The Bureau Boss on Temperature Trends, Heatwaves, and Climate Change

March 31, 2015

On Sunday profile on ABC Radio on Sunday 29 March, the Director and CEO of the Bureau of Meteorology was interviewed.  The whole interview is here:

For a scientist who claims to be only interested in science and not in advocacy, he certainly sounds like a fervent Global Warming Enthusiast.

Here is the feedback I sent to the Bureau.


Dear Dr Vertessy

I was interested to listen to your interview on Sunday Profile on ABC Radio yesterday, 29 March 2015. I was particularly interested in your comments regarding public criticism of the Bureau’s adjustments to temperature data, and on the increasing frequency in heatwaves.

Several times you stated that the adjustments “make no difference at all” to temperature trends, that the raw temperature data “tell exactly the same story”, and that we see “the same result (in temperature trends) for the whole continent” as for raw data.

You also stated that heatwaves are becoming “one of our most serious natural disasters”.  They are “a bit of a silent killer- it’s the number one cause of death.”  You also said, “We are probably seeing of the order of five times as many very serious heatwaves today as we did in the middle of last century.”

I have some questions.

Q.1: Can you please supply me with a reference to your data that show that the number one cause of death is heatwave?  I was sure it was cardio-pulmonary disease usually associated with very cold weather, with mortality rates much higher in winter than summer.  Perhaps you meant heatwaves are the number one cause of death when compared with other natural disasters, which is debatable.  This was not at all clear and must surely have misled some listeners.

Q.2:  Can you please supply me with a reference to your data that show five times as many very serious heatwaves today compared with the middle of last century?  Could you also please tell me your criteria for a very serious heatwave.

For the next question I refer you to Table 1 on page 14 of On the sensitivity of Australian temperature trends and variability to analysis methods and observation networks  (CAWCR Technical Report No. 050), R.J.B. Fawcett, B.C. Trewin, K. Braganza, R.J Smalley, B. Jovanovic and D.A. Jones , March 2012 (hereafter CTR-050).  This shows that quadratic change in mean annual temperatures from 1911 to 2010 in adjusted data of the ACORN-SAT network (+0.94C) is 36% greater than in the ‘unadjusted’ data of the AWAP network (+0.69C). For maxima, the change is 38.9%, and for minima is 34.1%.  In this paper the authors claim that the rise in unadjusted data is “somewhat smaller” than in ACORN-SAT.

Q.3:  In what way can 38.9%, 36%, or 34.1% difference in quadratic change be interpreted as “no difference”, “exactly the same story”, or “the same result”? 

Perhaps you should have told your listeners that the similarity was only since 1955, and that before this, raw data show temperatures (especially maxima) were cooling, but then 60 years is not such a long climate record for making trend analyses, and this may be confusing to those who cannot understand more than a simple climate narrative.

In the Concluding Remarks of CTR-050, p.50, the authors state that “further work will be undertaken to characterise in more detail these changes, particularly at the monthly and seasonal level”.

Q.4:  When can we expect to see the results of this further work published on the ACORN-SAT website?  If it is available elsewhere please refer me to it.  I am particularly interested in any difference in quadratic change in summer maxima between AWAP and ACORN-SAT, as this is relevant to heatwave analysis.

I look forward to your reply.”

For an explanation for my interest in comparison with AWAP data, see my analysis of monthly and seasonal differences in trends between AWAP and Acorn from October last year.  My calculations indicate a 200% increase in trend in summer maxima.

One might think that if Australia wide there has been a five-fold increase in the number of very serious heatwaves, there should also be some discernible increase in the number of very hot days.

To illustrate my incredulity about this claim, here is the timeseries graph of very hot days (BOM definition: >40 degrees Celsius) straight from the Bureau’s website:

 Hot days graph BOM

The linear trend (for what it’s worth) shows an increase of 0.02 days per decade.  That’s 0.2 of a day per hundred years, or 2 days in 1,000 years.  Scarey hey.

I will be following up on the hot days and heatwaves analysis in coming posts.

First Meeting of Technical Advisory Forum

March 27, 2015

The first meeting of the Clayton’s Review Panel was held yesterday.  The press release about the proceedings is here.

“The Technical Advisory Forum held its first meeting today to advise on part of Australia’s official climate record – the Australian Climate Observations Reference Network Surface Air Temperature (ACORN-SAT) data set.

Today’s discussions were robust and productive, and the Forum would like to thank the Department of the Environment for managing the Forum’s membership and assisting the Chair. The Forum would also like to thank the Bureau of Meteorology for providing information and answering questions on the ACORN-SAT data set.

The Technical Advisory Forum was appointed by the Parliamentary Secretary to the Minister for the Environment, the Hon Bob Baldwin MP as an independent advisory body to provide greater transparency and an impartial framework for quality assurance tests and analysis of the Bureau’s data sets.

In line with its Terms of Reference, the Forum discussed three aspects of ACORN-SAT: the extent of the public availability of the ACORN-SAT information; developments since the 2011 Independent Peer Review of the Bureau’s data and analysis methods; and the scientific integrity and robustness of the Australian climate record and the homogenisation process. The Forum considered a broad range of information and data concerning the management and development of the ACORN-SAT data set provided by the Bureau.

The Forum also received some additional information from members of the public about the data set. Members of the Technical Advisory Forum were appointed to provide advice on the basis of their formal expertise, and the Terms of Reference do not therefore require the Forum to receive public submissions. However, this additional information was provided to all Forum members to ensure that Forum members were aware of public concerns regarding the Bureau’s management of ACORN-SAT during their deliberations.

The Forum will deliver its report, including detailed recommendations, by June 2015.”

I’m not holding my breath.

How Hot Is Brisbane? With new, ‘improved’, daily benchmarks!

March 15, 2015

Recently I briefly posted about Brisbane, using 29.5 C as the heatwave bench mark.  Silly me- after checking Dr Perkins’ site (Scorcher) I now understand that Dr Perkins was not calculating a single bench mark temperature for a heatwave.  She calculates benchmark temperatures for every single day of the year.  Temperatures in the top 10% for each date for three days in a row constitute a heatwave- including in winter.  Thus Brisbane can have a heatwave in mid-July if the temperature is over 22.4 degrees for three days.  Some heatwave!  A real scorcher!  Only one winter ‘heatwave’ has ever been over 30 degrees (33.7 C in 2009).

It is interesting that Dr Perkins’ website, Scorcher! has this caveat on the front page: is intended as an educational site. It should not be used for any policy decisions and comes with no warranty.”

However, silly as it sounds for winters, I decided to attempt to replicate Dr Perkins’ heatwave findings for Brisbane.  For the years 1961 to 1990, I sorted each month’s days from hottest to coldest, and found the value for the top 10 percent of each month.  I then used this value for each month, and calculated a centred 31 day running mean to give a benchmark value for each day of the year.  This value is used to flag days hot enough, and if there are three or more in a row, there’s your ‘heatwave’.  Rough and ready, and probably a very different method to Dr Perkins’, but values for heatwave benchmarks should be similar- if anything, a little low.  Results are graphed below.

Fig.1: Seasonal count of heatwave days, showing that yes, the number of heatwave days has increased.

hw days seas bris

Fig.2: Spring count of heatwave days (Benchmark ranges from 24.8 to 30.1 C)

 hw days spring bris

Fig.3: Length of spring heatwaves

spr hw length bris

Fig.4: Temperatures of spring heatwaves

spring temps daily bm brisbane

Fig.5: Number of heatwaves in spring

 cnt spr heatwaves bris

So yes, by this definition it is correct to say in spring Brisbane has slightly more heatwaves, which have more heatwave days, are slightly longer, and very slightly warmer, if you trust linear trends as an analysis tool for such non-linear data- (without the linear trend line and equation, it is difficult to see any trend in heatwave days, duration, or temperature).

However, the results for other seasons are also revealing.

Fig.6: Temperatures of summer heatwaves  (Benchmark ranges from 30.1 to 30.9C)

summer temps daily bm brisbane

Fig.7: Temperatures of winter ‘heatwaves’ (Benchmark ranges from 22.4 to 24.7 C)

 winter temps daily bm brisbane

Fig.8: Temperatures of autumn ‘heatwaves’ (Benchmark ranges from 24.3 to 30.2 C)

autumn temps daily bm brisbane

Winter, summer, and spring all have slightly more and warmer heatwave days (summer heatwaves are 0.2 degree warmer after 65 years, winter ‘heatwaves’ will be half a degree warmer in 900 years!), but autumn is declining in all areas- length, temperature, and number.   Funny that Dr Perkins didn’t mention Brisbane autumns.

And the most important thing about Brisbane heatwaves is not included in Dr Perkins definition- humidity.

ACORN updated, but not improved. Result: the same bogus data.

March 14, 2015

Early this month the ACORN-SAT dataset was updated to 31 December 2014.  A quick check shows that despite much public criticism, nothing has changed since March 2012 when it was first released.  There have been no corrections.

There are still thousands of days of missing data.

There are still gross errors.

Adjustments remain, and so still have the effect of making many individual sites LESS comparable with neighbours, and of making national trends much greater than those for raw data.

However, in this post I am merely focussing on one of the many erroneous features in Acorn, that of the instances of minimum temperatures exceeding maximum temperatures.

Why bring this up again, as it was first spotted in 2012?  Because it is more evidence of unjustified adjustments, and Acorn’s authors have done nothing about it.

Although a specific check for errors in recording raw maxima and minima was conducted before homogenising, this check could not have been done with the homogenised data. It might be claimed that this feature is normal and due to a cold change arriving after 9.00 a.m.  This would be especially evident in winter at high altitudes such as at Cabramurra, with 212 occurrences.  If so, it would have been evident in the raw data.  However, a check of 22 of the sites, comprising 84% of the instances (803 of 954) reveals there are NO instances of maximum less than minimum in the raw data for Cabramurra, or anywhere else.  All instances occur in the adjusted data.

Fig. 1: Min > Max in Acorn (803 of 954 instances)

 minmax Aust

 Fig. 2: The same test for raw data at the same sites as Fig. 1

minmax Aust raw

 Further, despite the Bureau being aware of the problem since at least 1 July 2013 when Blair Trewin, lead author of ACORN, assured readers of the blog Open Mind (sic) at that “in the next version of the data set (later this year), in cases where the adjusted max < adjusted min, we’ll set both the max and min equal to the mean of the two” (which merely hides the fault caused by adjustments), the problem still exists- 212 occurrences are still in the ACORN record for Cabramurra, and there are 954 in total.  Trewin has done nothing about it- mainly because he would have to redo the whole dataset to do it properly.  The problem exists because adjustments have been too vigorous and too erratic.  The algorithm produces bogus data.  Merely averaging maxima and minima on those 954 days hides the underlying problem.

Here is one example, the worst I have found, (circled in Fig. 1), from Alice Springs, 15 June 1932.

Fig. 3: Raw vs Acorn at Alice Springs

Alice june32

Notice there are NO adjustments in Acorn to raw maxima, but Acorn minima mostly runs along one or two degrees less than raw- except on 1 and 2 June and most notably on 15 June, when Acorn minima shoots to 4.8 degrees above the maxima for the day!  This is caused by an adjustment of 6.4 degrees to the raw temperature.  The Bureau lists the following sites as neighbours used for this adjustment:  Boulia, Tibooburra, and Tennant Creek.  Ignore for the moment that these are all hundreds of kilometres from The Alice, just visually compare their data.

Fig. 4:  Alice vs ‘neighbours’

Alice june32 v neighbours

Tennant Creek and Boulia, hundreds of kilometres north and northeast, are usually several degrees warmer than The Alice, except when Acorn makes those huge adjustments.  Which is the outlier?

This Alice Springs instance is a rogue adjustment due to a leading “1” being inserted; 18.1 entered instead of 8.1.  This is not uncommon in Acorn and crops up repeatedly.

While this is annoying and farcical, it can be easily explained and easily fixed.  The real problem is with data that is not so grossly erroneous, such as at Cabramurra.  This graph shows raw minima and maxima for May and June 1962.  Note that often maximum is not much above minimum: there is little room for error.

Fig. 5: Cabramurra raw

raw Cabramurra

But see what happens with adjustments.

Fig. 6: Cabramurra raw and adjusted

raw vs adjusted Cabramurra

Maxima have been adjusted down, minima up, without any thought for the consequences, so four times in May and June 1962 minima exceed maxima.  Merely averaging maxima and minima for those days will hide the symptom of the underlying disease.  The adjustments cannot be justified because they result in nonsense data.  Exactly how do you explain negative Diurnal Temperature Range?

The problem of a tiny percentage of the record featuring this error is not minor:  it indicates that many adjustments are too great, and also that no quality assurance checks have been performed, and therefore many other types of errors are likely to exist.  And they do.

Now how is it that my laptop can find these errors, whereas the Bureau’s supercomputer can’t?

Here is the full list of sites and instances of bogus data:

Station, Number of days with minimum temperature exceeding the maximum temperature.
Adelaide, 1. Albany, 2. Alice Springs, 36. Birdsville, 1. Bourke, 12. Burketown, 6. Cabramurra, 212. Cairns, 2. Canberra, 4. Cape Borda, 4. Cape Leeuwin, 2. Cape Otway Lighthouse, 63. Charleville, 30. Charters Towers, 8. Dubbo, 8. Esperance, 1. Eucla, 5. Forrest, 1. Gabo Island, 1. Gayndah, 3. Georgetown, 15. Giles, 3. Grove, 1. Halls Creek, 21. Hobart, 7. Inverell, 11. Kalgoorlie-Boulder, 11. Kalumburu, 1. Katanning, 1. Kerang, 1. Kyancutta, 2. Larapuna (Eddystone Point), 4. Longreach, 24. Low Head, 39. Mackay, 61. Marble Bar, 11. Marree, 2. Meekatharra, 12. Melbourne Regional Office, 7. Merredin, 1. Mildura, 1. Miles, 5. Morawa, 7. Moree, 3. Mount Gambier, 12. Nhill, 4. Normanton, 3. Nowra, 2. Orbost, 48. Palmerville, 1. Port Hedland, 2. Port Lincoln, 8. Rabbit Flat, 3. Richmond (NSW), 1. Richmond (Qld), 9. Robe, 2. St George, 2. Sydney, 12. Tarcoola, 4. Tennant Creek, 40. Thargomindah, 5. Tibooburra, 15. Wagga Wagga, 1. Walgett, 3. Wilcannia, 1. Wilsons Promontory, 79. Wittenoom, 4. Wyalong, 2. Yamba, 1.

(From Willis Eschenbach at . Butlers Gorge has a further 37 days.)

Here is a plot of their occurrence- the big lump after 1960 is due to Cabramurra.

Fig. 7: Running 365 day count of min > max (803 of 954)

count incl cabramurra

Letter to Bob Baldwin re Acorn Inquiry Terms of Reference

March 13, 2015

Dear Mr Baldwin

I am shocked, horrified, and angry after reading the Terms of Reference and Format of the Technical Advisory Forum.
The proposal that this august body can investigate the many flaws in ACORN-SAT in a one day sitting once a year is beyond ridicule, beyond parody.  It is sillier than the funniest episodes of “Yes, Minister”.
It is apparent that it has been set up not to investigate but to verify the Bureau’s claims.  It is a political solution to a perceived embarrassment.  It has been set up to appease the Bureau’s critics by seeming to examine their concerns without actually doing so, and without upsetting the Bureau or the ALP or Greens.  It will be a whitewash.
The Technical Advisory Forum needs real teeth.  It should be able to call witnesses, examine documents and analyse data.  Instead it will listen to platitudes from the Bureau in the morning and have a discussion in the afternoon.  What a joke!
Meanwhile, ACORN-SAT has been updated to include data to 31 December 2014.  It has not been improved.  It still has thousands of days of missing data, 954 days when minimum temperature exceeds maximum, still has gross and obvious errors anyone can recognise, still over-adjusts raw data so that ‘homogenised’ data is LESS comparable with neighbouring sites, and still results in Australia wide trends that are much greater than those shown by raw data.
I assume that you have acted on advice from your department.  You have been hoodwinked.  And unless you swiftly act to rewrite the Terms of Reference and Format for the Technical Advisory Forum to give it some real investigative power, this will be added to the list of bad decisions of a gutless government, more concerned about the opinions of the ABC, the press gallery, and the cross benches than its once loyal and hopeful supporters.

Yours in disgust
Ken Stewart

Limited Blogging for a While

March 6, 2015

Due to family reasons, I can’t give much attention to this blog for a while.

I’ll leave you with two graphs to contrast climate alarmism about heatwaves with reality.

Dr Sarah Perkins of the Climate Change Research Centre and Dr Karl Braganza were in the news this week claiming that

“we’re definitely seeing more heatwaves across Australia”


“the number of days that belong in a heatwave each season — has been increasing since the 1950s.”


“Brisbane has been getting heatwaves in spring”

Let’s look at Brisbane.  Dr Perkins, a heatwave expert,  defines a heatwave as “three days in a row of temperatures in the top 10 per cent”.  The top 10% of days are those with a temperature above 29.5 degrees Celsius.  Yep, 29.5 C is apparently an extremely hot day at Brisbane Airport!  This shows consecutive days above 29.5 C- the red line is the 3 day heatwave benchmark.  Note the linear trend.

spring consec 29.5 brisbane

This shows the highest temperatures reached in 3 day heatwaves:

spring temps 29.5 brisbane

Just two graphs refute all of the above claims about Brisbane:

definitely NOT more heatwaves

definitely NOT more hot days each season since the 1950s

and Brisbane has ALWAYS had heatwaves in spring, but not so many nowadays.

There has been a slight increase in the number of standalone days above 29.5 C (0.17 days in 66 years), but ‘heatwaves’ have been getting COOLER.

Enough said.

How Hot is it in Boulia? (Hotter Than News Limited Can Remember)

March 3, 2015

Yesterday I posted about the ABC’s short term memory about heat at Longreach.  Today it’s Boulia’s turn, this time due to a News Limited article, “Relentless, endless summer. The Queensland town where it’s been 40 degrees pretty much forever”. 

Time for a reality check.

Using the ACORN dataset and current data for Boulia up to 2 March, here is a 180 day running count of days above 40 degrees in Boulia.  (This captures days in one summer, but not two).

Fig. 1: 180 day running count of days above 40 C at Boulia.

boulia 40

Even with 43 missing days of data in December and January, this summer is not unusual.

The next graph shows the length of 40+ heatwaves.

Fig. 2: Consecutive days above 40 C at Boulia.

boulia 40 consec

This is definitely NOT the longest ever heatwave.  This one is very mild- 31 days in 1973 is going to be hard to beat.  Note that there is NO TREND in the length of heatwaves.

Finally, this graph shows the number of days above 45 degrees, which is hot in anyone’s language.  The missing data may have had some 45+ days, but there were many days with rain in this period.

Fig. 3: 180 day running count of days above 45 C at Boulia.

 boulia 45

It’s hot in Boulia, but not unusually so.

Tomorrow I will look at the latest claims about Australia getting hotter and cities getting more heatwaves.

How Hot Is It at Longreach? (Hotter than the ABC can remember)

March 2, 2015

Today on ABC Radio’s Country Hour, there was an interview by Lydia Burton with a grazier from the Longreach district in Central Western Queensland, which has been having some pretty hot weather this summer, as has most of Western Queensland.

Peter Whip is from Royston near Longreach and says over the last 50 years he has noticed an increase in the amount of days over 35 degrees and that is a real concern for all landholders.

“In the last couple of years we have had some really long hot dry spells when the feed quality has been terrible and we have seen cattle conditions slip really quickly.

“That is certainly something that is changing over time and….it is something that in 20 years time it will be a significant impact.”

The gist of the segment was that cattle were struggling in the heat, with so many days above 35 degrees Celsius, and especially with days over 40 degrees, let alone 45 degrees.

As I have said before, I have close links to agriculture.  My parents, grandparents, and great-grandparents were farmers; one brother still farms, another is retired, and my brother-in-law is a farmer also.  As well, several nephews and nieces are on the land or in agri-business.  I have great sympathy for those on the land at all times, especially in drought conditions.

However, is there any evidence that Longreach is having an unusually hot spell, or that Longreach is getting hotter, or that it will get hotter in the future?  And how could an intrepid rural reporter check the facts against one grazier’s anecdote?

The very simplest check would be on Longreach’s mean maximum temperatures.  From November to February, the mean monthly maximum is never below 35 C, so we might expect many days to be well above 35 C.  Longreach Aero’s highest ever maximum was back on 28 January, 1990, at 47.3 C.

Let’s have a look at the number of days above 35 C, 40 C, and 45 C.  I’ll do this for 180 day periods, which will capture days in a single summer, but not two.  I use the official ACORN dataset, plus daily data right up to yesterday, 1st March, straight from the Bureau of Meteorology, so there can be no accusation of cherry picking or of using “unreliable” raw data.

Fig. 1:  Number of days in summer above 35 C at Longreach.

longreach 35

And for the last 50 years….

Fig. 2:  Number of days in summer above 35 C at Longreach since the summer of 1964 -65.

longreach 35 since 6465

What about the length of heatwaves?  The next plot shows the number of days in a row where the temperature has exceeded 35 C.

Fig. 3:  Number of consecutive days above 35 C at Longreach.

longreach 35 consec

Mr Whip may be a little mistaken.  Since 1965, and since 1910, there appears to be no trend in the number of days above 35 C in Longreach, and there is no trend in the length of heatwaves.

But the ABC is always talking about extremes- what about days over 40 C?

Fig. 4:  Number of days in summer above 40 C at Longreach.

longreach 40


Fig. 5:  Number of days in summer above 45 C at Longreach.

longreach 45

Oops again.

Now Lydia Burton is a good rural reporter, I frequently listen to her on the Country Hour and the Rural Report, but if only she had checked the facts, she could have saved her listeners a lot of angst.  Instead, “our” ABC continues to peddle nonsense about dangerous climate change.

It’s no wonder that people have lost faith in the ABC.

An “Eye Witness” Account of TC Marcia

February 28, 2015

A week after the event, now that our power has been restored, internet seems back to normal, and our cleanup is mostly complete, I can turn to an analysis of Tropical Cyclone Marcia.  Much initial analysis has been done, especially by Jen Marohasy at  who was in contact with me soon after the cyclone, and also by the Cyclone Testing Station of James Cook University.

However, perhaps I can add to the discussion with some additional suggestions and some personal observations.  I was there.  The eye passed over our house for about 25 to 30 minutes.  I have felt the impact in Rockhampton, experienced the electricity supply loss for 6 days and 6 hours, and cleaned up tree damage to our own 1.5 acre block, the local Heritage Village and a nursing home where a family member is resident.  Close family members were also without power.  I can also relate the experience to previous cyclones when I lived in the Mackay area.

Tropical Cyclone Marcia was a very small, intense, and short lived Complex Cyclone, with maximum winds about 30 – 40 km away from the centre, and well outside the eye wall.  At all times the Bureau of Meteorology was reporting its strength as at least one Category (more likely two) above what ground observations showed.  There is no surface evidence to show it was ever Category 5.  The only surface observation available indicates Category 3.

I watched on various websites the approach of TC Marcia throughout the day on Thursday 19th, and every couple of hours that night (few people got much sleep), and for as long as I had landline and then mobile internet connection (and power in my laptop) on Friday 20th.

Marcia rapidly intensified during Thursday.  By 5.30pm it was near Creal Reef and about 175 km  northeast of Mackay with a clearly visible eye on weather radar.

marci 17301902

The Tropical Cyclone Technical Bulletin reported it as having central pressure 958 hPa, 10 minute winds of 155 kmh and gusts to 220 kmh.  By 10.22pm it had central pressure 940hPa, winds 195 kmh with gusts to 270 kmh, and had slowed to 7 kmh but heading southwest, straight for Mackay.

By 3.20 am on Friday it was just to the east of the Percy Islands, where Middle Percy Island recorded winds from the south at 145 kmh gusting to 180 kmh.  The maximum winds here were reached between 4.00 and 4.30 am when winds reached 156 kmh gusting to 208 kmh.  Central pressure was 971.6 hPa.  It is interesting to note that the apparent temperature was 1.3 degrees C- it is very cold in a cyclone if exposed to the wind.

marci 03002002

Strongest winds and heaviest rainfall are usually in the eyewall.  Notice that the heaviest rain in the above image (yellow) is to the west of the eye- in the vicinity of Middle Percy Island.  This is what made me strongly doubt the forecast strength.  The Technical Bulletin reported it as having central pressure 929 hPa, winds 205 kmh gusting to 285 kmh, moving south at 13 kmh- straight towards Rockhampton.

The next image is from 6.00 am, about two hours before landfall, travelling south at 30 kmh.

 marci 06000220

Again note the heaviest rain outside and to the west of the eye wall.  The eye is about 25 km across.

Here is an interesting model of winds around Marcia when close to crossing the coast at 7.00 am on,-20.38,1821.

 winds 07002002

Strongest winds are shaded red.  Interestingly, they are to the northwest and northeast of the eye.  (Note also the deep trough from the Sunshine Coast to the border.)

After this I was busily checking final preparations, having breakfast, and making and receiving phone calls, as well as watching the websites.  By 10.00 am the wind was becoming much stronger, and the Bureau was still saying it was Cat 5 and heading straight for Rocky.

marcia forecast 10000220

Here is the radar image:

marci 10100220

Note the yellow area of heavy rain (and presumably strongest winds) south of Samuel Hill and close to Byfield.

The wind strengthened and was at its strongest by about 12.30 pm, consistently from the east.  There was one final prolonged gust, the strongest, at about 12.35 and then from about 12.50 the wind died away to nothing by 12.55 pm.  Like many others I went outside and had a look around, but didn’t go far.  It was raining lightly, heavily overcast, with puffs of wind from different directions.  At 1.25 pm we could hear the wind coming, and it picked up again over about five minutes, blowing directly from the west, but didn’t reach the same strength as before.  At a speed of 22 kmh, the eye was about 11 km across at this stage.  This screenshot shows the eye right over us at about 1.10 pm.

marci 12500220

The eye is just visible, and strongest rain and winds are to the south and east.

But this one shows it just 10 minutes later:

 marci 13000220 eye collapse

Note the eye has disappeared- it collapsed while over us.

Rockhampton Airport is about 12km south-southwest, so 30 minutes later the eye should have reached there.  But it didn’t.  Here is a screenshot of observations at Rocky Aero from 12.41 to 4.30 pm.

rocky eye

I have circled key information.  At 12.50 pm the wind was ESE at 72 kmh, with a gust to 113 kmh.  Between 1.13 and 1.52 (arrowed) the wind swung from SE to WSW, but never dropped below 24 kmh.  Lowest pressure was 975.7 hPa.  The wind steadily picked up and was strongest around 2.45 to 3.00 pm, reaching 82 kmh with a gust to 113 kmh.  This was different from my experience, where the wind was consistently from the east, then very light and variable, then directly from the west, but not as a strong.  The eye had collapsed before the centre was near the airport, and the remnants passed a couple of kilometres to the east.  The cyclone was rapidly degrading.

My last screenshot before the internet died and my battery went flat, shows the BOM forecast at 2.12pm:

marci forecast 14120220

It shows Marcia still Cat 3 at 2.00 pm.  But Rocky Aero data show winds were no more than Category 1 at this time.  At our place, the winds at this time were from the west and still damaging trees but not nearly as strong as before the eye passed.  By 3.00pm the winds were light and we were out in the street talking with the neighbours.

On eastern slopes or hills, and closer to the coast, winds would have been much higher.  I estimate gusts at our place, funnelled up a gully on an east facing slope, were around 120 – 130 kmh- similar to our experience of TC Ului at Sarina some years ago.

The Cyclone Testing Station of James Cook University has released a preliminary report at .  They will produce more definitive analysis in a few weeks.

To summarise, this report states Yeppoon experienced Category 2 winds, and Rockhampton Category 1.  The report goes on to state:

“A community that receives an over-represented wind speed report may have potential for complacency in preparation or building standards in the future. Further research is required to determine if this is an issue.”

But was TC Marcia a Complex Cyclone, one in which maximum winds are concentrated away from the eye?  All radar images show heaviest rain, and therefore probably winds, away from the eye wall.

Byfield community is 30 km northwest of Yeppoon, half way between Yeppoon and Samuel Hill.  There are reports of every tree stripped for kilometres, which appears to indicate strong Cat 3 to Cat 4 winds.  Radar image at 10.10 am shows heaviest rain just to the west of Byfield and just to the south of Samuel Hill.  Yet Samuel Hill did not experience winds of this strength.

And here’s a photo of part of our block taken at 3.30 on the afternoon Marcia passed.


Note every tree has at least 50% of its leaves stripped and several branches off, and several were completely uprooted, and these are tough old ironbarks.  So allowing for 30 km closer to the sea and two hours earlier, I wouldn’t be surprised if winds were 150 – 160 kmh at Byfield- but not 200 kmh.

Until JCU can do their complete assessment, we will not know exactly what happened.  No doubt their engineers will also visit Stanage Bay, where Marcia crossed the coast.  Tropical Cyclone Marcia was a very small, intense, and short lived Complex Cyclone, with maximum winds about 30 – 40 km away from the centre, and well outside the eye wall.  At no time was it Category 5.  Despite its relative weakness compared to some other cyclones, Marcia did tremendous damage in Central Queensland.

The main thing is no one was hurt.

My Submission to the BOM Review Panel

February 9, 2015

The Hon. Bob Baldwin MHR

Parliamentary Secretary to the Minister for the Environment

PO Box 6022

House of Representatives

Parliament House


Dear Mr Baldwin

Re: Recommendations for the Review Panel appointed to review official national temperature records

As a “citizen scientist” who has been researching Australia’s climate, and the ACORN-SAT record in particular, over the past several years, I am concerned about errors in the work done by the Australian Bureau of Meteorology, in particular warming bias introduced through homogenisation and the apparent general lack of quality control.   I draw your panel’s attention to these issues listed under four categories in the following submission.  More information, including supporting charts and tables, is provided in numbered attachments.

 1Adjustment Issues

 1.1 Homogenisation distorts temperature records causing warming bias at most locations, indicating the methods as stated in the CAWCR Technical Reports (see CTR-049) are not followed or do not work as designed. Homogenisation should lead to candidate sites having trends in temperature anomalies that are more like their neighbours’. However in many cases this does not occur, and homogenisation has resulted in wide disparities.  This is obvious from a simple visual inspection of a plot of ACORN data at candidate sites versus raw data of the listed neighbours (Attachment 1.1).

 1.2 A better but still simple method of comparison involves differencing. Differencing (anomaly data of candidate site minus data of reference sites) should show improved results following homogenising, with differences closer to zero.  Importantly, even if the differences fluctuate, there should be zero trend in differences. Yet at a number of sites, homogenising has produced worse results. (Attachments 1.2a, 1.2b, 1.2c).

 1.3 The most extreme examples of homogenising arising from warming adjustments, result in trends of candidate sites greater than the homogenised trends at neighbouring ACORN sites. (Attachment 1.3a). The converse applies where extreme cooling adjustments result in trends less than the homogenised trends of ACORN neighbours (Attachment 1.3b). This indicates over correction, resulting in the creation of artificial trends warmer or cooler than the neighbours’.

1.4  Data are homogenised by reference to up to 10 best correlated neighbours. Some of these neighbours may be hundreds of kilometres away, and with completely different climates.  Deleting the two most distant neighbours greatly improves data comparison between Mackay and its remaining neighbours. (Attachment 1.4a, 1.4b).

1.5 The Bureau has belatedly tried to explain adjustments with the release of the 28 page PDF file of all adjustments ( , and has also provided Summaries of Adjustments for six sites as a further explanation. ( ). Incredibly these Summaries don’t agree with the adjustments in the 28 page document. (Attachment 1.5).

1.6 The Bureau claims that sites exhibiting Urban Heat Island (UHI) effect are “excluded from downstream products such as the calculation of national and regional temperature anomalies for the analysis of large scale climate change” (CTR-049, pp.71-73.) These sites include Townsville, Rockhampton, Laverton RAAF, Richmond NSW, Sydney, Melbourne, Adelaide, and Hobart.  Unfortunately they certainly are used as comparison sites when making adjustments.  Several Queensland sites, including Cairns, Charters Towers, Mackay, and Bundaberg, have Townsville and/or Rockhampton listed as neighbours used for making adjustments.  If a site’s temperatures are suspect due to UHI to the extent that they cannot be used for regional or national anomalies, it seems illogical that they can be suitable for comparison with neighbours.  This apparent contradiction needs explanation.

 2. Impact on trends

2.1 The Bureau of Meteorology has reportedly claimed “an extensive study has found homogeneity adjustments have little impact on national trends and changes in temperature extremes.”  (Weekend Australian, August 23-24, 2014). In support of this, the Bureau displays a plot on the adjustments tab at the ACORN-SAT web page purporting to show “temperature trends since 1910 from the unadjusted temperatures from more than 700 locations (AWAP), together with those that have been carefully curated, quality controlled and corrected for artificially induced biases at 112 locations (ACORN-SAT)” ( However, the AWAP (Australian Water Availability Project) network is not “unadjusted”, but according to CTR-050 p.41, ‘the generation of stable climatologies implicit in the AWAP … ( analysis) … goes part of the way towards removing the temporal inhomogeneities implicit in the raw data without the explicit application of temporal-inhomogeneity adjustments. … Hence  it is reasonable to describe the AWAP … (analysis) as “partially homogenised” rather than unhomogenised.’  It is therefore misleading to describe the above-mentioned plot of ACORN vs AWAP as being a comparison with “unadjusted temperatures”.

2.2 Moreover, the Bureau has made no attempt to compare ACORN data with minimally adjusted raw data (that is, adjusted only to combine two incomplete records into one through examination of overlapping data.) My comparison of annual ACORN data (1910-2012) with raw records (corrected only for overlap) at 83 sites for minima and 84 for maxima shows the increase in trend of ACORN over raw is 66% and 13% respectively (Attachments 2.2a, 2.2b).  (The remaining sites had no suitable overlap between discontinued and new stations).  Nearly two thirds of the sites analysed had trends increased (warmed).

2.3 A few very remote sites especially in Northern and Central Australia have an enormous impact on the ACORN record. This is demonstrated by differencing area averaged means (the official national annual means) and straight averaged means of the 104 ACORN sites (Attachment 2.3).  As 7 – 10% of the national climate signal is due to Alice Springs alone (International Review Panel Report, September 2011 p. 12, both the influence of remote sites, and the area averaging algorithm, need to be investigated.

2.4 The South East portion of Australia (the area south and east of the median of ACORN stations’ latitudes and longitudes) has the greatest number of sites, and also the greatest change in trend in minima from raw to ACORN- 232% (Attachment 2.4a). The trend increase for New South Wales is 245% (Attachment 2.4b) and for Victoria is 350% (Attachment 2.4c).  Homogenisation in the most heavily populated areas of the country cannot be described as having “little impact”.

2.5 I have also compared network wide ACORN data with AWAP data (1911 – 2013), for annual, seasonal, and monthly analyses (the last two the Bureau has not yet completed.) The results (Attachments 2.5a, 2.5b, 2.5c, 2.5d, 2.5e) are staggering and need urgent investigation.  In particular, the 200% increase in trends for Summer Maxima is relevant to claims of increasing summer heat, especially in light of the recent Climate Institute report.

3. The effect of rounding on trends, and uncertainties in neighbouring stations’ data used for homogenising

3.1 The Bureau admits that rounding of temperatures to whole degrees in the Fahrenheit era may have led to an artificial breakpoint of +0.1C in 1972, but claims this is lost in the noisy signals of the 1970s (CTR-049 p.70). This, however, is disputable and needs thorough investigation, as an audit by several colleagues and myself of daily minima and maxima data in the Fahrenheit and Celsius eras (474 data records and 8,580,583 daily observations) found evidence that indicates the impact on trends could be between +0.1C and +0.4C. (Attachment 3.1).

3.2 Our study also found homogenising is based on records with large amounts of uncertainty (Attachment 3.2). Significantly, the first of the 2011 International Review Panel’s Recommendations (A1) was “Reduce the formal inspection tolerance on ACORN-SAT temperature sensors significantly below the present ±0.5 °C …“ (See  It is inexcusable that no error bounds are given for ACORN data.

4. Lack of Quality Assurance

The apparent lack of quality assurance means ACORN-SAT is not fit for the purpose of serious climate analysis including the calculation of annual temperature trends, identifying hottest or coldest days on record, analysing the intensity, duration, and frequency of heatwaves, matching rainfall with temperature, calculating monthly means or medians, and calculating diurnal temperature range.

4.1 ACORN-SAT daily data (and consequently, monthly and annual means) provide many obstacles to rigorous analysis. Days of data are missing, slabs of data are offset by one day (daily data being assigned to the wrong date, usually one day early), and many adjustments show obvious glaring errors (Attachment 4.1).

4.2 Another glaring error gives Australia a new hottest day on record (Attachment 4.2).

4.3 Other researchers have reported at least 917 days where minimum temperature exceeds maximum (Attachment 4.3). Although a specific check for errors in recording maxima and minima was conducted before homogenising, this check could not have been done with the homogenised data. It might be claimed that this feature is normal and due to a cold change arriving after 9.00 a.m.  This would be especially evident in winter at high altitudes such as Cabramurra, with 212 occurrences.  However, there are no instances of maximum less than minimum in the raw data for Cabramurra.  All instances occur in the adjusted data before February 1999.  Further, despite the Bureau being aware of the problem since at least I July 2013 when Blair Trewin, lead author of ACORN, assured readers of the blog Open Mind at that “in the next version of the data set (later this year), in cases where the adjusted max < adjusted min, we’ll set both the max and min equal to the mean of the two” (which merely hides the fault caused by adjustments), the problem still exists- 212 occurrences are still in the ACORN record for Cabramurra.

In conclusion, ACORN-SAT is not reliable and should be scrapped.  ACORN-SAT shows adjustments that distort the temperature record and do not follow the stated procedures in the Bureau’s own Technical Papers, generating warming biases at a large number of sites, thus greatly increasing the network wide trends.  Furthermore, the Bureau does not take account of uncertainty, and the data are generally riddled with errors indicating poor quality assurance.  Finally, its authors have not followed up on most undertakings made more than three years ago to permit replication and improve transparency ( ).

I am delighted with the formation of the Review Panel.  I hope that this review will bring about much needed improvements at the Bureau of Meteorology in the way the Bureau collates, audits, analyses and reports on national temperature data.

Yours sincerely

Ken Stewart


 Attachment 1.1:  One example of many – comparison of Acorn anomalies (black) with neighbours at Carnarvon

(For further information and full explanation see )

Attachment 1.2a:  Differencing: Rutherglen minus neighbours

(For further information and full explanation see ) 

Attachment 1.2b:  One example of many- Differences after homogenisation show worse results

(For further information and full explanation see )

Attachment 1.2c:  As for 1.2b, showing the mean of candidate data minus neighbours. 

(For further information and full explanation see )


There are many other sites with greater differences after homogenisation in minima or maxima.  I have also checked Deniliquin, Bourke, Amberley, Carnarvon, Williamtown, Mackay, Kalgoorlie-Boulder and Wilcannia, which all show this problem.  Some sites may show improved differences.  An audit of all sites is essential.

Attachment 1.3a: One example of many of warming adjustments over correcting-  Amberley Acorn vs nearest Acorn neighbours’ (mean of nearest Acorn neighbours’ homogenised data).

(For further information and full explanation see )


Attachment 1.3b:  An example of cooling adjustments over correcting- Acorn Tarcoola shows decreased trend compared with nearest Acorn neighbours’.

(For further information and full explanation see


 Attachment 1.4a: Mackay maxima differencing including all listed neighbours

(For further information and full explanation see )


Attachment 1.4b: Mackay differencing with 2 most distant neighbours excluded, showing improved differences.

(For further information and full explanation see )


Attachment 1.5: The Bureau’s lists of adjustments at six stations are different- except one at Orbost.

(For further information and full explanation see


Attachment 2.2a: Mean of Tmin annual anomalies at 83 sites- minimally adjusted raw data vs Acorn, 1910 – 2012 data. 

(For further information and full explanation see  )


Attachment 2.2b: Mean of Tmax annual anomalies at 84 sites- minimally adjusted raw data vs Acorn, 1910 – 2012 data. 

(For further information and full explanation see  )

Attachment 2.3: Differencing shows the effect of area averaging using very remote sites, 1910 – 2012 data.

(For further information and full explanation see  )


Attachment 2.4a: Tmin increase in trend in different regions, 1910 – 2012 data. 

(For further information and full explanation see ).

  Median network position map adj results

Attachment 2.4b: Increase in Tmin warming in NSW, 1910 – 2012 data.

(For further information and full explanation see ).


Attachment 2.4c: Increase in Tmin warming in Victoria, 1910 – 2012 data.

(For further information and full explanation see ).

  Vic chart

Attachment 2.5a: ACORN vs AWAP comparison- by month (1911 – 2013 data)

(For further information and full explanation see


Attachment 2.5b: ACORN vs AWAP comparison- annual and seasonal (1911 – 2013 data)

(For further information and full explanation see

summary table seasons

Attachment 2.5c: ACORN vs AWAP comparison- by season- Tmean (1911 – 2013 data)

(For further information and full explanation see

mean table seasons

Attachment 2.5d: ACORN vs AWAP comparison- by season- Tmin (1911 – 2013 data)

(For further information and full explanation see

  min table seasons

Attachment 2.5e: ACORN vs AWAP comparison- by season- Tmax (1911 – 2013 data)

(For further information and full explanation see

tmax table seasons

Attachment 3.1: Comparison of percentage of observations recorded in values from whole (rounded to .0) to 0.9 in the Fahrenheit era, with that of the Celsius era at continuing sites, with irrelevant sites deleted, indicating suitable conditions for creation of artificial warming.  (For a full discussion see ).

Attachment 3.2:  Analysis of the impact of rounding on trends and uncertainties.  (For a full discussion see ).

Our study concluded:

“As more than half of all sites in Australia had rounding probably greater than 50%, truncating at significant levels (33%, 50%, or 100%) before September 1972 would cause artificial warming of between +0.1C and +0.4C per 100 years.”

“Many …. sites have recorded large amounts of data in recent  years that may be in error by up to 0.50Celsius, being rounded to whole degrees, and more than half of the sample studied have recorded erroneous data at some time in the past 40 years.”

“As well, the vast majority of sites … inaccurately recorded observations in the Fahrenheit era by recording in whole degrees. For nearly half of all sites, this amounts to at least 50% of their total observations. It is probable that more than 50% of all Australian observations were rounded. This alone means that temperatures before 1972 may be inaccurate by up to 0.250 C.”

“The large amount of uncertainty in the records of so many sites means that homogenisation as practised by BOM researchers must be in question, and with it all analyses of Australia’s temperature trends.”

Attachment 4.1: Obvious errors indicate poor quality assurance.

There are numerous glaring errors for individual days at many sites.  The following graphic shows Rutherglen maxima at Climate Data Online for September to November 1926 compared with ACORN-SAT maxima from 30/09/1926 – 05/11/1926.

cdo v acorn max ruth oct26

Also on 13/10/1926, Acorn minima records -1.2 (adjusted down from +6.9).

Data for days such as 13/10/1926 with an obvious error, possibly the result of a missing leading digit, are not unusual and are found in the records of many stations.

The Acorn record for Rutherglen has some other peculiarities as well.  There are several separate periods where Acorn’s maxima record frequently does not match with data from Climate Data Online, and is one day too early.  These are:

1/11/1920 – 19/3/1940,

1/12/1940 – 31/10/1944,

1/5/1946 – 31/10/1947, and

1/12/1947 – 31/1/1948.

Attachment 4.2: Another glaring error- one of many

Australia’s hottest temperature is supposed to be 50.7C recorded at Oodnadatta on 02/01/1960, but ACORN-SAT has a temperature of 51.2C at Albany on 08/02/1933.  Many days have been adjusted by more than +6 degrees C, resulting in this ludicrous figure which has passed quality assurance.

  albany max 1933

Attachment 4.3: List of 69 stations with ACORN minima exceeding maxima.

Station, Number of days with minimum temperature exceeding the maximum temperature.

Adelaide, 1. Albany, 2. Alice Springs, 36. Birdsville, 1. Bourke, 12. Burketown, 6. Cabramurra, 212. Cairns, 2. Canberra, 4. Cape Borda, 4. Cape Leeuwin, 2. Cape Otway Lighthouse, 63. Charleville, 30. Charters Towers, 8. Dubbo, 8. Esperance, 1. Eucla, 5. Forrest, 1. Gabo Island, 1. Gayndah, 3. Georgetown, 15. Giles, 3. Grove, 1. Halls Creek, 21. Hobart, 7. Inverell, 11. Kalgoorlie-Boulder, 11. Kalumburu, 1. Katanning, 1. Kerang, 1. Kyancutta, 2. Larapuna (Eddystone Point), 4. Longreach, 24. Low Head, 39. Mackay, 61. Marble Bar, 11. Marree, 2. Meekatharra, 12. Melbourne Regional Office, 7. Merredin, 1. Mildura, 1. Miles, 5. Morawa, 7. Moree, 3. Mount Gambier, 12. Nhill, 4. Normanton, 3. Nowra, 2. Orbost, 48. Palmerville, 1. Port Hedland, 2. Port Lincoln, 8. Rabbit Flat, 3. Richmond (NSW), 1. Richmond (Qld), 9. Robe, 2. St George, 2. Sydney, 12. Tarcoola, 4. Tennant Creek, 40. Thargomindah, 5. Tibooburra, 15. Wagga Wagga, 1. Walgett, 3. Wilcannia, 1. Wilsons Promontory, 79. Wittenoom, 4. Wyalong, 2. Yamba, 1.

(From Willis Eschenbach at .  Another study  claims a total of 954 days.)


Get every new post delivered to your Inbox.

Join 55 other followers