. It’s is amazing what one can do with numbers, & it would appear that our Colonial Cousins in Australia’s BOM are having some fun. I always wonder what the Global Average Tempreature really means in practice. Some areas experience warmer than usual weather patterns, other areas experience cooler than usual weather patterns, simples. Ooooops! Better be careful I was begining to sound like a highly accurate computer generated Wet Office climate prediction in the year 2100!!!!:-))

May 19, 2011 2:17 am

Great post Chris. it’s very exciting to see my old home town represented in such an international forum.
To be honest though, I don’t see anything too suspect, other than the usual specious claims of ‘hottest summer on record’. Since the records don’t go back very far, and since the warming trend was already in evidence, it is not really a big deal. If you look on the temperature anomalies map, the whole of the West coast of Australia is considerably warmer than usual compared to the rest of the world that is considerably cooler than usual. I tried to find some kind of link between La Nina and Perth’s weather but couldn’t. I suspected that the effect of La Nina is to cause particularly hot dry conditions for the West coast of australia.
I regularly go home to Oz and I have noted that summers in the earlier part of the decade were disappointingly cool, which show up on the graph. My memory of growing up is that March in particular was extremely hot. I do understand though, the recent summer was very unpleasant.
But you raise, inadvertently, another interesting issue which is to do with the way average, or mean temperature is calculated. My understanding is that the maximum and minimum temperatures are simply taken, but that says nothing about the mean day time and mean night time temperatures. In Perth at least, that would be highly dependant on the sea breeze. If the sea breeze does not come in, as is often the way in March, the temperature can stay high most of the day, but not necessarily as high as it can get by 11am-12pm before it kicks in in January or February.
Simply averaging the Max and Min temperatures can lead to a misleading result.

Alan the Brit
May 19, 2011 2:25 am

Hello, folks. I forgot to add, an interesting little piece on last night’s BBC1 magazine show, The One Show, on Sir William Ramsey, a British chemist who eventually discovered neon. This apparently came about because he had analysed natural nitrogen in the atmosphere & found it to be different from the nitrogen produced by chemical reactions, even though that difference was only 1/10th of one percent!!!! Now, where have I heard that a change of 1/10th of one percent can’t make a difference? TSI perhaps? Funny things, numbers!

May 19, 2011 2:26 am

You got to love the way the warmists will pluck just any random set of data out of the air and then proclaim a record.
We should be going along with it and saying “oh look, its 0.3Celsius warmer than it was 30 years ago. No need to rush is there?”. At that rate, even it it was based on any kind of realistic comparison, we’d run out of fossil fuels long before we had to worry about climate change.

Nick Stokes
May 19, 2011 2:42 am

You can get the daily data for Mt Lawley from here. It is station 94608. For some reason it’s in Fahrenheit. But by averaging daily max I get, yes, 31.4C for the first 4 months of 2011. And 30.2C for the first 4 months of 1978.
I actually remember the summer of 1978 – I was there. It sure was hot, with a late blast from Cyclone Alby – but apparently not as hot as 2011.

May 19, 2011 2:45 am

As one moves inland from the coast, the max is going to get warmer and the min is going to get colder. What is so special is that people still go to the beach to escape the inland heat during a hot summer day.

May 19, 2011 2:58 am

Nick Stokes says:
May 19, 2011 at 2:42 am
You can get the daily data for Mt Lawley from here. It is station 94608. For some reason it’s in Fahrenheit. But by averaging daily max I get, yes, 31.4C for the first 4 months of 2011. And 30.2C for the first 4 months of 1978.
Interesting. So merely the Celsius to Fahrenheit conversion is also raising the temperatures by rounding.
That is real interesting. Now convert back to Celsius from NCDC numbers then average, does it go back down, stay level, or another step up?

May 19, 2011 3:03 am

I’ve lived in Perth for 39 years. I think this is the first time anyone has considered the move in the “official” weather site from East Perth to Mount Lawley. Rainfall as well as temperature is measured at the official site. However rainfall in Perth is very localised and the 4Km move is very likely to have a significant effect on the rainfall reading for Perth. I don’t think parallel readings for temperature or rainfall were compared for say 2-3 years after the move although it is apparent such comparisons would have been highly desirable from a scientific point of view. Perth, which is very isolated, has about 10 continually monitored weather sites dotted around the metropolitan area. A better record of temperature and rainfall would be gained by considering the results from all of these sites and obtaining composite results that reflect the temperature experienced (and rainfall occurring) over the entire metropolitan area. Will this occur? Unlikely.

May 19, 2011 3:08 am

Regardless of whether it holds the record or not, it is an absolute impossibility that any person could objectively (or subjectively) tell the difference of a monthly average of 0.5 degrees.
So the alarmist newspaper stories may interview people who ‘know’ it is getting hotter, but plainly they are talking out their downward facing hole.
Every year it gets hot in summer and the media finds some nitwit sweltering on the beach that says ‘it gets hotter every year’. No, pea-brained-vox-pop commentator, you subconciously absorb dross from the alarmist media and think it is getting hotter.

Hector Pascal
May 19, 2011 3:24 am

“Mt Lawley station resembles a football field size square of dry brown dirt. Is that what all of Perth is like?”
Pretty much. The climate is Mediterranean. Perth itself is on a coastal sand plain. During the last glacial the coastal plain very dry and was an active dune field. Now it is wetter, and the dunes have been stabilised by scrubby vegetation. Eucalypts and banksias, mainly. “Sandy” is the word.
It’s more to do with the Indian Ocean Dipole than ENSO. I can’t find much in the way of meaningful information, but I believe there is a decadal trend towards increasing dryness. It needs more research: send me money 😉
Perth summers are hot. I was there (without aircon) for the max max, 46.8C iirc, in about 1990. It was hot.

Stephen Harper
May 19, 2011 3:30 am

No Perth does not look like this. I took the photograph of the Mount Lawley station a few days ago. All around was green – but reticulation was being used. However, the plot of land on which the station stands is particularly arid. It surprised me how barren it was. Perhaps there is a herbicide regime in place to keep down any vegetation. This summer was a hot one – so any growth which might otherwise have been around the gauge had little chance of surviving without the relief of shade. I do wonder if there is a feeback mechanism in play here. It is hot, so the vegetation dies back. This exposes more sand. The sand heats up and hey presto! the temperature (reading) rises yet further.
On another point: the 2010/11 Perth summer did not have a single day over 40 degrees C as I recall. Just lots of days consistently over 30 degrees C. So it was a different kind of hot summer to the usual. This may indicate the effect a one-off weather event (a particularly warm Indian Ocean off the coast of Perth this year) more than it indicates that our usual summers are becoming hotter.

richard verney
May 19, 2011 3:32 am

Interesting. Very good work.
Suggests that a significant rise in temps is nothing more than siting issues and a step change associated with the re-siting. I really have little faith in the cogency of the official temperature records and the validity of the adjustments made to them. And from this ‘corrupted’ (perhaps I should say bastardized) data we are supposed to draw the conclusion that the world is warming by miniscule (but yet said to be significant) amounts such that we should spend trillions of dollars addressing the issue.
I suggest that we get the raw data straight before we draw any conclusions still less spend trillions on what is likely to be a non proble.

Roger Carr
May 19, 2011 3:38 am

Journalist, or detective? Nice work, Chris. Many thanks.

Christopher Hanley
May 19, 2011 3:41 am

Apparently 1978 was notable year, climate-wise, at two (almost) antipodal points of the globe:

Bob in Castlemaine
May 19, 2011 3:42 am

Great work Chris. The BOM’s alarmist agenda exposed once more and what an opportune time to see their feet held to the fire.

Hector Pascal
May 19, 2011 3:45 am

Richard Verney is right. Perth itself is quite pretty, and very much a garden city. If (big if) you add enough water, you can grow just about anything. Without irrigation, you get bare sand and scrub.

Hector Pascal
May 19, 2011 3:54 am

Oops @ Stephen Harper.

Pete in Cumbria UK
May 19, 2011 4:02 am

It looks like the ‘Rules of Rounding’ have changed since I leaned them.
From my schooldays (40+ years ago), you rounded towards the even number/digit and not always upwards as seems to happen now. The rule was in the case of a final digit 5, to make the final digit an even number after rounding.
31.55 rounds to 31.6…upwards as nowadays
31.45 rounds to 31.4 ….downwards in my system but would be 31.5 in the present way of doing things.
Is it an ‘American’ vs ‘British’ thing or what…..

Ken Harvey
May 19, 2011 4:27 am

Look up at the skies tonight. There are a zillion things to see. Something that you will not see is an average. Not in the entire universe. An average, even the simplest, is an anthropological concept – an artifact if you like. No one was ever saved from frostbite nor heat stroke by a tempered average for the day. Old Mother Nature does not concern herself with averages. It is her extremities which bite.

May 19, 2011 4:32 am

The big problem here, as it is with the models, is that the standard deviation is not ‘carried’ through with the further processing. An average figure of just 23.5 for a month could mean every day of that month was 23.5, or half were 24.5 and the other half 22.5, or straight line trend going through 23.5 in the middle etc… without the standard deviation being carried through you have no idea how ‘significant’ or reliable that figure really is.
As for rounding on rounding – it just shows how much they are trying to squeeze significance out of figures when there really isn’t any to be had..

May 19, 2011 5:18 am

Hector Pascal:
“It’s more to do with the Indian Ocean Dipole than ENSO. I can’t find much in the way of meaningful information, but I believe there is a decadal trend towards increasing dryness. It needs more research: send me money ;-)”
That’s very interesting – the Indian Dipole. We should perhaps ask Earl Happ to look into this since he is in Western Australia (making fantastic wine) and has some very interesting things to say on such matters.
One of the things that residents (friends and family) were saying is that the minimum temperatures during summer were very high. But I think the main issue was consistently high temperatures day after day with very little change – a blocking we experienced here in Europe during December with the opposite effect. If the Indian ocean was very warm (and swimming off Rottnest I noticed it was particularly warm from just a year ago), there would be less differential between the inland desert and the ocean, meaning the sea breeze should have been weaker, and leaving slightly elevated and more humid (and therefore hotter) air.

May 19, 2011 5:35 am

Pete in Cumbria UK says:
May 19, 2011 at 4:02 am

It looks like the ‘Rules of Rounding’ have changed since I leaned them.
From my schooldays (40+ years ago), you rounded towards the even number/digit and not always upwards as seems to happen now. The rule was in the case of a final digit 5, to make the final digit an even number after rounding.
31.55 rounds to 31.6…upwards as nowadays
31.45 rounds to 31.4 ….downwards in my system but would be 31.5 in the present way of doing things.
Is it an ‘American’ vs ‘British’ thing or what…..

That is “bankers’ rounding”. It eliminates the slight bias towards higher numbers you get when rounding fixed decimal place numbers up to the next significant figure when a 5 is found. Half the time it is rounded down instead.
In nature, where numbers are not of a fixed number of significant figures, (often called ‘real’ numbers) then the bias should not exist. In practically every situation, there being an infinite number of numbers between any two numbers, if a 5 exists, there will always be some more ‘less significan figures’, so always rounding up is valid. For example 31.5 could be something like 31.500000065.
In accounting you pretty much always deal with a fixed number of significant figures, so need to adjust to avoid bias.
That is my understanding anyway, but I have only programmed it, not fully investigated it.
A remark by René Rhéaume, 21.09.01

This “Banker’s” method uses the Gauss rule that if you are in a perfect half case, you must round to the nearest digit that can be divided by 2 (0,2,4,6,8). This rule is important to obtain more accurate results with rounded numbers after operation.

Which I maintain is not applicable with when dealing with real numbers. In ‘real’ numbers, there is ‘perfect half case’. Nature does not do ‘straight lines’ or ‘perfect half cases’ IMO.

Stephen Harper
May 19, 2011 5:55 am

@Hector Pascal
“Oops @ Stephen Harper.”

May 19, 2011 6:01 am

@Stephen Harper – you’re right – this summer lacked any of those over 40C days – I have to say that because of that, I found this summer to actually be cooler than the few summers before it. We could use some more rain though..

May 19, 2011 6:34 am

As mentioned above by Ian, Perth does have a number of other nearby weather sites which could be used to produce a Perth area average (whatever that means). However, as always the devil is in the detail – how many of these are long record sites, unchanged since startup? For example Rottnest Island, some 20 km offshore, would have been ideal to get a long term trend from but its station was shifted in 1995. Perhaps Perth airport? (just kidding).
As regards the Mt Lawley site, I sometimes cycle past there – and yes it is unattractive vegetationless dirt. Even weeds struggle in Perth with lack of rain. It would be real interesting to have a duplicate station a few metres away with grass underneath for comparison. Dirt is noticeably hotter than grass during the day especially, which is why our houses have lawns.
OTT, it’s great to see Greens Leader Bob Brown squirming under the heat of the local press over his party’s carbon tax policy. Brown claims the press is being unfair because it gives both sides of the debate airtime, but then if we are being asked to pay billions in some new tax, he has to convince us of its merits. Which he can’t.

Philip Bradley
May 19, 2011 6:40 am

Speaking as another Perth resident, we have a particularly strong sea breeze known as the Freemantle Doctor. It causes a sharp drop in temperature when it arrives typically mid-afternoon on hot days, around the time the daily maximum typically occurs.
The timing of the sea breeze at a particular location is determined by its proximity to the ocean and to a lesser extent the Swan River. I believe the Swan River channels the breeze.
Annual comparison between sites even a few kilometers apart in the Perth Metro area will be meaningless unless the effect of the sea breeze is taken into account. And that is a difficult task.
I’d expect the sea breeze to arrive later at Mount Lawley than East Perth and hence Mount Lawley should have higher maximum temperatures.
There should be little variation in night time land breezes over central Perth and hence comparisons of daily minimum should not have the same locality issues as daily maximums.
Nice work, Chris.

Tim Folkerts
May 19, 2011 6:42 am

Interestingly, the first four months of this year ALSO set the record for mean maximum temperature at the nearby Perth Airport meteorological station (32.3 C), and this record also beat out 1978 as the next highest mean maximum (31.74). Here the record go back to 1944. And here the difference is quite dramatic. (These are averages of the daily values for highs from Jan 1 through Apr 30 each year)
I have no idea how the siting of the airport’s station may have changed, or how UHI might have affected the readings, but it does give some credence to the claim that 2011 was hottest in Perth. (The next hottest years for Jan-Apr are 2010, 1972, 1988, and 2009)
There are several other official stations in the area, but most only have very short time-frames (a decade or two). Still, it would be interesting to see if they also show 2011 as the hottest.

Tim Folkerts
May 19, 2011 6:54 am

Pete in Cumbria UK says: May 19, 2011 at 4:02 am
… you rounded towards the even number/digit …

Of course, the slightly better method is to keep a few more digits during intermediate steps and round at the end. The actual results for the months are:
JAN: 32.49
FEB: 34.15
MAR: 31.87
APR: 27.31
AVERAGE: 31.46
This DOES indeed round to 31.5!
Of course, the mistake now is that they should have done a weighted average. The shortest month (FEB) had the highest temperatures. By giving the shortest month the same importance as the longer months, they threw off the average. The true number is indeed 31,42, which should round to 31.4.

Hector Pascal
May 19, 2011 7:01 am

@ Stephen Harper
Sorry. I made a comment above mistakenly addressed to Richard Verney, agreeing with your comment that Perth doesn’t look like a sandpit. It doesn’t, Perth looks fine, as long as it is reticulated. The Mt Lawley site looks like a sandpit because it is not irrigated.

May 19, 2011 7:43 am

You will pardon me, but I have a hard time understanding why the mean of the maximum means anything. The time weighted average means something since that is a measure of the total heating/cooling requirements. The maximum or minimum mean something because they are a measure of the worst case stress on the heating and cooling. But the average of the maximum? A nice phrase for people that have no understanding of statistics I suppose, but unless somebody can illuminate me with a real use for this statistical nonsense, I must regard it as simply that; nonsense.

Tim Folkerts
May 19, 2011 9:47 am

ShrNfr says: May 19, 2011 at 7:43 am
>You will pardon me, but I have a hard time understanding why the mean
> of the maximum means anything. The time weighted average means
> something since that is a measure of the total heating/cooling requirements.
Yes, a time weighted average would be best. But most records only show daily max & min, so that is what we need to work with.
> The maximum or minimum mean something because they are a measure
> of the worst case stress on the heating and cooling.
OK. Sounds good.
>But the average of the maximum? A nice phrase for people that have no
>understanding of statistics I suppose, but unless somebody can
>illuminate me with a real use for this statistical nonsense, I must
>regard it as simply that; nonsense.
Well, if the daily maximum means something (as you said in the previous sentence), then knowing the typical value for daily maxima also means something. It is the simplest summary of a measure that you think is valuable. It would give a typical stress on cooling systems over the course of a few months.

John Gorter
May 19, 2011 10:03 am

I remember the Perth summer of 1990 well. Just moved to Perth, wife arrived to look at accommodation, air conditioning in rented house at Swanbourne broke down, airconditioning in new car broke down, visited Maylands area where we were thinking of living, temp 45 deg or there abouts for several days, wife says ‘I’ll stay in Sydney’ (luckily she didn’t), and I don’t think its been hotter since – but thats just my memories.
John Gorter

May 19, 2011 10:08 am

Chris, any way to get the monthly average minimums, maximums and relative humidity for those four month period for 1972, 1978, 1988, 2009, 2010, and 2011? Just curious if the amount of water vapor in the atmosphere had anything to do with this during warmer years. Or have you already followed that same line of thought?

May 19, 2011 10:26 am

Nick Stokes says: You can get the daily data for Mt Lawley from here. It is station 94608. For some reason it’s in Fahrenheit. But by averaging daily max I get, yes, 31.4C for the first 4 months of 2011. And 30.2C for the first 4 months of 1978.
Thanks for drawing my attention to that ncdc/noaa (WMO) data re Perth Metro. However, digging into their records makes me wonder if we’re talking about the same location. Station 94608, wherever that is, started in 1978 and the mean max for that full year was 24.2C. This compares to an annual mean max of 24.5C since the BoM’s Perth Metro (Mt Lawley 9225) began in 1994, and indeed the last couple of years were 25C and 25.3C.
However, the ncdc/noaa 1978 figures show an annual mean min of 15.1C. This compares to an annual mean min of 12.6C since the BoM’s Perth Metro (Mt Lawley 9225) began in 1994. I doubt it. The Perth Regional Office had an annual mean min of 14.7C in 1978, compared to an average 14.1C from 1968 to 1991 (post Observatory so the same location for a change).
I’ve compared the Perth Metro figures for 1994, although the ncdc/noaa data only show 346 days for that year. However, they seem to be fairly random missing days so:
9225 average annual max = 24.9C;
ncdc/noaa average annual max = 24.7C;
9225 average annual min = 12.9C;
ncdc/noaa average annual min = 11.9C.
It’s good to see the ncdc/noaa average min come down from 15.1C in 1978 to 11.9C in 1994!
How about 1995 from ncdc/noaa? Nah, they’ve only included 318 days of the year and that’s getting a bit too sparse for a reliable comparison.
1996? Eureka! Every day of the year:
ncdc/noaa average annual max = 24.4C;
9225 average annual max = 24.5C;
ncdc/noaa average annual min = 11.8C;
9225 average annual min = 13C.
I’ll jump forward to 2010 and give it a final roll of the dice:
ncdc/noaa average annual max = 25.3C;
9225 annual average max = 25.3C;
ncdc/noaa average annual min = 11.2C;
9225 av annual min = 12.4C.
Why don’t we have a look at another station in the Perth metropolitan area – Perth Airport, which is 94610 in ncdc/noaa-land and 9021 in BoM-land. For 2009 (I’d compare 2010 only the BoM hasn’t yet calculated last year’s December temp (?):
ncdc/noaa average annual max = 25.5C;
9021 average annual max = 25.5C;
ncdc/noaa average annual min = 11.1C
9021 average annual min = 12.3C
I’ve got no idea how ncdc/noaa has created a Perth Metro station dating back to 1978 as there’s no record of Perth Metro before 1994 on the BoM website. A cursory study of the ncdc/noaa Perth Metro data suggests it’s as reliable as an ice cube in an oven, and their Perth Airport figures also suggest they’re getting their temperature data from a dodgy operator – particularly their minima.
Agnostic – check for a variety of explanations re Perth’s temperature patterns and current dry spell. The BoM does attribute some of the past year to La Nina but, like Stephen Harper above (thanks for the pic, Steve!), I’m intrigued by the fact that the southern coastal sea surface temperatures off Western Australia have been well above average compared to the rest of the world for some time now… see Our west-east weather flows have to push southerly cold fronts across that warm ocean water if we’re to get rain, and the BoM weather outlooks usually show they’re on their way four or five days beforehand but give up as they get closer to the coast. The regularity and/or strength of our afternoon sea breezes might also be affected. I’m no expert on SST and I’ve got nothing to back up my belief except for what looks like an obvious explanation. Also note that across the 2.6 million square kilometres of Western Australia, this past summer recorded the highest-ever rainfall and below average maxima.
Tim Folkerts – I’m not sure if you’ve repeated the BoM rounding mistake but the average max of every single day in the four months was 31.4241666666667. That’s 31.4 when rounded.
Re Perth’s summer hitting 40C – we just got there on two occasions and there’s nothing particularly remarkable about that if you go through the BoM records. I also thought it was a comparatively cool summer because our “hot” days were mostly around 37C with none of those 42-43C scorchers, although a long run of warm minima above 20C made it a bit sticky at night. It was (almost) Perth’s record hottest summer even though it didn’t get particularly hot.
Thanks to all who’ve made complimentary comments.
BoM sources:
Perth Metro max annuals –
Perth Metro min annuals –
Perth Airport max annuals –
Perth Airport min annuals –

May 19, 2011 11:01 am

wayne says: Chris, any way to get the monthly average minimums, maximums and relative humidity for those four month period for 1972, 1978, 1988, 2009, 2010, and 2011? Just curious if the amount of water vapor in the atmosphere had anything to do with this during warmer years. Or have you already followed that same line of thought?
Nope, I haven’t looked into relative humidity. There are some longer term averages for Perth Metro at and you can track down temps, rainfall and solar exposure for each year back to 1992 at (search for Perth Metro at 9225 and Perth Regional Office at 9034) – but I don’t know where to find yearly or monthly rh stats.

Tim Folkerts
May 19, 2011 12:16 pm

Chris Gillham comments: May 19, 2011 at 10:26 am
Tim Folkerts – I’m not sure if you’ve repeated the BoM rounding mistake but the average max of every single day in the four months was 31.4241666666667. That’s 31.4 when rounded.
Actually, I was agreeing with you.
* If the average is done for the four months separately and those four numbers are averaged, then the answer would be just over 31.46, so it would very legitimately round to 31.5. But as I said, they should actually have done a weighted average, not a simple average.
* When the average is done the correct way, then I posted the same numbers above as you — that the correct average is 31.42, which could be rounded to 31.4 if you only want to keep one decimal place of precision.
The more significant point ( in a comment at May 19, 2011 at 6:42 am) was that the data from the Perth airport the first four months of this year ALSO set the record for mean maximum temperature. And the last three years are all in the top 6 hottest years.

Tim Folkerts
May 19, 2011 12:25 pm

Wayne asks: May 19, 2011 at 10:08 am
Chris, any way to get the monthly average minimums, maximums and relative humidity for those four month period for 1972, 1978, 1988, 2009, 2010, and 2011?
Weather Underground has daily data going back a while for various locations. I saw RH & Dew Point back about 10 years for Perth. But the data is presented for each day — I don’t think there is a way to download data for a month or year at a time, so extracting the data could be painful! And I don;t think you will find ’72, ’78, or ’88.

May 19, 2011 12:59 pm

Wooot! Perth on the map … yes it was a fairly toasty summer, but then I have only lived here since 1999. We get a fair few hot summers. As I live in the centre of town (near that red dot on the right actually) I shall be happy to take a few photos of those red dots as they are currently if you like.

May 19, 2011 4:35 pm

Quick question. If they rounded their previous average in the same way as the new one, then shouldn’t the difference be all that matters? I think your point about the rounding only matters if they changed method doesn’t it?

May 19, 2011 6:09 pm

But no-one in the rest of Australia cares about Perth.

Greg Beasley
May 19, 2011 6:44 pm

Perth – 2011 Record Summer Temperatures
When examining historic temperatures, one needs to consider: (1) whether the weather station has remained in the same location for the duration of the historic records; (2) whether the weather station site has been impacted by surrounding development (i.e., urban heat island effect or UHI), and (3) the impact of nocturnal cloud cover. Chris Gillham (WUWT, May 19) has given an excellent account of the consequences of a change in weather station location in Perth, Western Australia.
In 1947 Perth supported a population of 303,000. Today, the population of Perth has increased to nearly 1.7 million.
Now, Perth Airport is 10 kilometres east of the Perth CBD. Land was purchased to the south of Guildford and east of Ascot for the construction of the airport (YPPH) in 1938. The airport opened in 1940; initially as an RAAF base. A few years later (1944) the airport supported a mix of military and civil aviation. After WW2 the airport attained international status in 1952 and now comprises two intersecting runways and separate domestic and international terminals. Development of the airport, therefore, coincided with the post WW2 “Baby Boom” (1946-1963) and an influx of migrants from Great Britain and Europe.
In parallel, the Perth CBD shot up off the back of the Western Australian mining boom of the 1960s and 70s. At the same time there was an influx of new residents from the eastern states of Australia. Consequently, the airport that used to be on the outskirts of metropolitan Perth is now surrounded on three sides by residential and industrial development and to the east by remnant buffer of vegetation and more residential development.
A direct consequence of these changes is that temperatures in both the Perth CBD and at Perth Airport are now very much infected by UHI.
Weather station data at Perth Airport extends back to 1944, so we have a continuous record on how temperatures have changed from the end of WW2 right up to the present day.
The whole AGW argument is now largely being driven by a bias towards “warm” weather stations and continued increase in UHI in CBDs and areas of urban sprawl. This is also true of the Perth CBD and its nearby airport.
More recently, and in response to weakening sunspot activity, there has been a corresponding increase in low level cloud. This has exacerbated the nocturnal component of UHI. This is evident in the fact that, whilst the average daytime temperature at Perth Airport was 1º C above the long-term monthly mean maximum for January, the night time mean temperature for January minimum temperature was 3º C hotter than the corresponding long-term mean; 20º C cf. 17º C!

May 19, 2011 7:07 pm

Nice work, Chris!
I had the privilege to live in Perth from 1984 to 1988. Saw the America’s Cup won by Conner in Stars and Stripes 87, and managed the development and operation of a gold mine out near Wiluna. What a fantastic place that is!
Perth, in my mind, is one of the three most beautiful cities in the world: Perth, San Francisco, and Sydney (Vancouver is up there, too). Great place, great people, and for a geologist, great work. My very best to you folk down there.

Philip Bradley
May 19, 2011 8:07 pm

Interesting paper on rainfall changes on the GNANGARA MOUND (just outside Perth).
Main conclusion is that land use changes have had a greater effect on rainfall than GHG changes over the last century, and cyclical effects are the second most important.
They introduce the paper with,
Fully accepting climatic change
Clearly trying to protect themselves from AGW proponent backlash because their science doesn’t support the GW ideology.

Lawrence John
May 20, 2011 12:09 am

Great research.
Averages are dangerous, I always remind myself that if a man stands with one foot in ice and the other foot in boiling water, “on average” his feet are a comfortable 50 degrees.

May 20, 2011 12:55 am

Interestingly, while Australia’s southwest has had quite a hot summer, today’s “The Australian” newspaper reports that the snow has come early and for the first time in living memory, a Snowy Mountains ski field has turned its lifts on in May. Just goes to show how bad that Murdoch Press is, reporting something that doesn’t support the prevailing paradigm.

May 20, 2011 7:25 am

I wonder if you guys are perhaps all missing a point. The issue is not whether we have temperatures rising. I am missing here whether or not it is maximum temperatures that are driving up the mean temperature or whether or not it is the minimum temperature that is driving up the mean temperature.
I would love to hear all your comments on the balls that are currently on my pool table:
(you do not want to know how much work this little compilation was)
if you can add the Perth data in the same form as on my pool table, I would be much obliged.

Ian George
May 20, 2011 2:10 pm

Yet for Australia as a whole the temps have been down.
February was the 5th coldest Feb maxima on record.
March was the coldest maxima on record.
April was the 14th coldest.
And May has had one of its coldest starts so far on record.
When the La Nino and Indian DIPOLE come together as they did in 1974/75, Perth (and the SW corner of WA) will usually receive a hotter and drier summer while the rest of Australia has a cooler and wetter period.

May 20, 2011 4:16 pm

Did these records stand up outside the greater Perth area ?

Philip Bradley
May 20, 2011 4:41 pm

Rather an abrupt shift yesterday from warm humid sub-tropical air to cold Antarctic/Southern Ocean air.
So much so, it caused a tornado about 6 kms south of the Perth CBD. Very unusual for here.

May 20, 2011 5:27 pm

Nicely done! Instead of blindly turning the computational crank on the station data we are fed, stations records neeed to be vetted before using them for analysis of climate variations. Climatology should not be confused with urbanology.

May 20, 2011 7:02 pm

You can make colour contour maps of WA for actual max, min or mean T over various periods – 3 months being most useful. I am surprised colours are not further up the hot scale.
It is interesting to checkout what the BoM predicted.
For summer – max was supposed to be normal and min was hotter.
But for Jan to March there is a real bolt out of left field.
Max is predicted very cold. And min a bit warm. Rare to see such an intense cold prediction over such a large area.
How can the BoM model get things so EXACTLY wrong ?

May 20, 2011 10:34 pm

It looks as though the region around Perth outskirts was mostly not as hot as Perth – comparing the first 4 months of 2011 with 1978. If you click on my name you can see my new post with several comparisons.

May 21, 2011 2:49 pm

Monopole says:
May 19, 2011 at 6:34 am
You have to get completely outside the UHI area for meaningful indications of the strength of that effect. Comparing regressional trends at Perth R.O. and York WA showns the former rising ~.5K faster than the latter through 1991. This is roughly half the statistical average for 20-th century UHI enhancement for cities of Perth’s present size.

May 21, 2011 2:56 pm

My omission. That shoud read ” ~.5K PER CENTURY faster…”

May 21, 2011 3:47 pm

Regressional trends are highly time-dependent and I’m still messing up in conveying the contrast with York! From the beginning of the intact Perth R.O. record in 1897 through 1991 the trend difference relative to York is +1.2K/century. That’s pretty much in line with statistical expectations and an indication that record summer highs were unlikely in the much greater non-urban area surrounding Perth.

May 21, 2011 6:14 pm

Unbelievably sloppy work from the BoM. The people they employ in the climate section are becoming increasingly political and decreasingly scientific. I suggest the BoM makes all of its climatologists do some basic courses in statistics.

June 28, 2011 9:18 pm

dont matter what anyone says on these posts its just too darn hot for white people in a perth summer, in the end the heat and the dried up landscape are just too dam boring!

June 29, 2011 11:24 am

well, except God, there really is nobody to blame for the apparent warming there….

%d bloggers like this:
Verified by MonsterInsights