Guest essay by Bob Fernley-Jones
Our taxpayer funded broadcaster here in Oz, the ABC, has recently aired a scary story in its “science” TV show Catalyst, entitled Climate Extremes. (For the brave, the video and web page plus comments is here). For extra emphasis on TV there were repeated shots of several heat stressed old ladies on hospital stretchers, stacks of coffins (caskets), and much thunder and lightning etcetera.
But hey, there was this more technical screenshot sequence that rather caught my eye at about 3 ½ minutes in:
The narration elucidated how these bell curves clarified why the weather had become more extreme in the past decade, and, being a tad curious I searched around for the source, but without success. My closest find is contained in a report by our Oz government funded Climate Commission entitled The Angry Summer. (2012/3 DJF) This august body is headed by Prof Tim Flannery and amongst its expert advisors is Prof David Karoly, about whom I guess many overseas readers have heard?
But, engineers like me tend to be suspicious, and one thing I puzzled on was that the change in global average T of 0.80 C took place over a period of ~160 years according to HadCRU, and that the Catalyst show implied that the alleged effects were concentrated into the last decade.
I also worried that Carl Gauss would probably writhe in his 18th century grave to see his statistical normal distribution curve applied to an extremely complex system by our modern wizards of CAGW. In reality, the real distribution must be very lopsided around the global average of 150 C, (288K). And, the absence of units and scale also seems to be a touch odd and I pondered if it might err into an exaggeration. Putting aside the great media reverence for our Climate Commission, it disturbed me enough to compile this stuff below:
I found that not all weather extreme bell curves are equal. For instance, David Karoly is senior co-author in a recent paper entitled; The human role in our ‘angry’ hot summer, which uses a different concept. It employs some modelling projections and is based on a regional sample of Australia only and summer only, (and maybe with distribution closer to normal?). Perhaps it could be fun for anyone who might wish to explore it. Karoly’s retracted Gergis et al SH hocky-stick paper and his conclusions on a biology paper of modelled 10-day early emergence of Melbourne butterflies based on Avalon Airport temperature history, etcetera, do not fill me with confidence though!
Above is an interesting variation in figure 1.9 on page 44 of the 2nd draft AR5 IPCC report. It is a combination of two other bell curves (a) and (b).
Strangely, the Climate Commission’s claim of their source as from AR4 (S. Solomon et al) is not found in the relevant chapter 3. (E.g. 3.8.2 Evidence for Changes in Variability or Extremes – not there)
Part 2:Some more economies in material facts and relevance:
Back to our recent TV show:
NARRATION: This past year in Australia, we’ve seen plenty of heat. At the Bureau of Meteorology, forecasters have been watching record after record tumble.
The first figure from our Climate Commission below left is compared with data from our BoM. Although the BoM have apparently discarded some hotter old records, still yet, NONE of the remaining State or Territory records were broken in the angry summer according to the current BoM table, (right, modified to fit):
But, back to our TV show again quoting a wise doctor of authority in CAGW:
Dr Karl Braganza [BoM]: January was the hottest month on record. The summer was the hottest on record [in Oz]…
The next two graphs give the BoM time-series records for January and summer monthly maximum average temperature data for the crucially important Murray-Darling (rivers) Basin (MDB) “food bowl”. There was nothing special about summer in that vast area covering substantial parts of four States; a map is available here. (BTW, Oz is roughly the same size as the contiguous USA). According to the BoM the volatility of monthly past extremes in the MDB was greater than in the last decade. The master link is here and it has drop-down menus that enable visit to other regions and seasons.
|
So what about the individual States and Territories of Oz? Well here they are (treating the Capital Territory or ACT as part of New South Wales or NSW) for the reportedly hottest month of January, but cut off short for compactness and to ease hotness comparison:
|
|
For more commentary on the earlier unbroken temperature records map above right see this, and here is another reference which includes UAH satellite data for Australia also showing the so-called “Angry Summer”, as rather ordinary.
The drop-down menus enable research of the BoM time-series variously by regions and seasons or months, and my conclusion is that monthly average temperature extremes were greater in the past, and so too was monthly volatility. It would take much space here to demonstrate that but the drop-down menus provide the capability for those that may be interested to research it.
Part 3:Breaking the mood with something almost amusing:
I’ve also submitted a wider ranging formal complaint to our taxpayer funded ABC, concerning the bias and other stuff in this story, (the ABC is required by statute to serve the public, and breached its own editorial policies). I closed the complaint off with this:
Ms [Anja] Taylor was the declared presenter, producer and researcher for this show. She presented almost entirely extreme views with an apparent lack of investigative journalism. With the exception of Dr Fischer’s input about the warming effect of dry soils, (which is not controversial if we ignore Prof David Karoly), all other topics were either demonstrably false or controversial. Not content with presenting scientific material facts and balance, (the Editorial Policies require impartiality), she adds inappropriate drama and irrelevance including these images:
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
vukcevic wrote: ”In case anyone is interested to know what distribution of the world longest temperature record looks like: http://www.vukcevic.talktalk.net/CETd.htm”
Yes, it is interesting and a bit more towards normal distribution because of the small average annual temperature range in central England, and of course the extremes are smoothed out by the annual averaging.
Here’s another grave consequence of global warming in Australia. 🙂
“The sun is shining, the flowers are in bloom – and undertakers everywhere are tearing their hair out. Sydney’s unseasonably mild winter, the warmest on record, might be a joy for most of us but it’s making life tough for funeral directors, who are experiencing their slowest season in 25 years.
”We’ve seen the biggest drop in business in a generation,” said Andrew Smith, chief executive of InvoCare, the largest private funeral, cemetery and crematorium operator in the Asia-Pacific region. ”Winter is usually our busiest time, but there’s been no real flu season this year and no real cold snaps, and that’s being reflected in a big drop in business.”
Read more: http://www.smh.com.au/national/business-dead-as-undertakers-feel-heat-from-a-warm-winter-20130822-2segz.html#ixzz2d4CSANBo“
Hooray! Someone who can distinguish between “not all are” and “all are not”.
Bob Fernley-Jones says:
August 28, 2013 at 4:31 pm
vukcevic wrote: ”In case anyone is interested to know what distribution of the world longest temperature record looks like: http://www.vukcevic.talktalk.net/CETd.htm”
Yes, it is interesting and a bit more towards normal distribution because of the small average annual temperature range in central England, and of course the extremes are smoothed out by the annual averaging.
>>>>>>>>>>>>>>>>>>>>
The distribution of means will approximate a normally distribution. Central Limit Theorem
(There is that sample size of 30)
ferd berple says: “One of the problems in using averages – and more often anomalies – is that the processing that goes into creating the average and anomalies can create statistical trends where no such trends exist in the underlying data. For example, in general what the temperature records show is that minimum temperatures are increasing, but maximum temperatures are not.”
Not surprising. It’s difficult to shift the right hand part of the distribution curve further to the right, since any surface rise sheds heat to space by T⁴, and the right side already has a much higher T than the left.
As shown by Gumbel in his monograph on “Statistics of Extremes,” their distribution is not likely to be gaussian.
Bob,
Do you (or anyone else) know how the BOM calculate its daily/monthly temp averages? Do they use only the data from the ACORN stations or do they average every w/s in Australia?
And when they do compare temperature averages to get ‘an angry summer’, do they only compare data with w/s that have been operating from 1910 (or do they include some recently installed w/s that have no long-term records)?
Both January 1896 and January 1939 seemed to have some very hot weather but don’t feature in their literature. For instance, Bourke had 17 days in a row of +40C in 1939 (though ACORN has readjusted these temps down) and 22 days of +40C in 1896. We didn’t see that this year.
Ian … Australia ACORN-SAT temperature fraction audit at http://www.waclimate.net/round/australia-acorn.html explains a few of the ACORN quirks, particularly temperature rounding before and since decimal conversion.
vukcevic, copy to Gail & 1sk1:
I see that your CET distribution; http://www.vukcevic.talktalk.net/CETd.htm has an average annual temperature of 9.2 degrees 22 times as averaged from 1659-2012 and that despite the small CET spatial coverage and an annual average temperature range of only about 4 C, (which surprises me*), there is nevertheless a pronounced longer tail on the cold side and throughout the series there are erratic jumps. For instance; 9.8 degrees has 18 events below maximum and 9.9 degrees has only 7 events in the 353 years.
* The annual averaging eliminates outliers though.
Gail & 1sk1;
Concerning distribution of global average temperatures, there are complex temporal and spatial variations, but what is generally described as THE global average (or if you prefer; mean) is ~15 C or ~288K. My main point is the simple observation that various sources give spatial (zonal) annual averages as up to about 30 or 35 C on the hot side of average, and down to very, very much colder the other side. Thus, using annual averages the cold tail MUST be very much longer than the hot side and thus not Gaussian. Throw-in the unsmoothed outliers and it becomes even more pronounced. Here is another source that gives a greater range of average temperatures to that I cited in the essay, but it does not damage the concept:
http://www.physicalgeography.net/fundamentals/7m.html
I don’t have a big acceptance of some statistical theories and prefer to leave it to the experts whom in my opinion sometimes get their nickers in a twist. For instance, although Steve McIntyre et al did BRILLIANT work with statistical stuff and whatnot on the Manna hockey-stick, I think it complicated the issue. There were some six much simpler ways of debunking it that might have been more easily understood by lay people.
Did you know that about seven (?) years ago there was an alarming cluster of breast cancers in staff at the Brisbane studios of the ABC? Despite extensive testing etcetera, no explanation of cause could be found. Nevertheless, the staff moved to a new location to evade the unknown.
Catalyst (Jonica Newby) also made a big thing last year in her story; “Measuring Our Temperature” of a cluster of over 200 deaths of an endangered Cockatoo mostly on a golf course near Hopetoun WA on a very hot day. It was enthusiastically spruiked on ABC radio in The Science Show declaring the new normal of worsening climate change. However, various evidences give that whilst the single hot afternoon may have been a tipping factor there must have been some other prime cause such as disease.
Many states built, now mothballed but still being paid for, billion dollar sea water desalination plants based on Flannery’s predictions about “larger more common and severe drought”. Almost all dams throughout the country are pretty much full. We had severe flooding in 2010/2011 (Not as severe as the 1894, 1905 and 1974 Queensland floods). We also now have Karoly stating sea level rises, and that is GLOBAL sea levels, have reversed because of the big wet in Aus in 2010/2011.
Two days ago, all across the media we were towards the end of the warmest winter on record. Then it was the warmest in 150 years. Now it’s the warmest 3rd or 4th winter on record. My guess is they haven’t a clue beyond their salaries and pensions.
1sky1 says:
August 28, 2013 at 7:52 pm
As shown by Gumbel in his monograph on “Statistics of Extremes,” their distribution is not likely to be gaussian.
>>>>>>>>>>>>>>>
Yes it runs into the Willis’s Thunderstorm Hypothesis except in the deserts.
waclimate says: @ur momisugly August 28, 2013 at 11:34 pm
Ian … Australia ACORN-SAT temperature fraction audit at http://www.waclimate.net/round/australia-acorn.html explains a few of the ACORN quirks, particularly temperature rounding before and since decimal conversion.
>>>>>>>>>>>>>>>>>
Jo Nova had quite a bit to add to that too. Australian Temperature Records: Shoddy, Inaccurate, Unreliable – Surprise/
Bob Fernley-Jones says: @ur momisugly August 29, 2013 at 12:10 am
vukcevic, copy to Gail & 1sk1:
I see that your CET distribution…. there is nevertheless a pronounced longer tail on the cold side and throughout the series there are erratic jumps…. * The annual averaging eliminates outliers though.
>>>>>>>>>>>>>>>>>>>
That is the Central limit Theorem in action. The more points averaged per mean the more the curve tends towards a normal distribution. However you are going to get that tail if the curve for the original data is overwhelmingly lopsided.
Sort of emphases the problems for climate are going to probably be from the cooler side not the warmer.
Its kind of funny how records can be set when old data is discarded or how much warmer the present seems when you keep changing past temperatures to make them cooler! Whether its ACORN or the ASOS in the US that supposedly discovered a “cold bias” in older temperatures but somehow resulted in them being adjusted DOWNWARD, its scary they’re in charge.
I had that dilemma for the first time on Jan. 21, 1985. I didn’t have a registering thermometer and had woken up well after dawn, when temperatures had risen from their lows. The day was record cold and at my house could have been as low as -10F, which would be the coldest I’d ever recorded. I thought about it, but ended up entering -6F in my records as the low because it was the official reading at the local (but usually warmer) airport, because I didn’t feel right putting a number there I couldn’t “prove.” A similar process the officials are taking, but they’re fudging the #’s.
You’d have to forgive my moral dilemma back in ’85…after all I was only NINE YEARS OLD!
Bob Fernley-Jones & Gail Combs
………….
Hi
By plotting 1 degree C moving average, a better visual impression of the ‘CET bell curve’ is obtained
http://www.vukcevic.talktalk.net/CETd.htm
Thanks waclimate and Gail
I’m still at a lost as to how the BOM determine Aust average temps and what it compares that data with to come up with any conclusions. I ran the raw and ACORN data for Bourke (Jan 1939) and found a great deal of adjustment. I know they adjust temps with nearby w/s but can’t find any reason for the huge downward adjustments.
raw ACORN
38.9 38.4
40.0 39.1
42.2 41.9
38.1 37.9
38.9 38.4
41.7 41.5
41.7 41.5
43.4 43.0
46.1 45.7
48.3 47.9
47.2 46.8
46.2 45.8
45.7 45.3
46.1 45.7
47.2 46.8
46.7 46.3
40.0 39.1
40.1 39.1
40.0 39.1
41.9 41.7
42.5 42.1
44.2 43.8
36.7 36.5
40.3 39.2
36.6 36.5
29.4 29.5
29.3 29.4
28.8 28.9
30.6 30.5
35.6 35.4
38.6 38.3
48.3 47.9
28.8 28.9
40.4 40.03 Jan average
Note: All temps over 30C have been adjusted down, below 30C adjusted up. No rhyme nor reason.
Ian George,
Yes the clockwise rotation of time series temperature records is quite distasteful. GISS is perhaps the most famous as a long term “creeping correction”, and there was all that fuss over the NZ stuff. Chris de Freitus wrote an article here:
http://www.nzherald.co.nz/opinion/news/article.cfm?c_id=466&objectid=10833106
Funny how in different regions and whatnot the data in the past gets corrected downwards even when one would think we know less about it. It has the effect of reducing the cold tail on the distribution curve.
Karoly and Flannery are due to become past Australian History, along with their alarmism, after the forthcoming election which will sweep their unscientific detritus out of public memory.
Patrick wrote:
…Two days ago, all across the media we were towards the end of the warmest winter on record. Then it was the warmest in 150 years. Now it’s the warmest 3rd or 4th winter on record…
Yes, the media likes hot stuff. I was watching ABC TV news as July closed this year and our trustworthy weatherman informed us two nights in a row that July had been the hottest in melbourne on record, or something like that. Eh? I thought. I didn’t notice living where I do in a NE suburb. I’ve just checked again by asking Google:
melbourne + “hottest july”
and it gave 15,800 hits including substitutions such as warmest……………………….Hmmmmm.
An interesting comparison is that the BoM records show considerably higher temperatures than in 2013 for Victoria for July since 1975, for both mean and maximum:
Whoops,
I forgot the link for July temps:
http://www.bom.gov.au/cgi-bin/climate/change/timeseries.cgi?graph=tmax&area=vic&season=07&ave_yr=5
Bob
According to BOM, Melbourne RO had its hottest July at 15.9C beating 1975 by 0.2C.
However, Melbourne Airport had 14.7C whereas in 1975 it was 15.1C there.
So, at worst, Melbourne has had an 0.2C raise in July temps in nearly 40 years.
http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_display_type=dailyDataFile&p_nccObsCode=122&p_stn_num=086071&p_c=-1481742221&p_startYear=2013
I had to XL the data to get the average as they haven’t done it yet.
Other data bases for Melbourne have a mixture of Reg Office and the Air Port. I presume the ones I have here are all from Reg Off.
Ian,
Yes, an increase of 0.2C since 1975 is really scary eh? That’s especially so since 1975 was at the end of a global cooling cycle starting at around 1940. So then we had that great warming phase when not a lot happened in Melbourne.
It gets more farcical though and I nearly composed an analysis for Anthony Watts to review. I don’t know if you are familiar with Melbourne but it has a rather substantial east-west main road, Victoria Street, which narrowly bypasses the traditional CBD which has a grid pattern of streets, and branching off from Victoria Street at an acute angle is Latrobe Street. It busily traverses the north side of that city grid. There is thus this narrow tiny triangle of land where the Melbourne “Regional” Office instrumentation is located between much tarmac.
Towards the bottom of the following BoM PDF is a layout of the site at Aug 2008 and in earlier years. In reality it should be named City Office, not Regional.
http://www.bom.gov.au/clim_data/cdio/metadata/pdf/siteinfo/IDCJMD0040.086071.SiteInfo.pdf
I wanted to go down there with a camera around midday at the time when it was sunny to see the latest situation but failed to do so. It seems to be a classic UHI situation to me going by the maps.
Bob
Just checked it out on Google maps and, as you say, it does look like ‘a classic UHI’ site.
In your link to the station metadata it says it was opened in 1908, but records for the Reg Off go back to 1855. So was this the original site or was it moved?
Ian,
I think the BoM Melbourne Regional (= City) Office is a can of worms and wouldn’t contemplate any BoM data prior to 1910, given that that is when they start their time series graphics.
Melbourne temperatures are also rather contorted in the context of alleged AGW because of the notorious effect of wind direction and duration. We are the butt of jokes on this matter and wind direction can almost be estimated from the temperature data. An extreme example was on 7/Feb/2009 when we had tragic bushfires in the NE hills. After some days of hot north winds, (a heatwave), we broke max temperature records in Melbourne with ~46C on that afternoon, but come about 5:15 pm a cold front came through and T dropped to ~29C in a matter of minutes. About 17 degrees in minutes….. (Perhaps around ten degrees is more normal)! (Incidentally, the change in wind direction was also disastrous in the bushfire area, in a combination of bad stuff)
Interestingly, the latest Melbourne RO 2008 site-plan shows an anemometer right on the corner between the two busy roads.
http://www.bom.gov.au/clim_data/cdio/metadata/pdf/siteinfo/IDCJMD0040.086071.SiteInfo.pdf
However they don’t report it nowadays but instead use wind data from Essendon Airport presumably because of issues with turbulence not only from nearby large structures and trees, but also increasing road traffic around the rather enclosed RO site.
Groan. You’ve inspired me to try and get down there tomorrow with a camera.