Was 2013 the Hottest Year on Record? Update!

January 6, 2014

Update:  Warwick Hughes has reminded me of his post on 5 December at http://www.warwickhughes.com/blog/?p=2496 where he shows a distinct drift in UAH data compared with RSS, and in later posts he confirms this in southern Africa and the USA.  Warwick says:

"I have checked UAH against CRUT4 and GHCN CAMS for all Australia and it
looks like there was a drift in UAH 2005-2006.

Until UAH resolves the issue, I think their ranking of Australian hot
years is not worth repeating."

That may help explain the large divergence in recent years.  

I will leave this post as is, with the caveat that it is based on available UAH and Acorn data.


On Friday, 2 January, the BOM released its Climate Statement claiming 2013 as the hottest year on record.

The UAH dataset for lower troposphere temperatures has also been just released.

I have compared BOM monthly data with UAH by converting the BOM anomalies to the same reference period as UAH (1981-2010).

Here is the result:  UAH vs BOM 1978-2013 (12 month running means)uah v bom

It is plain to see that in the satellite era, Australian surface temperatures (as calculated by the BOM) reached a record last year.

For the 12 month periods to December, UAH agrees that 2013 was the hottest, just ahead of 1998 and 2009.

According to UAH, the 12 months period to October 2013 was just edged out by the 12 months to June 2010.

So, the BOM is right in saying 2013 was the hottest on their 104 year (and very much adjusted) record.

While the two datasets match reasonably well in most years, especially 1996-1999, they diverge markedly in recent extreme years.  It appears that the BOM area averaging algorithm accentuates extremes, probably because of the scarcity of observing sites in the remote inland, where warming and cooling are much greater.  Alice Springs, for example, being hundreds of kilometres from the nearest neighbouring site, contributes 7 – 10% of the national warming signal.

As well, the satellites’ remote sensors do not necessarily match the atmospheric conditions at ground level, depending on different seasonal conditions.  However, to quote Dr John Christy, “the temperature of the lower troposphere (TLT) more accurately represents what the bulk atmosphere is doing – which is the quantity that is most directly related to greenhouse gas impacts.”

So- if you are interested in the weather, how hot it is locally, consult the BOM- the old Weather Bureau.  If you are interested in whether the climate is changing due to greenhouses gases, consult the satellite data.

And yes, the weather has been hot (and still is where I live).

What did Chris Turney expect?

January 5, 2014

Professor Turney did not have to take an unsuitable ship full of “climate tourists” to Antarctica.  He could have just checked the Bureau of Meteorology’s website.

As the Aurora Australis will be calling at Casey base to deliver delayed supplies before returning the hapless Turney and the rest of the expedition to Australia, I thought I’d help with what conditions to expect at Casey.  I used official ACORN-SAT monthly data to 2011 and Climate Data Online daily temperatures since then.

Here are the actual monthly maximum temperatures at Casey for 2013:Casey max 2013

As you can see, temperatures were below the mean (calculated from 1970-1990) for most of the year, and the monthly mean maximum temperatures were above freezing (the straight blue line) only in January and December.  Monthly mean minimum temperatures never get above freezing.   (The highest daily minimum in 2013 was +1.7 C on 15 January.  The warmest minima this summer were on 29 and 30 December.  It got to +0.3.)

And has there been recent warming?

This graph is of maximum and minimum anomalies from the 1970-1990 means, smoothed with running 12 month means:Casey 1970-2013

Australia has three bases on the Antarctic coast, Casey, Davis, and Mawson.  Davis and Mawson show some slight warming:Davis 1958-2013Mawson 1958-2013

The mean anomalies of all three sites:Antarctic means

show a linear trend of about  +0.15 C- but the rise (such as it is)  is by no means steady.

To show how insignificant the warming is in Antarctica, here are annual mean anomalies compared with those of Australia:Antarctic-Oz comp

Remember, one of the so-called “fingerprints of greenhouse warming” is that warming should be greater towards the poles.

Professor Turney could have saved himself a lot of time, trouble, and embarrassment.

BOM’s Annual Climate Statement

January 3, 2014

This is just a quick comment on the 2013 Annual Climate Statement released today by the Bureau.

I have checked only minimum temperatures, and my calculation I released yesterday for the 2013 annual mean for minima was +0.82 C.  The Bureau has reported  an annual average of +0.94 C.

The BOM’s figures are derived from the Acorn dataset.  They acknowledge in the small print that:

Note that all values in this statement are as compiled from data available on 2 January 2014. Subsequent quality control and the availability of additional data may result in minor changes to final values.

What they really mean is that Acorn won’t be updated with 2013 data for several weeks, and that nearly all Acorn sites have many months of daily data that has not yet been quality controlled.  This means that the values they give can not be checked for several weeks.

As well, three Acorn sites in Western Australia ceased reporting in August 2012.  Bridgetown has years of data that is out by two days, and Rutherglen has years of data out by one day.   And on some 800 occasions the minimum temperature is higher than the maximum.  Quality checking is not as rigorous as you might expect.

Any changes may be only minor, but rushing to publish before the data can be checked is not a good look.

Weather predictions- final check

January 3, 2014

This is the final post in my series of checking a hunch that temperature change indicates a weather change 160 days later.

Back on November 3, I predicted:

2 to 10 unstable; 12-13-14; 16-17-18-19; 22-23-24; 26 to 31 unstable.”

I did not change this in December.

This is how  I went:

Dec2013 predictions check

I’ve marked with green bars the predicted dates of unsettled weather as above.  Red bars show the actual times.  They match.

And finally, here’s my graph showing predicted weather events for April to June.  Again, green bars indicate dates when weather events may be expected.Apr-June 2014

I will leave this topic for now, not because the method doesn’t work (it does!), but to concentrate on other interests.

2013 Minimum Temperatures Released

January 2, 2014

Ken Stewart, 2 January 2014

UPDATE 3 January: BOM has updated it’s time series graph, but not the raw data, which still finishes at 2012! See below.

I have calculated the annual 2013 minimum temperature anomaly for Australia, well before the Bureau of Meteorology.

Not including the 8 sites acknowledged as having anomalous warming due to the Urban Heat Island (UHI) effect, I calculate the straight mean (without area averaging) to be +0.82 C.  This puts 2013 as second warmest after 1998, and just ahead of 1973 and 1988.

I expect that the BOM will publish a figure of around +1.2C, and claim 2013 as the warmest on record for minima.

I calculated this by using daily Acorn data for 1910 to 2012 from http://www.bom.gov.au/climate/change/acorn-sat/  , plus daily minima for 2013 for these same sites from Climate Data Online.  I used Acorn data from 1961-1990 to recalculate monthly means for each site, and then calculated running centred 31 day means to estimate daily means for the same period.

Then I calculated daily anomalies for each site, and amalgamated these into a straight mean for Australia.

The result is as follows:

Fig. 1:  365 day running mean of daily data.acorn 365d 1910-13 no uhi

I will analyse Fig. 1 in some detail later.  But first, how does my calculation stack up against the BOM super computer?

Fig. 2: Annual (31 December each year) means of minima 1910 to 2012.Acorn ann v me 1910-12

My calculation is in green, BOM in red.  As you can see, the match is pretty close, and of course I have not used any area averaging.  But you would expect the results to be close, as I have used exactly the same data.  You will notice that the major differences occur in years of higher or lower than normal minima.  These appear to have become larger in the last 40 years.  The official annual figures show greater extremes, as shown above.

I have also calculated trends for the 1910 to 2013 period, and hope that this will persuade you of the futility of using linear trends for temperatures, and that if you cherry pick you can prove just about anything.  The next graph is a plot of the continuous running trend from 31 December 2013 all the way back to 1 January 1910.  That is, the linear trend through datapoints between any selected date and 31 December 2013.

Fig. 3: Continuous running trend, daily minima anomaliescont trend Oz no uhi

The vertical axis measures trend in degrees Celsius at particular points in time.  Note the rapid fluctuations at the right hand end.  I’m sure no one would be silly enough to calculate trends of only a few years’ data.

As the time period increases (moving from right to left) the fluctuations smooth out.  Note that Australia has had zero trend in daily minima since 21 July 1997.  Interesting, but no predictor of the future.

Moving further back in time, the plot shows the temperature trend increasing until the early 1940s.  Up until then the long term trend is fairly stable.  Since 1910 the trend is about 1.1C per 104 years.  The maximum trend can be calculated from 1922. Therefore, a cheerful cherrypicker can choose whatever time frame they like to produce a linear trend that suits.

Back to my graph of the 365 day running means of daily temperatures. Figure 1 again:acorn 365d 1910-13 no uhi

Note that the 365 day mean peaked in early November 2013 and has dropped since then.  The peak was at +0.94C, which is still below that of 1998 and 2006.

But also note that the rise of about +1.1C over 104 years is by no means steady.  There are several sharp rises and falls along the way.  Let’s have a closer look at these.

Fig. 4: Step changes in temperatureacorn 365d 1910-13 no uhi stepups

I have shown (starting in 2014) how the minimum temperature record of Australia features a series of sharp step ups, followed by slow declines.  I have indicated the start of these periods and the linear trend lines of each one.  There may have been one in 1926, and 2013 may (or may not) be the start of another such period.  They are more frequent and more pronounced in the past 40 years than in the first 60 years.  This appears to show a link to natural climate forces, such as the El Nino- Southern Oscillation.

I will analyse these results further in future posts, and may do the same for maxima as well.  (People are interested in maxima because “that’s how hot it is”.  I like minima because they tell you more about climate e.g. if they increase faster than maxima this may indicate greenhouse warming.)

Watch for the official 2013 minimum temperature anomaly:  probably +1.2C.

Update 3 January:

Here is the official BOM graph to 2013:

timesereis tmin to 2013

and it looks like a bit over +0.9C  +0.94 C, so less than I expected and closer to mine.

No Warming in North Australia for 31 Years

December 23, 2013

I’m nearly a year late with this, but I’ve only just noticed.

According to the Bureau of Meteorology’s official temperature records, for all of the Northern Australian region- the half of the continent north of 26 degrees South- the minimum temperatures are steadfastly refusing to rise.  From 1982 to 2012, the linear trendline for minima is on the decreasing side of dead flat.

Acorn tmin Nth Oz 82-12

This is longer than the 3o years regarded as the minimum period for analysing climate trends, and in spite of the massive increase in amount of CO2 emissions.  Note that 1982 and 2011-2012 were almost equally cooler than normal.

Remember  that one of the fingerprints of greenhouse warming is that minima should be increasing more than maxima.

Here is the 365 day running mean of daily minima anomalies of all Acorn sites in Northern Australia (more about this next year) up to early December this year:

tmin nth aust 1910-13a

Rather than a smoothly rising trend, the record is characterised by 10 to 15 year rapid rises and falls, responding to events in the Pacific and Indian Oceans.

This is a diagram of Australia’s climate regions:summer1213  regions

After New Year I will post about minima for other regions and Australia as a whole.

Merry Christmas to all.

It’s Official: a 27 Year Hiatus

December 13, 2013

Australia’s Bureau of Meteorology (BOM) has been publishing its brand new (less than 2 years old) temperature dataset, the Australian Climate Observations Reference Network- Surface Air Temperatures (ACORN-SAT, or Acorn), and using this to make scary pronouncements about our climate.

Back in March, I showed that there had been an 18 year pause in Australia’s mean temperatures.  In April I showed how Dr Braganza’s “Fingerprints of Global Warming” (from an article in The Conversation on 14/06/2011) are absent in Australia.  I included a passing reference to minima, which really needs a separate post, so here it is.

Why minima?  To quote Dr Braganza,

…Patterns of temperature change that are uniquely associated with the enhanced greenhouse effect, and which have been observed in the real world include:


  • greater warming of night time temperatures than daytime temperatures

Greenhouse theory suggests that greenhouse gases slow the normal rapid night time radiation into space, therefore minima should be increasing.  Australia as a region is roughly the same size as the USA, or Greenland, or Europe, or the Arctic or Antarctic, and is supposed to be at great risk from the effects of global warming, so the behaviour of minima in Australia should be of some interest.

So let’s see if this unique pattern of temperature change has been observed in the real world.

This is a graph of annual minimum temperature anomalies across Australia from 1986 (updated!) to 2012, sourced directly from Acorn data at the BOM website at http://www.bom.gov.au/web01/ncc/www/cli_chg/timeseries/tmin/0112/aus/latest.txt:Acorn tmin Australia

Which clearly shows that minimum temperatures have been cooling, not warming, for 27 years.

Strange, but true.  You won’t see that mentioned in any BOM publication.

Weather predictions: December

December 1, 2013

At the start of November, I said:

5 to 10 unstable; 13 to 21 unstable with several events; 26-27-28-29-30 unstable.”

All correct, 1 miss.  Instability with some very wild storms marked much of November especially in the South-East of the state.

Now I suppose anyone could have predicted storms for November.  But remember, back in August I had said:


5-6-7, 9-10, 13-14-15, 17-18-19-20, 27-28-29.”

Here’s a chart showing August predictions in light green and early November predictions in dark green.octdec13resultsnov

5 right, I miss.  I should have stuck with my original predictions!

So the method is holding.

Predictions for December to 31 March remain the same as I predicted last month.  As well, I expect weather events around these dates in April and May (+/- 1 day):

2,4,7,11,15,20,23,25, May 1, 8,11.

April should have unstable weather, and I would not be surprised if we get significant rain.


Weather Predictions November – March

November 3, 2013

At the start of October, I made the following predictions:

Sep 30- 1-2, 8-9-10, 14-15-16, 20-21-22-23, 25?, 28-29-30-31-Nov 1 (possibly 2 events).

Changes were detected on:- 2-3, 8-9, 13-14, 17-18, 22-23- 24, 28- 29-30-31 (the”possibly 2 events” was true indeed with a series of storm fronts that persisted.)Oct 13 results & pred

Dark red is Tmin, pink is Tmax, black is inverted pressure, blue is rain, green is the 160 day lagged Tmin, yellow is 40 day lagged pressure.  I may try 40 day lagged temperature in future.   November and December predictions are shown as red ovals.  So there were 5 correct predictions, 1 wrong, and 1 miss (although the wrong event would have been a “correct” if it had been a day earlier).  Not bad for 160 days out, I console myself.  I also suspect there has been a slight acceleration of weather systems, so perhaps I should be looking at events around 158 days ahead instead of 160.  No matter for now but I will keep an eye on it.

Predictions for November and December (slightly adjusted from last month):

5 to 10 unstable; 13 to 21 unstable with several events; 26-27-28-29-30 unstable.
2 to 10 unstable; 12-13-14; 16-17-18-19; 22-23-24; 26 to 31 unstable.

 January to March:

Here is the graph I use for predicting weather changes, which shows 2nd derivative, or acceleration/deceleration, of minimum temperature.

jan-mar 2014 predictions

Think of a cool change moving through from the west.  A large temperature differential, or a fast moving change, both have the same effect.  When the green line goes below zero, the temperature has decelerated, and a change (or unsettled weather) occurs in the time period indicated.  Several rapid oscillations appear to relate to unsettled weather.  My only problem is I am averaging over a very wide region, and the time lag may be changing.

There is always something to learn.

UAH for September

October 19, 2013

The UAH lower tropospheric temperature anomaly for Australia for September was +1.56C, up from +1.12C in August.  The 12 month running mean has now reached +0.72C, which as commenter barry noted is not statistically different from equal highest since 1979.  Nor is it a record- the 12 months to September is now the third hottest (after June and May 2010).  Uncertainty for UAH is +/- 0.1C.  We may yet see a record.

uah 12m sept13


2010 and 2013 marked.  Mean and +/- 2 standard deviations shown.


Get every new post delivered to your Inbox.

Join 35 other followers