The Need for a New Paradigm, Including for Rainfall Forecasting

The following paper was delivered by Jennifer Marohasy at the Ninth International Conference on Climate Change in Las Vegas on Wednesday 9th July 2013 in Panel 13.

Jennifer Marohasy at 9ICCC

Jennifer Marohasy at 9ICCC

SCIENTIFIC disciplines are always underpinned by theories that collectively define the dominant paradigm. In the case of modern climate science that paradigm is anthropogenic global warming (AGW). It defines the research questions asked, and dictates the methodology employed by the majority of climate scientists most of the time. AGW may be a paradigm with little practical utility and tremendous political value, but it’s a paradigm none-the-less. The world’s most powerful and influential leaders also endorse AGW. Even Australia’s Prime Minister Tony Abbott said in June, “I don’t think people should run around pretending there is disagreement when none exists. President Obama and myself both take climate change very seriously.”

AGW has even dictated the question for this session, “How is climate change affecting sea level, rainfall and water availability? Will a warming planet complicate or alleviate water challenges?”

But what is the utility of such a question, particularly given there is growing evidence that planet earth has already entered what could become a protracted period of cooling? This is not what we hear in the mainstream media, but this view is supported by evidence in the unadjusted temperature data for my state of Queensland in Australia, and is also what some astrophysicists have been forecasting for some time.

If mainstream science operated under a different paradigm, one where researchers believed it was possible to forecast weather and climate not just 3 days in advance, but with a high level of skill 3 months, or 3 years in advance, and if their primary focus was not justice, equity and curbing greed, but rather the provision of useful information to Joe Citizen, the question for this session might have been very different. “How much rain is forecast to fall on California’s Central Valley each and every month for the next 2 years?” Answering such a question could not only aid food production, but also facilitate planning for floods and drought. The work I have been undertaking with Professor John Abbot, Central Queensland University, is attempting to address this type of question for Queensland, Australia.

The General Circulation Models (GCMs) that underpin the paradigm of AGW have difficulty generating rainfall forecasts with any real level of skill more than 4 days in advance. If a fraction of the billions spent developing these simulation models, had instead been invested in a theory of climate underpinned with state-of-the-art statistical models based on an understanding of natural climate cycles, I believe we would be much closer to being able to mitigate climate variability across the globe through better rainfall forecasts with very significant benefits for all of mankind, but particularly subsistence farmers in places like India. In short, what we really need is a new paradigm for climate science, underpinned by new tools with some utility.

So, what I will do in the limited time available today is tell you something about how artificial neural networks (ANNs), a form of Artificial Intelligence (AI), have potential as a new tool for a new paradigm, and in particular their application to rainfall forecasting. I will then be in a position to better answer the question for this session concerning rainfall and water availability in a warming world.

ANNs are massive, parallel-distributed, information-processing systems with characteristics resembling the biological neural networks of the human brain. Imagine a computer with the capacity to search and find the complex linear, and also nonlinear relationships, which may exist between local temperature, rainfall, and Pacific Ocean phenomena including changing patterns of sea surface temperatures and pressure. Image a computer with a powerful and versatile data-modelling tool that is able to capture and represent input and output relationships acquiring knowledge, through learning from multiple examplars deciphered from vast arrays of historical data, storing this knowledge within inter-neuron connections strengths known as synaptic weights. A computer that can learn relationships, model and measure relationships, then use this information to forecast rainfall.

This is essentially what the ANN that John Abbot and I built to forecast rainfall for 17 locations across Queensland can do. In a paper by us published two year ago by the Chinese Academy of Sciences’ journal Advances in Atmospheric Science (Volume 29) we detail the model and demonstrate how much more skilful medium-term monthly forecasts from this model are, relative to forecasts from the Australian Bureau of Meteorology’s best GCM. In a more recent publication in the journal Atmospheric Research (Volume 138) we show how forecasts for Queensland are improved with inclusion of the Inter-decadal Pacific Oscillation. In a conference paper presented a year ago in Southampton (River Basin Management VII, WIT Press) we show the potential of ANN to forecast extreme rainfall, specifically the devastating flooding that submerged Queensland’s capital Brisbane, in January 2011.

We have another paper currently in press at the International Journal of Sustainable Development and Planning, which considers the affect of a 3 degree C increase in temperature on rainfall, a component of the question for this panel. What did we find? In our cast study when maximum and minimum temperatures were increased by 3 degree Celsius there was a decline in summer rainfall and an increase in winter rainfall at a placed called Nebo in central Queensland.

The bottom-line is that ANNs can already provide better medium-term rainfall forecasts for Queensland, Australia. This has a real practical value. ANNs can also provide an independent method of GCM validation under future climates with results from Nebo suggesting a smoothing of the annual variability in rainfall rather than more climatic extremes assuming global warming.

ANNs use the existence of recurrent patterns in historical data to inform the rainfall forecast. That our rainfall forecasts drawing in most cases on about 85 years of temperature data show considerable skill, means natural climate cycles must persist. Climate is not on a new trajectory as suggested by many proponents of AGW.

A great advantage of using ANNs is that they can easily be adapted to test and incorporate additional input data series, as climatic knowledge develops. If the relationship can be quantified it can be modelled by an ANNs, including potentially lunar, solar and planetary cycles.

While AGW is a demonstrably failed paradigm, it will be replaced only when a critical number of practicing scientists start working on something new. New paradigms always have their own questions and their own tools. In the same way that GCMs underpin AGW theory, ANNs could underpin a new paradigm based on better quantifying the drives of natural climate cycles.

Henry David Thoreau wrote, “In the long run, men only hit what they aim at.” What are we as a community of sceptics aiming for? Just the overthrow of AGW, or can we aim much higher, including for skilful rainfall forecasts?

Thank you.


I would like to thank The Heartland Institute for the opportunity to present at the Ninth International Conference on Climate Change. All the research on the application of artificial neural networks to medium-term rainfall forecasting detailed here has been done in collaboration with Professor John Abbot funded by the B. Macfie Family Foundation.

A video of the talk can be found at this link…
Scroll to find image of me in red, click on the link. The tiny URL for this blog post is


130 Responses to The Need for a New Paradigm, Including for Rainfall Forecasting

  1. kevin long July 11, 2014 at 8:09 am #

    Well done Jen, now you are on target and getting much close to what really maters, seasonal forecasts for rainfall and future rainfall trends, that is what I have been working on for more than a decade, the results of my extensive research can be found on my web site I hope you enjoy reading about my journey to help the farmers of the MDB and all that depend on them.

  2. Larry Fields July 11, 2014 at 8:28 am #

    Hi Jennifer,
    There’s a parallel between your work on seasonal rainfall forecasting in Queensland, and the work of Wassily Leontief, who was awarded the Nobel Prize in Economic Science in 1973. BTW, his Harvard colleagues were really PO’d at him.

    Leontief pioneered Input-Output Analysis in Economics. Like the Artificial Neural Networks that you use, IOA is computationally intensive. And it mercilessly bypasses the dead wood in prevailing theory.

    After WW II, most of the big-name American economists of the day were predicting a return to the conditions of the Great Depression. As we all know, they were dead wrong. Leontief was the one of the few who got it right.

    Like the AGW faithful, specialists in Macro-economics have been a hidebound lot. One of their articles of faith was that it’s necessary to keep unemployment high, in order to control inflation.

    That belief system was shaken, but not stirred, during the Stagflation of Jimmy Carter’s presidency. During the Reagan presidency, Fed Chairman Paul Volker stated that it would be unwise to ever let unemployment drop below 7%!

    I have not kept close tabs on the Macro crowd. My general impression is that they still pay lip service to their tin gods, but they are less dogmatic about it. Ironically, we still have the Phillips Curve, but without Phillips, and without the curve!

    Within the field of Economics, IOA has been relegated to the theory of International Trade. It’s unfortunate that Leontief’s work has been marginalized in this way.

    Leontief was one of the rare individuals who could make important contributions by thinking outside of the paradigm. But as we all know, zombie paradigms (like AGW) have remarkable resiliency.

    My main sources of information are conversations with the late systems scientist, Glenn Burress, who successfully predicted all of the economic ups and downs of the 1970s. Unfortunately, Glenn was too in-your-face, and was marginalized to a much greater extent than Leontief. Here’s a link to a short biography of Wassily Leontief.

  3. Ian Thomson July 11, 2014 at 9:38 am #

    Well put together.

  4. Luke July 11, 2014 at 10:07 am #

    It’s all just spruiking and hand waving unless there is a web page with a forecast there each month.

    Simple as that.

  5. Emily Frazer July 11, 2014 at 12:52 pm #

    Dr Marohasy has articles in the best journals and a big vision. If the BOM worked with her, we would already have better monthly forecasts on the tax-payer funded website.

  6. sp July 11, 2014 at 4:52 pm #

    Nothing more than hand waving from Luke as usual. Simple as that.

  7. Luke July 11, 2014 at 8:17 pm #

    Emily – a Google Scholar search of …. ann rainfall forecast neural network …. gives 18,500 results. So BoM may be somewhat aware of the broader literature.

    There is of course as an industry backed route for direct promotion.

    So sp – you’re not curious to see the system running for real – surely even a hard man (maybe a woman) like yourself is interested ?

    However if we can’t see the system running in real-time I reckon it doesn’t work (being a sceptic you see and believing nothing works well enough).

  8. Mr Koala July 11, 2014 at 8:20 pm #


    You seem to have no understanding at all to test a predictive model.

    It has nothing to do with waving your arms around or having a website.

    What it requires is testing the model on a data set that has not been used to train the model.

    Then you can evaluate the predictive skill, and make comparisons with other models.

    Abbot and Marohasy have already published such comparisons in recognized journals. In particular, the BOM’s general circulation model POAMA does about as well as the Brazil football team against Germany. (Germany 7, Brazil 1)

    POAMA struggles to even beat climatology – that’s just the long term average – or in football terms – Iceland.

    Poor Like, your team just isn’t very good. Suggest you try another sport. maybe one that involves arm waving.

  9. Luke July 11, 2014 at 9:28 pm #

    Bearsy mate – The said publications weren’t convincing at all as previously discussed in detail. And surely you’re not going to appeal to journal publication when sceptics have been ranting for years how broken peer review is. Appeal to authority now? Weally?

    Jo Nova now has the gold standard – here’s our model lock stock and barrel. But of course rainfall forecasting may have financial value so one might not want to publish exact details. One can appreciate that.

    However no punter is going to go with something with no for real track record. You can hindcast all you like. BoM tell us that all the time and all Debs looks at is the latest forecast.

    And hey perhaps Nova and Evans are right the world is cooling so maybe baseline indices used to build models aren’t stable. Who would know?

    Pundits want to see the real thing.

    If it doesn’t exist – why? Probably coz it doesn’t work – it’s the only serious test.

    And the longer it doesn’t exist the more wonders why not? It’s not a trick question but having no real-time site is pretty tricky !

    Whether it be ANNs, lunar, solar, planetary do-hickies – I say it ‘s all bunkum – so show me ! We don’t want gonnas – gonna do this and gonna do that !

    (team – I don’t have a team – I’m just some contrarian bastard)

  10. Debbie July 11, 2014 at 11:00 pm #

    Part of your argument is precisely the point!
    A lot of money & time has been wasted on something that has NOT delivered anything particularly useful.
    At least Jen & Abbot & others are prepared to look outside the box (or the prevailing paradigm).
    It may not deliver either . . .but your sooking over how it is or isn’t presented is missing the point.
    Quite clearly. . .The science wasn’t settled!
    As I have frequently commented. . .there will be no one happier than us farmers & all the other businesses that work with the weather/climate all the time when seasonal variations & climate cycles are better understood and become easier to predict.

  11. Luke July 11, 2014 at 11:31 pm #

    Yes yes Debbie no need to go on about all that – all understood, and govt money pissed against the wall as you say – tsk tsk tsk – but you’re in the new era now with the white knights – its new blue Omo Debs – so let’s move the discussion on.

    POAMA is crud – so let’s see how those 18,500 papers on neural nets stack up.

    So where’s the real time system for Qld? Aren’t you curious how these last seasons would have gone with good old ANN – you’ve have some big wets and sneaky droughts to give it a test on.

    Debs – surely of all people you are vitally interested? Surely? Don’t tell me you’re just a tyre kicker. Me – I just reckon it doesn’t work and it’s pure bunkum. It’s called being sceptical. Learnt that here I did.

    What would Popper say if you wouldn’t submit your system to the real world. Yo’all have been lecturing me about that for years. Empirical evidence as you dearly love Debs. No Socratic irony scrap metal here Debs.

  12. cohenite July 12, 2014 at 10:33 am #

    I thought you had retired luke?

    You’ve made these complaints before luke:

  13. Debbie July 12, 2014 at 1:41 pm #

    I am vitally interested Luke.
    You really do write some nonsense sometimes.
    It’s time to move on. All of us out here in the real world need to get off it and on with it.

  14. Luke July 12, 2014 at 4:11 pm #

    So if you’re vitally interested you’d want to see it in action wouldn’t you Debs. Go on say say “yes”.

    And it is time to move on? So what are you moving on with Debs – your wet thumb checking the wind direction. Surely you’d like a nice neural net forecast? Wouldn’t you just love one on streamflow !

    Cohenite – yes I have retired but occasional voluntary guest appearances are possible. Basically the AGW wars are over – the world will move on this in the next few years. It’s all over I’m afraid.

  15. Toby July 12, 2014 at 5:28 pm #

    “Basically the AGW wars are over ” yup agreed, carbon tax will be thrown out, temp hasnt climbed and climate shrills are increasingly looking like scaremongers and delusionalists. Funny how you think they are won when most forecasts associated with CAGW have not occurred. What a deluded world we live in

    The world is not acting in any significant way, the key emitters such as china and india get very little energy from renewables (less than 1% if you exclude biomass {cow dung etc!!}. and what the west has done and is doing will achieve ntg except lower living standards and reduce our ability to cope. Lomborg discusses this in the Oz today

    By all means stay deluded Luke, because that s what you are if you think anything significant is being done to curtail CAGW.

    A few more years of no warming and even zealots will have to acknowledge how awful the theory of CAGW actually is.

    Real scientists would acknowledge it already, but climate scientists continue to do immeasurable damage to science with their pseudo garbage activism dressed up as real science.

    Only the ignorant or activists with ulterior motives could truly believe in CAGW. The facts do not lie, rapid rises in co2 have not caused rapid changes in temperature, neither correlation or causation has been proved and only the blinded could miss this!

  16. Toby July 12, 2014 at 6:12 pm #

    a few years ago a number of “scientists” (a generous term for a pseudo science) used adelie penguins as proof of what would happen to humanity as the world warmed.

    But as usual catastrophists scares turn out to be worth less than the paper their ideas are written on.

    it wont be long until climate scientists and advocates, spruiking the catastrophic effects of co2 will be treated with the disdain they deserve. And yes that is what they deserve for the damage they have done to science and the scientific method.

  17. Luke July 12, 2014 at 6:17 pm #

    Unless Toby you only like looking at denier friendly satellite data sets. Unless you want to deny the ocean temperatures at depth. Unless you want to deny the palaeo evidence that 2xCO2 = 3C.

    And unless like Debbie you think we’re there yet. “Mum are there yet – I want to be there now – why aren’t we there now”.

    WG1 authors aren’t real scientists – oh come now !

    Lower living standards – pffft – well one million solar households and you’ve never had it so good. Stop sooking. Age of entitlement and all that …

    Blame the coal sucking right wing for shutting down manufacturing in this country.

    Anyway – lah de dah – let’s see some for real ANN forecasts !

  18. Robert July 12, 2014 at 6:37 pm #

    Jen, I hope you keep plugging away with your ANN. I won’t have an opinion on any of it till many seasons have passed, but I can’t see the harm in having a go.

    Even the BoM was entitled to have a go. Then they got religion.

  19. Debbie July 12, 2014 at 7:09 pm #

    Interestingly, I am in India at the moment.
    It is a fascinating place, full of paradoxes and sometimes quite confronting. We should remember how lucky we are in Australia. . . and how environmentally responsible we are.
    It’s time to move on from the gloom and doom predictions.
    SE Aust has not dried up, we did not have those millions of climate refugees, it still snows in the
    Alps, coastal areas are still above water etc etc
    Time to let go that old paradigm.
    Time to get off it and on with it 🙂

  20. Luke July 12, 2014 at 7:36 pm #

    Yes yes yes Debs – well put.

    Well why are you here then Debs if you’re getting on with it and off it – shouldn’t you be planting something or doing a service on the John Deere?

  21. Debbie July 12, 2014 at 8:06 pm #

    Learning about & experiencing other cultures.
    I think it’s still classed as doing something.
    We finished harvesting our summer crops & sowing our winter crops before we left.
    We also serviced the NH header, the MF tractor, the truck and all the other machinery, but we don’t have a JD.
    It’s called time management Luke.
    Us business owners have to be good at it, especially us farmers. We have fairly tight seasonal windows in our profession.
    On an irrigation farm July is fairly quiet and the best time to gain other off farm experiences.

  22. Toby July 12, 2014 at 8:07 pm #

    “Unless Toby you only like looking at denier friendly satellite data sets ” all the global data sets Luke! not ones denialist’s (realists more like!) advocate at all..They were the same data sets you mob of catastrophists used as “evidence” of warming. But like all good pseudo scientists when teh facts dont fit…well create some new “facts”, no matter how iffy (infact the more iffy the better)…thats why now you are asking us to look in the deep oceans.

    You do realise we have no idea what has happened to the deep oceans in the past and have little idea about what is happening now?!!….but of course that allows pseudo scientists to claim with even greater certainty they are correct in the catastrophic forecasts!

    as for the paleo record. It clearly shows co2 lags temperature rise except for one or maybe two periods in history.

    The Earth is evidence that negative feedback dominates, sure that doesnt mean positive feedback effects never occur, but clearly negative feedback dominates

    most of us agree with the physics that double co2 and you get about 1c increase (1.2) in temp.

    But like all good scientists with a cause, the dominant evidence can be thrown out for a good cause and so now doubling co2 leads to 1.5-4.5 c increase in temp……….ie they are ignoring even the possibility of negative feedback effects.

    climate science only deserves our disdain and contempt because that is what they have brought on themselves with their continual stream of exaggerations and non events. A real science is falsifiable, but every wrong prediction by climate science is treated as yet more proof of its veracity. Pseudo science is a generous term.


  23. Luke July 12, 2014 at 10:34 pm #

    Toby what utter drivel – get yourself updated on palaeo – ice core lags are utterly irrelevant. For heavens sake. How many times have we been over this.

    Ignoring possibility of negative feedbacks – errr no.

    You’re just parroting denier memes. So tedious. Including the word “catastrophic”. Mate if it was already “catastrophic” somehow we wouldn’t be even debating. “Are we there yet”?”

    “A real science is falsifiable” yup and that’s why we need to see real time neural nets – back on topic.

  24. Toby July 13, 2014 at 12:20 am #

    fair go Luke, without “catastrophe” the warming case has ntg? ie if realists are correct and co2 doubling only causes 1.2 c or under temperature increase then the whole thing is shown for what it is………..the whole premise of “climate change/ global warming” rests on additional warming having significant impacts…realists recognise this, the real denialists are those who hang their hat on the dangerous consequences.
    How can the climate sensitivity be 1.5-4.5c if the phsyics states a doubling of co2 causes about 1.2 c increase in temp? Why are they completing ignoring the possibility of negative feedback effects? You do agree that it is negative feedback effects that actually dominate the Earth’s system in general I assume? (otherwise we wouldnt be here would we?)

    As a general comment I assume you would agree that warmer is better than colder? and I seem to recall even the IPCC agrees the benefits of warming are largely positive until 2070 .

    You are right, ntg I mentioned has not been discussed by us over the years.

    I disagree with you and i appear to have the evidence on my side. You state my points are drivel (they are true however{eg 1.double co2 = +1.2c, 2. co2 lags temp rise, 3. negative feedback effects in general dominate, 4.all the global data sets that are generally referred to show no warming for up to 18 years etc) you just chose to select other “facts” as more important ( like the heat is hiding in the oceans…seriously come on Luke, if you had a foot in the sceptics camp you would laugh at this surely!?)….but that is all i seem to hear coming from warmists mouths unfortunately.

    From where i sit the science is questionable and the economics of what we are doing is appalling. I have been saying this for a decade now and given no warming for so long who has been right since we started debating this crazy topic?) I suspect in 20 years time you will be very embarrassed you have fought so hard for this pseudo science.

    On the topic of this thread…the need for a new paradigm…couldnt agree more. Perhaps its time you considered the same?

  25. Debbie July 13, 2014 at 5:25 am #

    Are we where yet Luke??????
    You are arguing with those straw man thingos again.
    Who claimed ‘the science is settled’?
    It certainly wasn’t Marohasy & Abbot.
    Who /what behaves as if they own & manage the weather/climate?
    Once again. . .It wasn’t those you call ‘deniers’
    Perhaps you could ho back to deltoid and play devil’s advocate over there?
    They need help.
    Along with SKS and hotwhopper and Rabbet etc they are withering on the proverbial vine with their inability to let go of a failed paradigm.
    Or maybe you could come visit India and get a dose of reality?

  26. Luke July 13, 2014 at 8:32 am #

    “all the global data sets that are generally referred to show no warming for up to 18 years” errr no BTW – get minimally updated

  27. spangled drongo July 13, 2014 at 10:49 am #

    Luke shows his climate integrity once again by linking to computer generated pretties
    that he wants us all to believe instead of being somewhat sceptical of the worlds adjusted data for the last century or so.

    If there’s one thing current arguments over data manipulation have shown us it is 1) there is very little happening other than natural variability, 2) past nat var has been more extreme and 3) to send the poor of this world into more poverty to try and rectify what is very likely not only a non-problem but possibly an advantage is just the carpet baggers like Clive Palmer and Al Gore in full cry.

  28. Toby July 13, 2014 at 10:53 am #

    interesting link Luke….thx for confirming my point for me about a lack of recent warming. You dont agree all of those data sets show no real warming since at least 2000? Ignore the rainbow diagram and look at the actual lines in the graphs. One day you will wake up and feel so foolish.

  29. Luke July 13, 2014 at 1:03 pm #

    Toby – the land based data sets all show warming in the 17-18 years.

    More detailed analyses show seasonal differences. If you think nothing is happening you’re deluding yourself.

    As for palaeo demonstrations of powerful greenhouse forcing there are excellent examples – the Paleocene–Eocene Thermal Maximum, Mid-Eocene Climatic Optimum and Eocene Thermal Maximum 2

    Transiting from Snowball Earth by CO2 forcing (and dust) is another.

    CO2 following temperature in glaciation cycle cores is simply because changes in insolation not CO2 start the melting process with CO2 kicking in later. I am gob-stopped that any serious sceptic does not understand this.

    Spangled thinks its computer generated – of course it is silly – so you can do simple stats tests for any period with any data set. HOW HANDY !

    What is also computer generated is any satellite data – the temperature number reported is a long way from the brightness value measured by the on-board instrument adjusted for radiometric and orbital effects across a range of satellite over the years. It’s as COMPUTER GENERATED as you can get. BTW your tax return and bank account are also computer generated.

  30. Debbie July 13, 2014 at 1:24 pm #

    The topic is mostly about the need for a new paradigm as the current focus is not proving to be useful.
    Absolutely no one has said that nothing is happening.
    The delusion is yours on that score.
    Computers are a useful tool to help us make sense of the world around us.

  31. Robert July 13, 2014 at 1:35 pm #

    Toby states that CO2 increase follows warming. Luke is gob-stopped and points out that…CO2 increase follows warming.

    I’m different again. I say that CO2 increase follows warming. Don’t be gob-stopped now!

  32. Luke July 13, 2014 at 2:14 pm #

    Robert – too immensely stupid to reply to. You really do need to try harder.

  33. Robert July 13, 2014 at 2:40 pm #

    Supe, it’s like when you had to tell SD in no uncertain terms how back-up generators are used as back-up generators. I have to admit, I was one of the silly sausages who thought that back-up generators are actually used as back-up generators. And I firmly believed that CO2 followed warming, whereas all along CO2 followed warming. I’m such a dope.

    Chick at the local takeaway used to tell customers that fish pieces cost 80c each but they could have five for $4. Hombre, the hungry punters just grabbed that $4 bargain every time. Don’t know why you made me think of that.

  34. Debbie July 13, 2014 at 2:40 pm #

    BTW Luke,
    If I had done or not done most of what govt employees have told me I should or shouldn’t do over the last decade. . .I would no longer be farming.
    You are not my employer or my tour guide and your shoulding is getting a tad irksome.

  35. bazza July 13, 2014 at 4:35 pm #

    bazza July 13, 2014 at 2:24 pm #

    bazza July 13, 2014 at 12:05 pm #

    This must have been deleted twice.???? I will have to dribble the bits in to see what gets through this place of free and vigorous discourse.
    The criteria for research include that the results be reproducible – no argument there. It is not reproducible If you don’t have a rigorous process to pick the rainfall stations you use, who knows how much pick and flick was done. .
    Why was the A&M approach the first not to use anomalies and ignore that the main signal in rainfall data is a seasonal one and if you don’t get rid of that first you end up with very questionable results

  36. Debbie July 13, 2014 at 5:00 pm #

    Along with Luke, you are missing the point.
    A & M are actually working on reaching a better understanding of seasonal variations & doing some thinking outside the box that you seem to be locked inside.
    All results are questionable Bazza.

  37. spangled drongo July 13, 2014 at 5:19 pm #

    Yes Debbie, Luke’s box comes in handy.

    It’s his PETM box that he seeks when deafened by background noise that buries his blurred signal.

    It also protects him from those 8000 foot thick glaciers that existed while CO2 was in excess of 4000 ppm.

    In soft science you need a handy box like that.

  38. Luke July 13, 2014 at 6:47 pm #

    Bazza I wouldn’t bother attempting to comment. There is no discussion here and obviously proliferation of nonsense and poor science is preferred. A far cry from the vigour of earlier days.

    So sorry Debs and Robert – unable to engage. (Snip)

    Back to retirement and enjoy the echo chamber.

  39. spangled drongo July 13, 2014 at 7:52 pm #

    Try some great CO2 science before you close the lid Luke. Bound to cheer you up:

  40. Debbie July 13, 2014 at 9:05 pm #

    I think that’s called ‘flouncing’ Luke.
    But seriously. . .Science blogs like Deltoid and SkS are keeling over. . .they would probably appreciate some help and support from you.
    If we’re talking echo chambers. . .they are excellent examples.

  41. spangled drongo July 13, 2014 at 9:22 pm #

    And some vigour of earlier days:

  42. cohenite July 13, 2014 at 9:42 pm #

    See ya luke; temperature is flat no matter what tammy says; in the case of RSS for longer than Santer’s statistically significant period of 17 years for the period to be classified as climate not weather and the EEB is negative, more OLR, since 1998. No OHC and SST declining etc.

    By any reasonable standard AGW is as empty as Hansen-Young’s head.

  43. Luke July 13, 2014 at 10:26 pm #

    Nah I was just kidding. I’m back. Over that now – but bit hard to refute when your refutation is snipped. And has cost me a new screen after punching the last one out.

    Seems nobody wants to talk ANNs so I guess trading insults is what is comes too. Bazza making quality comments but no engagement.

  44. Robert July 13, 2014 at 11:22 pm #

    There are cycles, of a bewildering and messy sort. If someone can make a bit more sense of them for the cockies and planners, it’s worth a shot. If it’s light on cost to taxpayers, even better. If it doesn’t work…what’s new? The first thing is to exclude the opinions of those who believe in trajectories and unstoppable trends. Such people are great at sounding sciency, but they are just negative emotion with legs. Which is why they’ll say anything at all.

    The old Brisbane Courier correspondent of 1889 who researched the droughts and floods from Mathew Flinders onward had this to say:

    “Here we have 107 years of statistics, and who can discern from them the rule that guides the weather ? A matter which enters so largely into our health and comfort, happiness and prosperity, that I hope to be excused for thus dwelling upon it.”

    He knew of the Qld deluge of 1841 which came on the heels of that century’s worst known drought. Magnificent Wivenhoe’s chief hydrologist thought it was worthwhile knowledge after 1974. He built for 1841 and what might have come before 1841. It beats me how people can talk about climate as a short-term affair

    Have a go, Jen. I’m guessing it will be a hard haul for an uncertain result, but why not give it a shot? Beats sitting at home watching the Flintstones.

  45. cohenite July 13, 2014 at 11:25 pm #

    Happy to talk ANN; perhaps you can tell us all how a backpropagation algorithm differs from hindcasting? Then we can talk about how Koutsoyiannis’s techniques differs from those used by CMIP.

  46. Luke July 14, 2014 at 12:15 am #

    No idea Cohenite why don’t you impress us with a dissertation. I thought back propagation was how networks where trained with some optimisatiion method. Are Levenberg-Marquardt methods still kosher?

    But it’s all a bit deus ex machina.

    I simply want to know whether we have a useful rainfall forecasting system for Qld rainfall or not. i.e. the TOPIC !

    Given your obvious penchant for the subject perhaps you could advise Bazza and myself do we plant, agist or sandbag on the current ANN forecast. (And given Robert is still into sock theory Jaysee might be keen to know too)

    You see David Evans has told us to prepare for cooling so we’ve got that one on the fridge door.

  47. Luke July 14, 2014 at 12:17 am #

    Robert you might advise us if 107 years is enough of a sample. And what would you recommend if the base climate indicators seem to be drifting? Just asking?

  48. Robert July 14, 2014 at 12:58 am #

    It was 107 years at time of writing. We’ve now had over 230 years since Flinders’ reports – mostly of drought and fire – and that isn’t enough. And the base climate indicators will always be drifting. Tough gig, but if someone wants to take a shot (it won’t be me) I’ll say good luck to them. Anything beats the Byzantium of CO2 science.

    Just remember that a few years after the Hawkesbury rose to a phenomenal 101′ at Windsor (1806) you had to pay 8d for a bucket of water in parched Sydney. (It wasn’t Evian.) One year separated Black Thursday 1851, likely the world’s biggest known fire, and the Gundagai Flood. Our longest drought, oblivious of PDO etc, peaked in the such bad conditions in 1967 that nobody would have believed the transformation to come within a few years. Then there’s the withering spring/early summer of 2012 following our massive bucketing. (I could go on, but I know you hate this stuff.)That’s a sampling of what a serious researcher has to deal with and what I’m glad not to deal with. I’ll just be a tyre-kicking customer, thanks.

    If Jen and John have a stomach for this and if the cost is minimal to the rest of us, why on earth would I complain? And what on earth are you complaining about? Just asking.

  49. Joseph A Olson, PE July 14, 2014 at 6:23 am #

    I also attended the Heartland ICCC event in Vegas and was honored to her Dr Marohasy deliver her Rainfall Forecasting and her AGW Debate in Australia presentations. This is a three sided debate between the Darth BIG Warmists, the Luke LITTLE Warmists and the Obie NO Warmists. Both Darth and Luke believe in the magic gas hypothesis of back
    radiation warming, only disagreeing on the degree of warming.

    Traditional Thermodynamics does not support either position. My analysis of the ICCC #9 is posted here….

    Find and share Truth….it is your duty as an Earthling.

  50. spangled drongo July 14, 2014 at 7:47 am #

    Win for anti-boganism in the world cup:

    Germany 1 [no tatts]

    Argentina 0 [all tatts]

  51. Luke July 14, 2014 at 7:52 am #

    And for Toby – as for your assertion “well pretty well everyone believes in CO2=1.2C” – see Mr Olson above, Exhibit A, on more slayer type stuff. And another colourful US style web site in the “it’s all fraud” tradition. Very nice. However, one still does wonder how all those science people doing radiation balance studies seem to have got the right answer for the wrong reason, including the USAF, but better not think any harder or one might turn to stone. Won’t say anymore as this topic has had reams written about it. Wattsy, Spencer and Nova even shout “no more slayer stuff – pullease!”. But that’s OK – it’s all good.

    Robert – “I could go on, but I know you hate this stuff.” not at all – only when you use such to prevent any progress. And a “oh well somethin’ happened somewhere sometime before so I’ll use that as an answer to any analysis” – hardly a statistical analysis. But keep going – you’re a treasure trove and good fun.

    “And what on earth are you complaining about?” Basically spruiking that one has a rolled gold rainfall forecasting system when you can’t even see it in action. There would also be the many science questions in previous posts unanswered. And the gratuitous rhetorical fluff accompanying the said proclamation. Otherwise why blog it – turn off comments and so we only may make polite applause.Perhaps a one way “Like” button.

    Same with all the lunar and solar stuff rainfall forecasting guff – it’s all bunk – you can never really “see it” in action, often samples size of 1 (Bazza’s n=1 comment), and usually surrounded with a nice gratuitous wad of anti-AGW rhetoric for a diversion. You could cut all that filler out and tell us more about the actual forecast system. This style of science makes one wonder if it’s really just an anti-AGW beatup. As for “new” paradigms Robert – there’s only about another 18,000 Google Scholar examples of ANN climate style science.

  52. Luke July 14, 2014 at 7:53 am #

    Spangled – just more n=1

  53. cohenite July 14, 2014 at 8:50 am #

    Patterns in climate are everything; so far no perturbation of the patterns by CO2. So luke what are the PDO and IPO doing; throw in a bit of SOI and your rainfall forecast should pop up before your bloodshot eyes.

  54. Robert July 14, 2014 at 8:57 am #

    Firstly, when I talk of stuff which happened I am not offering statistical analysis (or Persian miniatures or wigwams). I am talk of stuff which happened. Got that?

    As for Jen and John, I assume you’ll see something in action WHEN it is in action. You will not see something in action UNTIL it is in action. Got that?

    Supe, this is getting like the back-up generator and the five 80c fish pieces for $4. It’s like your thundering contradiction of Toby where you weren’t actually contradicting (just thundering).

    Let’s try one more time.

    Even if there is the promise of rolled gold rainfall forecasting system – though there has been no mention of “rolled gold” – you will NOT see it in action UNTIL it is in action.

    Same as you can’t have dinner at breakfast time or the Rugby League GF in February. Got it?

    Not my system, and has nothing to do with skeptical old me. It is Jen and John’s undertaking. I wish them well and if their forecasts hit home over a number of years I’ll start paying closer attention. And if they are light on everybody’s pockets and don’t make manipulative claims and excuses I’ll be happy to see them trying indefinitely. Even without immediate success they’ll probably learn a lot from facing what you and bazza dislike so much: stuff which actually happened.

  55. Neville July 14, 2014 at 9:02 am #

    Geeezzz ya gotta laugh. They now try and tell us that OZ drying out is caused by AGW. But southern OZ has been drying out for at least 5,000 years according to Patrick De Deckker’s 20 year study. And I suppose if OZ were to wreck our economy and reduce our co2 emissions by 5% we’d notice a change? What a mob of barking mad numbskulls.

  56. Neville July 14, 2014 at 9:21 am #

    BTW Luke, Brozek used Nick Stoke’s info to select for statistical significance and found no SS in all the data bases for about for 19.2 years. That’s the average of 4 trends.
    That doesn’t include the endlessly adjusted Giss nonsense of course. If you include Giss it drops to 18.2 years. See below………….

    Section 2

    For this analysis, data was retrieved from Nick Stokes’ Trendviewer available on his website Nick Stokes’ Trendviewer. This analysis indicates for how long there has not been statistically significant warming according to Nick’s criteria. Data go to their latest update for each set. In every case, note that the lower error bar is negative so a slope of 0 cannot be ruled out from the month indicated.

    On several different data sets, there has been no statistically significant warming for between 14 and 21 years.

    The details for several sets are below.

    For UAH: Since February 1996: CI from -0.017 to 2.347
    For RSS: Since November 1992: CI from -0.016 to 1.857
    For Hadcrut4: Since October 1996: CI from -0.010 to 1.215
    For Hadsst3: Since January 1993: CI from -0.016 to 1.813
    For GISS: Since December 1999: CI from -0.004 to 1.413

  57. bazza July 14, 2014 at 9:47 am #

    Robert suggests this sort of research is pretty harmless because it is light on the taxpayer . He must have innocently assumed that the Macfie Foundation which is a generous funder of A&M is not set up to achieve tax exemptions. There are also the huge overheads incurred by university researchers picked up by taxpayers.

    He is correct though in that there is no drain on other government or industry funding except perhaps for a bit from the energy industries. As Dr Bryant Macfie said “ Sadly work of this kind would be unlikely to receive funding from conventional sources,” . Perhaps if it used conventional rigorous processes for data analysis the odds would be better. Perhaps also if either author was prepared to respond to genuine queries on their methodology rather than rely on sundry sycophants. That is why research gets done – to get it out and tested – only then can you claim some expertise that some even dare to use in other arenas!

    As for this being a new paradigm most of the skill comes from the SOI and that or similar has been in use for a century.

  58. spangled drongo July 14, 2014 at 9:55 am #

    Neville, Luke’s mileage may vary.

  59. Neville July 14, 2014 at 10:30 am #

    A clear majority of Aussie’s want to get rid of the idiot’s co2 tax. I think Palmer and his PUP runts should wake up.

  60. Robert July 14, 2014 at 10:46 am #

    Sycophant is a strong word, bazza.

    Bazza, just review what I wrote and not what you imagined I wrote. IF it’s light on the taxpayer does not mean that it IS or WILL BE light on the taxpayer. “If” is not “because”.

    As to the level of my enthusiasm for any system, I think numbers are pretty dumb without sensible people to interpret them with great skepticism and climate numbers are stupendously dumb without etc.

    SOI is a perfect example of airheads using a useful but rough observation set as a reliable mechanism. The number of neutral and even La Nina years, especially from the mid 1920s to 1940, when drought has afflicted large slabs of eastern Oz can’t sway them from their literalism. They seem unable to drop the joystick and use their loaf.

    The problem is not with Walker’s foundational work. The problem is with shallow minded manipulation. As a total skeptic I couldn’t care less about some warming post-1970s or a pause post-1998. I find it all too frivolous for words. Sea level rise? Pretty paltry, requiring some particularly twisty manipulation. You could walk from Melbourne to Tassie a brief ten thousand years ago, so really…

    The real answer is for the heavily resourced BoM to drop the political guff (and Jones). If it gets some debate and competition from the likes of Jen and John, so much the better. But let us know if the pair are creating “huge overheads”. Maybe you can run the costs against a rusting tidal power plant sinking obscurely into the sand somewhere. Or a rusting de-sal plant chewing coal power uselessly.

  61. Luke July 14, 2014 at 12:26 pm #

    So again a bunch of diversionary OT comments. Bazza the cheer squad here will do anything to stop talking about ANNs for rainfall forecasting in Qld. Telling isn’t it. Time for a few Flannery’s and Al Gore’s to lay more smoke to escape a discussion.

    Come on Robert – it’s already claimed to better than BoM yet we can’t see the system in action ROFL and LMAO – probably coz it doesn’t work..

  62. Jennifer Marohay July 14, 2014 at 12:43 pm #


    We heard you the first time. We heard you when you said that my ideas on rainfall forecasting had no value until they were published in the peer-reviewed literature.

    Then we heard you when you said my peer-reviewed publications had no value because I had not benchmarked against POAMA.

    Then we heard you when you said I needed a fully fledged operational weather forecasting business or my benchmarking against POAMA, that had been published in the peer-reviewed literature, was pointless.

    While we appreciate that you have an opinion and goal posts that are forever changing, could you and Bazza please desist from misrepresenting my work, hijacking comment threads at this blog and generally being a spoiler.

    This blog is not about you.

    In the interests of diversity of opinion: Comments from you and Bazza that are repetitive will be deleted. If you persist I shall insist you post here under your real name. Because if you want to spoil my blog threads, and attempt to white-ant interest in my rainfall forecasting using artificial intelligence then you should at least declare who funds you… which interests you are so keen to protect and why.

  63. Neville July 14, 2014 at 1:06 pm #

    Good points Jennifer, but I don’t think upside down Lukey will ever wake up. His totalitarian instincts and fundamentalist religion doesn’t allow it. BTW a new study has found that the USA is responsible for 0.2c of global warming since 1850. Boy do they reckon??? What a mob of jokers.

  64. Mr Koala July 14, 2014 at 3:36 pm #

    You need to face reality Luke. POAMA’s rainfall forecasts have such low correlation coefficients when compared to observed monthly rainfall, that essentially its a way of producing noise. Its a very expensive form of noise production. All those government workers being paid to generate noise. Its this why you identify so much with the POAMA boys?

  65. bazza July 14, 2014 at 3:47 pm #

    I am selfie funded Jen and also highly sceptical about all new research of interest to me. I wrongly assumed as a scientist that you would welcome dialogue and even questions on your novel methodology as otherwise there is no point other than joining in and saying “whoopee” we have a “new paradigm”. I do note your claim in Vegas on natural climate cycles persisting – I reckon you mean the seasonal one and I hope it continues to persist.

  66. jennifer July 14, 2014 at 4:08 pm #

    Hi Bazza

    Its my understanding that your a statistician employed full-time by government. If this is not the case could you please send me your CV and/or other details confirming your identity and how your self-funded.


  67. sp July 14, 2014 at 4:09 pm #

    Well said Jennifer.

  68. jennifer July 14, 2014 at 4:15 pm #

    Everyone else,

    I’m not suggesting I have a “new paradigm”. I’m suggesting that there is a need for a new paradigm. I’m suggesting that artificial neural networks (ANN) could be a useful tool underpinning a new theory of climate that was concerned with better understanding and quantifying natural climate cycles.

  69. bazza July 14, 2014 at 4:27 pm #

    Jen, what a hide, as if you warrant an understanding of my status. The one you have is completely wide of the mark in any case. I told you I am self-funded and if you don’t believe me that is your problem.

  70. jennifer July 14, 2014 at 4:30 pm #


    I am continually accused of being a shill for big oil. If your a shill for government, then its time your declared as much.

    This is a private blog site, if you want to continue posting here, then you need to declare your interests.

  71. jennifer July 14, 2014 at 4:38 pm #

    Everyone else,

    If you want to know where I was born, where I went to school, what my maiden name is, what I’ve published, who I’ve worked for, how I am currently funded, then the information is here…

    You will see that my original training is as a biologist, that in the past I’ve worked for government and industry and most recently on climate science issues funded by the B. Macfie Family Foundation.

  72. Neville July 14, 2014 at 4:49 pm #

    Even more junk science , superstition and witchcraft to fool the population about increased extreme weather events, that don’t exist. That’s anymore extreme than one hundred years ago. And of course the poor will cop it in the neck by paying a higher cost for everything, while the banks, big business, insurance companies etc are smacking their chops because they’ll be making 100s of billions $ more at the expense of the little bloke. What a con and fraud and zip change to temp, climate and extreme events.

  73. Luke July 14, 2014 at 4:56 pm #

    I fail to see how asking pertinent questions is being a spoiler. Spoiling is letting the thread fill up with all manner of off-topic material which the others here love to do. And I am suckered into rebutting.

    I’m not interested in the protecting the status quo, POAMA or any system.

    Bazza of all people loves outliers and high odds horses. Myself I have even suggested CVAP funding!

    I haven’t been enthralled with POAMA since release but that is a very short period of time to make a serious judgement. BoM are frustrating in the lack of clear validation data albeit there is some.

    I cannot understand why some simple questions cannot be answered and being stonewalled results in frustration.

    Questions that have been asked are not to do an “ah-ha” “gotcha”. They are relevant to understanding. Bazza has made a point about the seasonal cycle and not using anomalies. He has asked about the process of station selection.

    I have asked previously if POAMA is used operationally in the way that you tested – how rainfall was downscaled and was this a “special” one-off. I have asked about the number of comparisons. I have asked about POAMA versions.

    Then there is the question of changing index baselines – i.e. what do we do about non-stationarity in the SOI and warming SSTs around Australia. I understand a warming Indian Ocean trend caused a rethink in an earlier BoM statistical model.

    Related – if there is an centennial warming trend and EOF 1 from this most basic analysis says there’s a biggie then won’t the ANN will train on that too. (EOF2 being the PDO).

    And if this new paradigm system is good enough to be showcased internationally I don’t understand why one cannot see it in action – perhaps firewalled as experimental and protected as access upon request if needed. Otherwise all the rhetoric about long suffering farmers is simply rhetoric (surely?).

    And POAMA doesn’t pretend to have good skill all the time. Which is why the probability maps need to also be interpreted with the skill map.

    Are any of these questions unreasonable or slick? It is simply about the science.

    But there are no answers or engagement. So the thread lead is simply a statement which you do not wish to have discussed. Discussion which used to be so good here.

    Anyway as you have now at least acknowledged that you have indeed heard the points so I will desist.

  74. Luke July 14, 2014 at 5:11 pm #

    And I missed this “I needed a fully fledged operational weather forecasting business” – no I didn’t say that all. Would a monthly run of the model for 20 Qld rainfall stations be a fully fledged operational weather forecasting business” like WeatherZone – I don’t think so !

  75. Mr Koala July 14, 2014 at 5:45 pm #

    POAMA tells us, for example that there is a 50 to 55% probability of above median rainfall. Are you really going to shut down the mine on the basis of this information? This is what happened in November 2010, then the heavens opened and the rains came pelting down in December.

    The problem with these types of probability forecast is they do not differentiate between 1% above the median, and 500% above the median. It allows the forecaster great latitude in self congratulation for producing a “correct:” forecast without actually providing very much useful information for the end user.

    Its like forecasting the winner of the Melbourne Cup as a horse with 4 legs and a tail. Its a “correct” forecast year after year but really useless if you want to place your bet and win some cash.

    As with so many government schemes – pink batts, set top boxes, school halls, assisting the aboriginal community, its all about government congratulating itself, not about efficient, cost-effective delivery to the paying public.

    That’s why the $25 set top boxes ended up costing $400 each!

  76. bazza July 14, 2014 at 5:49 pm #

    ” This is a private blog site, if you want to continue posting here, then you need to declare your interests.”
    I will declare my interests if that is the policy for all and if Jen can justify why I have been singled out. But why drop in if Jen cant handle questions? Come to think of it, I have nothing to declare apart from I am interested in the truth.

  77. sp July 14, 2014 at 7:04 pm #

    Je has declared her interests Bazz, so should you. It seems you cant handle questions.

  78. Emily Fraser July 14, 2014 at 7:49 pm #

    I have just emailed Dr Marohasy with my Resume.

    Bazza, You do carry on like you have a reason to undermine everything written here. It would be interesting to know why you do that. I assumed you worked for Al Gore.

  79. cohenite July 15, 2014 at 9:15 am #

    “Then there is the question of changing index baselines – i.e. what do we do about non-stationarity in the SOI and warming SSTs around Australia. I understand a warming Indian Ocean trend caused a rethink in an earlier BoM statistical model.”

    What about them luke? Accounting for natural variation doesn’t preclude accommodating for Stochastic aspects of climate. Various statistical technigues such as Bayesian and Monte Carlo can do this. The problem is the assumptions underpinning AGW taint any modelling used to predict on the basis of AGW. That is POAMA’s issue; it’s assumptions.

    As I understand it Jennifer’s model uses prior patterns of climate without filtering them through AGW style assumptions. Extracting the dominant causes is then related to establishing past correlations.

    Stewart Franks has been doing this for years:

    Your epiphany about AGW is overdue luke. Resist the dark side luke; come over to the light side luke! You can still satisfy your opposition defiance disorder at Deltoid!

  80. bazza July 15, 2014 at 10:17 am #

    Koala raises an interesting point back at 3.36 pm 14th “POAMA’s rainfall forecasts have such low correlation coefficients when compared to observed monthly rainfall, that essentially its a way of producing noise”.I think this was the point made in your latest Atmospheric Research about low correlations between predicted and observed for seasonal forecasts as evidenced by some interpolated output from the previous version of POAMA.

    I agree correlations are often are more intuitive way to communicate skill given they scale 0 +/-1. Correlations also show up important times of high skill whether due to chance from a small sample, or artificial skill, or hopefully a strong signal from an ENSO indicator. Seasonal forecasts are mostly of low skill as any attempt to gain understanding shows. The trick is not to throw out those high skill babies!

    I think your A&M article only mentions one ANN correlation, that is a high one for Cairns over a short period which on a quick read is probably for the 12 months pooled.
    If you have other correlations just of results for one forecast season at a time that would be useful to clarify Koala’s comments on noise etc. For example referring to Fig 2 p175 the lower left hand plot is for Ayrshire Downs. What is the correlation for the summer (DJF) forecast for the years plotted.

    If you don’t have that available, I would like to get a copy of the data as plotted for Ayrshire Downs pls.

  81. Minister for Truth July 15, 2014 at 2:51 pm #

    I think the blog owners comments above regarding Luke’s persistent and obsessive behaviour are absolutely spot on …its quite characteristic of vexatious litigants.

    This blog isn’t about him and his agendas and his abuse, and the childish attempts at satire only serve to underscore what may be the real agenda, and who his paymasters are.

    He reminds me so much of the ABC journalists …and because of their poor standards no longer bother watching any ABC current affairs programs …ditto any blog where vexatious idiots dominate

  82. Luke July 15, 2014 at 5:58 pm #

    “what may be the real agenda”

    Or not !

  83. Toby July 15, 2014 at 7:04 pm #

    Without Luke it would a be pretty boring blog, and Jen has posted to get feedback, criticism and ideas I would assume. How some of the questions are phrased could at times use a softer touch, but many of the questions do not seem unreasonable even if they may challenge Jen’s work.

    For instance it seems pretty reasonable for us to be shown forecasts for the future so we can try and gauge its usefulness.

    Most of us here question the reliability of models and feel that many of them are tweaked to be able to “hindcast” and show a “track record”. Clearly current forecast methods are not much use according to farmers and other who rely on them.

    A web site with monthly predictions sounds like a good idea to me? It is time for a change in paradigm and hopefully this may be a prompt for such change!

  84. Mr Koala July 15, 2014 at 7:57 pm #

    With any model it is only possible to compare observed (i.e. past values) against modeled (or predicted values). Values currently observable in the future inevitably become past values at the point where they are available to make any comparison. Think about it.

    Perhaps Luke has a time machine so that he can voyage into the future and send us back rainfall observations before they actually occur. Queensland government workers can perform miracles! Perhaps they should have tried this approach to operating the Wivenhoe Dam and all those poor folk would not have been drowned.

    Shao and Li (2013) report even lower correlations between observed and computed monthly rainfall forecasts for downscaled POAMA 1.5 output for 35 locations in the Murray Darling Basin. They report a mean r = 0.38 for monthly rainfall forecasts 1 month ahead, dropping to r = 0.192 at 3 months. These results independently strongly support the conclusions of Abbot and Marohasy that the downscaled forecasts from POAMA 1.5 struggle to even match climatology 1 month in advance. Such low correlation coefficients are indicative of generating little more than random noise. If the later versions of POAMA are greatly improved- why wont they publish these results. Put up or shut up.

    Shao and Li, 2013. Stochastic Environmental Research and Risk Assessment 27(4) 819-830

  85. Toby July 15, 2014 at 8:26 pm #

    Surely the purpose is to make a prediction and then when we reach that point in time see if reality matches the prior prediction!?

    If it does make an accurate prediction consistently (or at least a more accurate prediction than POAMA etc), then surely that would confirm the potential validity of the AI model.

    I wouldnt trust a climate model that showed it had accurately predicted the past because I would just believe parameters had been tweaked to reach the reality.

    When I ran an FX dealing room I was forever being shown models that would make me and the bank a fortune. Sure they were able to show how accurately their systems had “predicted” the past, but did I trust them? Of course not, if their models are that good why sell them and not just use them to make a fortune for themselves?

    So show us some predictions for the future, and then when we actually get real data for that time period check for quality/ accuracy. I never saw a model that was able to predict the future with any consistency, but I saw heaps who got the past “perfect”.

    The same problem faces climate models.

    And to answer a statement from Luke from a few days ago, of course models and computers have a purpose, when it comes to engineering and known parameters etc then modelling can be useful. I wouldnt want to drive across a bridge that hadnt been “modeled” , or be in a sky scraper where the same wasnt done. But I would not trust an economic model, or financial model because there are so many unknowns.

    Hence my lack of faith in climate models for very similar reasons.

  86. Luke July 15, 2014 at 9:44 pm #

    Jeez Bear – you’re amazing. Drownings !? – get atlas – is Grantham flash flood affected by Wivenhoe. You remind me some sort of idle university student that is sketchy on basic facts? Talk about making stuff up.

    Like bogus POAMA comparisons. Does BoM use POAMA 1.5 downscaled data (or more precisely some cherry picked personal cooked up variant) as its online prediction method.

    Just spare us the bunk and answer the question !

    Is the method used by A&M also used by BoM in current operational forecasting.

    Must be so onerous to publish a table for 20 rainfall stations each month from an existing model. So much work. Universities with their laid on taxpayer paid access to the national computing infrastructure. Drowning in spare cycles.

    Mate people have been coming with rainfall forecasts since Federation. All great for a yarn.

  87. Luke July 15, 2014 at 9:52 pm #

    Toby – one can ask questions but no answers are forthcoming. It’s appeal to authority of “it’s published”. Which is what sceptics routinely reject.

    Toby – laid on facilities for unis. linked up with

  88. Luke July 15, 2014 at 10:00 pm #

    Toby – in terms of your comment, “models tweaked to give a tuned hindcast”. Typically one excludes a sub-set of data for independent testing.

    Now it depends how strong you feel about that science – i.e. if it’s ready for field use. If the level of spruiking is pretty high one would assume it would be available. And it’s not reasonable to allude to the poor suffering farmers if nothing is offered for them to to try.

  89. jennifer July 15, 2014 at 10:20 pm #


    Our focus has been, and will continue to be for the immediate future, on improving on our system of forecasting/our prototype model and getting stuff published in the peer-reviewed literature. We are also now attempting to forecast ENSO.

    We also remain keen to benchmark against POAMA, including by way of future predictions.

    But to do this we need the BOM/POAMA to also provide deterministic (as opposed to probabilistic) forecasts for some real locations in Queensland (as opposed to 250 square grids) for some period into the future.

    We have repeatedly said that we would like to benchmark against the latest version of POAMA for one month periods up to three months in advance.

    In July 2011 the Bureau at least provided us with hind casts for 17 sites in Queensland against which we could benchmark. The Bureau has since refused to provide us with any information by way of specific rainfall totals forecasts for any location. Their official forecasts are all presented as probabilities, which is nonsense.

    This refusal to enable us to benchmark, even against hindcasts with output from their latest version of POAMA, was most recently communicated to me by way of an email received on July 4, 2014.

  90. Luke July 16, 2014 at 12:59 am #

    So BAM3 underlying POAMA is at a 250km resolution.

    If so this means Cairns to half way across Cape York are in one grid box, Rocky to Emerald,

    Brisbane to the other side of the Darling Downs.

    So the rainfall gradients across that distance are large. So a method of downscaling has to be developed to get to a point. What is that method? Is that what BoM uses operationally or a custom hack to get some data. How good is that downscaling approach.

    Downscaling is a whole field in itself – a major field with many approaches – e.g.

    So this is essentially beating a version of POAMA into shape with another analysis. Really a test of the downscaling technology more than anything.

    So we have a grid box POAMA with a screwed on downscaling analysis vs an ANN analysis. Surely this is total hookey in terms of a meaningful comparison.

  91. spangled drongo July 16, 2014 at 7:52 am #

    “I wouldnt trust a climate model that showed it had accurately predicted the past because I would just believe parameters had been tweaked to reach the reality.”

    Toby, yes but we’re supposed to believe these that are so far out on the past as well. The only point they agree upon is the point of start-up.

    Feed ’em the known knowns and they still fail:

  92. Mr Koala July 16, 2014 at 9:13 am #


    Does the Queensland government actually pay you to sit there and post on blogs all day?

  93. Luke July 16, 2014 at 10:10 am #


    Does some uni or big oil employ you to waste time?

    For you it’s all just some game. For these folks it’s not.

    So instead of having a quite reasonable science discussion you’d like to deflect any discussion. Having come a cropper on bogus drownings Bearsy tries the ad hom. And still no answers to reasonable questions of clarification.

  94. Luke July 16, 2014 at 10:14 am #

    Spangled – and what you only see in that try-on graph are model means not the individual ensemble runs (i.e. a mean is made of up of averaged individual ensemble runs).

    You are comparing a single ensemble run i.e. actual real world observations not against what individual runs may have produced.

    It’s slick sophistry and you do need to always get some other perspective.

    However, any reasonable modeller would still concede it’s on the low side.

    Ask Bear to explain it to you – he has all day to play tag with me,

  95. sp July 16, 2014 at 11:02 am #

    Luke – post under your own name and declare who funds you, and if you work for the public service.

  96. Luke July 16, 2014 at 12:27 pm #

    sp – do the same declare your big oil interests

  97. Luke July 16, 2014 at 12:34 pm #

    It’s interesting isn’t it – sp could attempt to discuss the science but in typical cheer squad fashion tries to shut down the debate. I guess that’s what’s you expect isn’t it. If cornered and found wanting – go the ad hom. You see sp – I don’t care who you are only except what you have to offer the discussion.

    Which is nothing.

    Do you want to discuss science of simply be a part of a tribe and sense the inner warmth from the group. Self-reinforcing high fives and back slaps.

  98. sp July 16, 2014 at 1:32 pm #

    You are here as spoiler Luke – you are not contributing.

    You have again implied Jen and Bear are shills for big oil.

    For credibility you need to declare who you work for (government), who funds you, and your credentials and experience in “climate science” – you do come across as a bit of a know-it-all expert with a habit of providing useless links.

    You just dont like it when anybody challenges your notion of settled science. You get all cranky.

    So, why dont you answer one question – do you work for government? A simple yes or no will do.

  99. Luke July 16, 2014 at 2:24 pm #

    Please apologise – where I have accused Jen of being a shill. More lies from you sonny. If you weren’t a content free you would have noticed over the years I have said that Jen isn’t a shill – she believes in what she believes in. Her world view is hers and she is entitled to it. Easy enough to accept.

    Stop being a content free bully and talk some science or get off the thread. Last time I will respond to you if comment is not science.

  100. sp July 16, 2014 at 3:56 pm #

    Luke. Check your comment at 10.10. Stop bluffing. Stop lying.

  101. bazza July 16, 2014 at 4:15 pm #

    Jen, back at 10:17 am on the 15th, I asked whether the data on which Fig 2 was based was available.? Could you pls let me know and I will follow up directly with the corresponding author if that is what is required.?

  102. Debbie July 16, 2014 at 4:17 pm #

    Luke @ 10:10.
    What has any government funded climate research done to help agriculture in a useful or practical manner?
    You seem to be suffering from a memory lapse.
    Have you forgotten what CSIRO & BoM have been proclaiming for Eastern Australia over the last decade?
    They put out lots & lots of stuff and claim they’re being useful. . .but the actual farmers are becoming more & more amused and less & less impressed.
    Your ‘just a game’ comment applies accross the board.

  103. Luke July 16, 2014 at 5:02 pm #

    sp 10:10 does not refer to Jen – so pls apologise and last engagement with you if not making science comments.

    Debs – well far be it from me to defend the entire domain of govt applied climate research for agriculture or POAMA or BoM or CSIRO – but I look forward to your detailed critique of which would have roots back to the 1990s. With some apparent engagement by producers and agribusiness.

    Clearly some people are impressed enough to participate and build relationships but perhaps they are the ones who actually may wish to engage positively and produce some decision analytics – vis a vis sledgers and political spectators.

    And one must say under the leadership of the much maligned D Jones one seems to have a 1000% improvement in data availability. As compared to stricter times. Perhaps of little value but handy for those wanting undermine or constructively critique BoM’s assessments.

    Of course there may be many long unpaid lost hours and weekends making urgent climate business cases by state agencies to convince a cynical Treasury to provide billions of dollars in drought aid over decades. Billions maybe not of much use….

    As for a drier Eastern Australia – well the long term anomalies are still there. And we are not “there yet”. Droughts and flood cycles play out over decades. For those who do follow the science there does seem to be some long term movement in the sub-tropical ridge intensity and changes in the Southern Annular Mode and a warming Indian Ocean. Ongoing these will have some impacts. Floods and droughts will remain but in what proportion. I think if you carefully read the science you will not derive that “it will never rain again” . Press clips, doorstop interviews with tired staff, op-eds are not considered science nor are they decision analytics. I assume you don’t run your business on press clippings. And the science is what it is. Sometimes forecasts simply won;t have skill.

    All off-topic of course. And such spoiler comments. Screaming shill and so forth. Weally?

  104. Robert July 16, 2014 at 5:31 pm #

    As for a drier Eastern Australia – well, we got there! Between 1895 and 1950, give or take some years.

    Don’t get excited, however, about those 1950s and 1970s. The history of Oz drought in those “good” decades is almost as hair-raising as what came before, after, and even in between. Almost.

    Of course, we can only guess about Eastern Australia’s parched years which corresponded with a profound change to a wetter climate in North America between 1825 and 1840 (about which we can speak in some detail). Bloody climate change!

    Silly climate. It never reads any of the literature, holds very loosely to a few rules…but just does whatever it wants when it suits.

    Sack the climate.

  105. Luke July 16, 2014 at 6:03 pm #

    If one notes the southern wet season rainfall decile map from 1996 here

    It shows a significant rainfall deficit period (ongoing Debbie) in southern Australia and SW WA. Despite some record rainfall. It is not however, a drought map of current conditions.

    I hope we don’t have to go over deciles 101 again but it clearly shows areas that are lowest on record. i.e. that this 17 years period is drier than any other. Could be just bad luck as Robert would tell us. Or there may be some underlying influence and that influence may have causes which are natural or anthropogenic or a combination.

    Climate modelling says greenhouse forcing is having an effect on STRi and SAM. And that explains some of the deficit.

  106. sp July 16, 2014 at 6:25 pm #

    Luke at 10:10 (to Mr Bear)

    – “Does some uni or big oil employ you to waste time? For you it’s all just some game. For these folks it’s not.”

    Me at 1:32

    – “You have again implied Jen and Bear are shills for big oil.”

    Stop bluffing. Stop lying. Do you work for government – yes or no?

  107. Robert July 16, 2014 at 6:41 pm #

    “…there may be some underlying influence and that influence may have causes which are natural or anthropogenic or a combination.”

    Well, let us know if it’s a bit anthro. We can do something about that. “Doing something” won’t involve whirlygigs, feed-in tariffs, carbon taxes, EU rascals, Timmythermia etc. Too much good money and too much good coal going to waste on all those white elephants.

    As for the disastrous conditions of the late early 1790s, late 1830s, Fed Drought, WW2 drought…such droughts may not have been just bad luck either. When you think that those were just the biggies, and that there were shocking droughts all through the 19th and 20th centuries…

    Hey, I know some people hate discussion of stuff which actually happened but…

    Maybe they had SAM and IOD and ENSO and PDO and all kinds of “influences” back then! Maybe the influences came first and the acronyms came later! Scary.

  108. Luke July 16, 2014 at 7:44 pm #

    Well Robert I know you would never like to know the origin of the black plague and just like to regale us with how many people died at each outbreak and what they were were wearing and religious preferences. Probably bad air hey. But you know science is curious and likes to know a bit more of why and how.

    Why oh why didn’t we just stop at phlogiston. Was much simpler and nicer.

    And how dare Watson and Crick come up DNA – how boring.

    Yes state of the art science says it’s a bit anthro.

  109. Luke July 16, 2014 at 7:51 pm #

    Experiments for Robert – get dressed in black – go out into the midday sun and report how you feel, then repeat with all white. Report back.

    and look Robby – no hands – magic ! bad air….

  110. Robert July 16, 2014 at 8:27 pm #

    Supe, leave Crick and Watson out of it. It’s a bit like people on the skep side calling things after Galileo, and biochar salesman Turney daring to use the name of Mawson. We need to lose the heroic posturing.

    Anyway, you and I don’t seem completely apart. I believe in all those climate influences, including anthro, and you believe in them. It’s just that, like the climate in general, you think they’re new because they have new names. You also on occasions seem to believe they can be manipulated by a new collectivist weapon called, mystifyingly, The Market.

    That link you suggested? I won’t hit it. Just to keep you guessing – and save myself a likely irrelevance.

  111. Luke July 16, 2014 at 10:00 pm #

    No never hit links – you can be turned to stone.

    No I don’t “believe” in them – some evidence has been demonstrated. But if you didn’t hit the link I wouldn’t worry.

    new because they have new names – errr nope. Because they are trending and trending coz why? Of course maybe some of the names may not even really exist – they might be simply statistical debris artifacts – sort of like seasonal forecasting 🙂

    “Manipulate” errr nope – implies precision. More like – can we not keep throwing large rocks on the frozen pond surface. I’m not sure how many more it can take and cracks are appearing.

  112. sp July 16, 2014 at 10:38 pm #

    “I’m not sure how many more it can take and cracks are appearing.”

    Pure alarmism.

    AGW -the wonder product that does so much – can heat or cool, make wetter …. Heres one for you Luke, a hole in the ground, yep, the cracks are appearing:

    “The cause of the hole’s sudden appearance in Yamal – which translates as ‘the end of the world’ – in the far north of Siberia is not yet known.

    There has been web speculation about the crater indicating ‘the arrival of a UFO craft’.

    But one Russian expert says the cause is more likely to be global warming releasing gases under the surface, which then explode like a champagne cork.”

  113. Robert July 16, 2014 at 10:40 pm #

    To think all those influences did it in the past without, errr, trending! They were there in 1902 etc, but they were just hanging about. Even when they were baking India as well as us, back in the Fed Drought years. An Englishman called Walker who was heading up Indian observatories thought there must be “influences” behind monsoon failure and the 1899 famine. But there’s no way those influences would have been, errr, trending. They were probably just doing the Lambeth Walk in whalebone corsets – or whatever one did back then.

    Anyway, that was a long time ago, way back in the pre-trend era. Right now, we must find that “Robby” who’s indifferent to medical research on plagues and is bored by the DNA work of Crick and Watson. He sounds like a reverse-Deltoid. But where is he? (Probably out chucking rocks at the climate.)

  114. Luke July 17, 2014 at 1:35 am #

    Well the TREND with the plague is for less due to anthropogenic influence, despite it always have previously been trending.

    As being worried about bad air may not have been that useful to solving the problem.

    And we certainly also seemed to have got in on with smallpox which was also a big trendsetter for a while.

    You can see a lot and never understand anything. So I guess they come up with this called science which … but then to do that you’d have to think they may be hidden forces at work and click random youtube links….

    Nah ! Its’ crap.

  115. Ulric Lyons July 17, 2014 at 8:42 am #

    Rainfall anomalies at down to weekly scales can be mapped many years ahead:

  116. Robert July 17, 2014 at 8:44 am #

    What’s he talking about?

  117. Luke July 17, 2014 at 9:44 am #

    Just mocking your philosophical position of nihilism in science advancement.

  118. Robert July 17, 2014 at 10:01 am #

    So there’s this anti-science guy who doesn’t want disease research and thinks climate isn’t influenced by anything…and this guy’s name is Robby?

    Really, Supe. Sometimes you lose more face by trying to save face. What is the point of inventing someone with my name who believes all kinds of odd stuff and then proceeding to argue against that fictitious person?

    Really. Why do that?

  119. Luke July 17, 2014 at 10:05 am #

    Well was only gentle caring mocking. Just pointing out inconsistencies and a high degree of science nihilism. We’re hardly going to make any progress just suggesting somethin’ has happened before and that’s all we can ever know.

  120. Mr Koala July 17, 2014 at 10:08 am #

    In ancient times, the Oracle at Delphi was consulted for advice on outcomes of future events.

    Any inconsistencies between prophecies and events were dismissed as failure to correctly interpret the responses, not an error of the oracle. Very often prophecies were worded ambiguously, so as to cover all contingencies – especially so ex post facto. One famous such response to a query about participation in a military campaign was “You will go you will return never in war will you perish”. This gives the recipient liberty to place a comma before or after the word “never”, thus covering both possible outcomes

    The Oracle at the Bureau of Meteorology is very similar. There is a 50% probability that the rainfall will be above median next season.

    There is a 100% probability that the next World Cup will be won by a team with 11 men, including a goal keeper. So place your bet early, Luke!

  121. Luke July 17, 2014 at 10:17 am #

    Bearsy – we agreed that early on. But this is about A&M’s very interesting new work. Luckily sometimes the probability is sometimes not 50%.

  122. Robert July 17, 2014 at 10:28 am #

    “… suggesting somethin’ has happened before and that’s all we can ever know.”

    Interesting. It certainly sound like me, that first part. In fact, I’ve been shocked by what certain presumably educated people don’t know about past events and conditions. How they could pretend to science by wilfully ignoring big slices of it is beyond me.

    But as to the second part, “that’s all we can ever know”. Now I find that very strange. Somehow I think certain people just get angry at mention of the past because it undermines their main agenda of establishing modern climate exceptionalism. The anti-science accusation is just tacked on for cred.

    Someone (very sciency but very warmie) was recently pointing out to me that Pine Island Glacier had no geothermal activity to explain its movement. (All part of the collapsing Antarctic blah.) It felt very odd for non-sciency me to have to point out that, like so much of Western Antarctica, PIG has an old but still active sub-glacial volcano with an ash sheet the size of Wales, at least according to airborne ice-sounding radar done by BAS a few years back. How could he single out PIG and ignore such a massive fact about it? Or did he just not know?

    People in the grip of a stern dogma cannot think or see. All the brains and science background in the world can’t help them. Isaac Newton and others may have believed all kinds of potty things, but they could keep their odd beliefs separate from their science. What’s happening now?

  123. Luke July 17, 2014 at 10:35 am #

    Well that might involve clicking on the odd youtube link. But not evens.

  124. bazza July 17, 2014 at 3:57 pm #

    Maybe some other contributor knows the answer to my query on data availability for the Abbott and Marohasy paper. Journals like Elsevier are keen on more open access and many researchers are too.
    Mr Koala sounds like he is up on the topic – maybe he knows?

  125. Luke July 17, 2014 at 4:18 pm #

    Robert – on anthro – the presence of natural forest fires does not imply that arson doesn’t exist.

  126. Debbie July 18, 2014 at 2:17 pm #

    Luke at 5:02.
    The issue is one of focus. It has NEVER been about undermining.
    You clearly either don’t understand, don’t care or are deliberately ignoring the importance of seasonal variations and climate patterns to businesses such as agriculture, mining, construction, logistics etc.
    You seem to be more interested in defending the status quo that sneers and smears Australian businesses.
    As Jen points out. . .time for a new paradigm .

  127. Toby July 18, 2014 at 5:15 pm #

    Hi Jen sorry for my delayed response it has been a long week.

    I sympathise with your problem with finding something to benchmark against, whether it be due to broad zonal forecasts or just not being given access to the data.

    Could you not however select some sites and make specific predictions for these at say 1-3 month intervals, and then benchmark yourself against reality to gain a good idea of the quality of your AI predictions?

    If your system works reasonably well over a period of time you have proved its potential worth and may find it easier to gain research funding and more widespread recognition for your work.

    I would imagine that if your system does improve our understanding of future weather it would not take long to get a reputation where it really matters….anybody who relies on accurate forecasts beyond the current week or maybe 2?

    A year or so of good forecasts openly available for all to see in advance would make everybody have to pay attention to your work?

    Keep up the great work, if you are successful it will be a great accomplishment.

  128. Joseph A Olson, PE July 25, 2014 at 11:23 am #

    @ Luke, 7/14/14, 7:52 am

    In the interest of full disclosure, i am a retired professional engineer who has seriously studied a wide variety of Earth science with college level understanding for over fourth years. I am self funded and co-author of the science text, “Slaying the Sky Dragon” and co-founder of Principia Scientific International, where i have peer reviewed every one of the articles posted in the Publications tab. My website is always in the top listing of millions in the category of Faux Science, which includes the elitist directed Carbon climate forcing, “sustainable” energy, “peak” oil and the big bang.

    We live in a false paradigm reality, bounded by faux science, fake history, filtered news and financed with a fiat currency. There is an emerging NEW PARADIGM, and it will happen whether you believe, participate or just quibble. As for Watts, Spencer and Nova, my linked article, “Mommie, Can We Play Obombie Truth Origami” addresses Tony and Roy from my ICCC #9 in person meeting, more to come. My first post at Canada Free Press, “One Pleasant Day in Runnymede” got a supportive reply from Joanne, who i consider a colleague, with only one minor disagreement.

    There is NO back radiation warming by any three atom gas. The Ptolemaic flat Earth model had workable mathematical models, until Galileo discovered a few extra moons. The IPCC does not even have “workable” models for daily weather forecasts. The principled scientists of PSI will continue to articulate the missing empirical evidence for the AGW hypothesis.


  1. Three Facts Most Sceptics Don’t Seem to Understand | Jennifer Marohasy - July 26, 2014

    […] I explained in session 13 at the conference: Scientific disciplines are always underpinned by theories that collectively define the dominant […]

Website by 46digital