OVER the last year, John Abbot and I have worked on a new method for forecasting rainfall based on the use of artificial intelligence.
We have benchmarked our forecasts against output from the Australian Bureau of Meteorology’s General Circulation Model (POAMA 1.5). Our model gives a more accurate forecast for 16 of the 17 sites in Queensland that have the highest quality rainfall data. Our methodology is detailed in a new paper that will soon be published in the international journal Advances in Atmospheric Sciences.[1]
Artificial neural networks can find existing complex relationships and patterns repeated in rainfall data. The neural network software that we run on our standard laptop computer, can perform millions of calculations very quickly, and in this way find natural patterns that repeat themselves.
In contrast, the Bureau uses general circulation models; the same models used to predict global warming. These models attempt to forecast seasonal and monthly rainfall by working from a particular theory of climate. Our system is radically different in that we let the model find the patterns. We let the model find the natural cycles and then forecast forward up to 9 months.
That our system works, proves that there are patterns – natural recurrent cycles – in historical rainfall data.
Contrary to some of the claims of some climate sceptics: the natural hydrological cycle is not totally chaotic.
Contrary to the claims of most warmists: these natural rainfall cycles have not been perturbed by the current elevated levels of carbon dioxide.
What we did is tune one of the most advanced off-the-shelf software packages for pattern recognition (Synapse, Peltarion) into the natural patterns and recurrent cycles in historic rainfall data.
We found a combination of historic rainfall and temperature data combined with historic data for three climate indices (Southern Oscillation Index, Pacific Decadal Oscillation and Nino 3.4) gave the most accurate forecasts. Inclusion of solar irradiance and sunspot number did not enhance the performance of our model. Inputting carbon dioxide concentrations doesn’t improve our model.
We believe we have barely scratched the surface in terms of potential for rainfall forecasting using this method for eastern Australia and are looking for significant additional funding to progress this work.
The work has so far been wholly funded by the B. Macfie Family Foundation and supported by the Centre for Plant and Water Science at Central Queensland University.
*****
[1] Email me at jennifermarohasy at jennifermarohasy.com if you would like an advance copy of our paper:
John Abbot and Jennifer Marohasy
Application of Artificial Neural Networks to Rainfall Forecasting in Queensland, Australia
Advances in Atmospheric Sciences, Volume 29, Number 4, Pages 717-730.
FOLLOWING IS SOME RECENT MEDIA:
Notebook neural network beats supercomputer model for predicting rain
By Stephen Withers. May 28, 2012
Researchers at Rockhampton-based CQUniversity have come up with what appears to be a better model for predicting rainfall in Queensland.
The Australian Bureau of Meteorology makes its forecasts from a mathematical model of the atmosphere (known as POAMA) that run on supercomputers.
But a pair of researchers at CQUniversity have applied neural network technology to the task and claim this gives better rainfall forecasts – and runs on an ordinary notebook computer.
The basic idea of a neural network is a collection of connected nodes that are modelled after the neurons in a brain. Like the biological structure on which it is modelled, a neural network is capable of ‘learning’…
*****
Software offers brain on the rain
Angus Thompson. May 30, 2012
ARTIFICIAL intelligence software has outperformed the Bureau of Meteorology in predicting long-term rainfall – and could soon be used for daily forecasts.
The neural networking technology proved more accurate in forecasting weather events in 16 out of 17 Queensland locations, including rainfall that led to the Wivenhoe Dam bursting, researchers from Central Queensland University found…
http://www.heraldsun.com.au/ipad/software-offers-brain-on-the-rain/story-fn6bfkm6-1226372995011
*******
Weather Forecast System Developed
By Austin King, May 30, 2012-05-30
Dr Abbot combined his interest in artificial intelligence for pattern recognition with Dr Marohasy’s knowledge of climate science to come up with a neural network system that predicts monthly rainfall from three to nine months in advance.
Their system is the focus of a study about to be published in the international journal Advances in Atmospheric Sciences.
They decided to put their heads together just after the January 2011 flood that isolated Rockhampton from the rest of Central Queensland.
The method is based on pattern analysis accepting that there are patterns like short- and long-term cycles evident in rainfall data…
http://www.themorningbulletin.com.au/story/2012/05/30/weather-forecasting-system-developed/
******
Original Media Release
Laptop beats supercomputers at forecasting rainfall – CQU
28 May 2012
Laptop beats supercomputers at forecasting rainfall, thanks to artificial intelligence approach
CQUniversity researchers – including one who honed his skills through share market trading – have been able to forecast seasonal rainfall more accurately than the Met Bureau by using artificial intelligence for pattern analysis.
A journal, Advances in Atmospheric Sciences, is about to publish a new paper by CQUniversity researchers Dr John Abbot and Dr Jennifer Marohasy. The paper shows the application of artificial intelligence to monthly and seasonal rainfall forecasting in Queensland, Australia, and compares their forecasts with forecasts by the Australian Bureau of Meteorology’s super computers.
In March 2009 the Australian government ordered two new supercomputers ostensibly to ensure Australia
is at the forefront of international weather forecasting and climate modelling.
The super computers run a general circulation model known as POAMA that is a mathematical representation of general atmospheric circulation patterns. In the West, attempts to improve rainfall forecasts from general circulation models have focused on improving our basic understanding of weather processes, most recently with a focus on the role of carbon dioxide as a greenhouse gas.
In other parts of the world (notably China, India, Iran), governments have also funded research into artificial neural networks for rainfall forecasting. This radically different method is based on pattern analysis accepting that there are patterns, for example short and longer-term cycles, evident in rainfall data. Neural networks, based on artificial intelligence, have the ability to consider large numbers of climate indices (eg. El Nino, Indian Ocean Dipole) and other inputs (eg. temperature, cosmic ray flux) simultaneously and make predictions independently of any understanding of, for example, the hydrological cycle.
Following the devastating floods of January 2011, with three-quarters of Queensland declared a disaster zone, CQ-based researchers Dr Abbot and Dr Marohasy combined their respective interests in artificial intelligence for pattern recognition with climate science to see if they could forecast the weather at least as well as the Australian Bureau of Meteorology.
In their first attempt at optimising their dynamic stand-alone time-delay recurrent neural network (TDRNN) they found that patterns within rainfall data alone could provide a forecast.
Their optimal model combined current and lagged rainfall, temperatures, Southern Oscillation Index, Pacific Decadal Oscillation and Nino 3.4 reflected in the highest Pearson correlation coefficient and lowest root mean squared error value (RMSE). This model was applied to 20 sites across Queensland generating monthly rainfall predictions three months in advance.
The Australian Bureau of Meteorology provided output data from POAMA enabling a direct comparison of the ability of the two models to forecast seasonal rainfall: the general circulation model (POAMA) developed from a theory of climate and run on a super computer with a large staff versus the neural network prototype based on pattern recognition theory and run off a laptop in a small office.
Outputs, as monthly rainfall forecasts three months in advance for the period 1993 to 2009, were compared with observed rainfall data using time-series plots, RMSE and Pearson correlation coefficients. The comparison showed the prototype neural network achieved a lower RMSE for 16 of the 17 sites compared, meaning it gave a better forecast for 16 of the 17 sites.
Dr Abbot, who honed his skills with neural networks through share market trading, considers the prototype design for rainfall forecasting very preliminary, with potential for significant improvement and application for anywhere in eastern Australia with at least 100 years of high quality historical rainfall data.
The findings have significant application to everyone affected by the weather but particularly for agriculture and mining with scheduling of mine activities in CQ significantly impacted by wet days. ENDS
Details and PICs via http://uninews.cqu.edu.au/UniNews/viewStory.do?story=9483
Jennifer Marohasy BSc PhD is a critical thinker with expertise in the scientific method.
If you live in Queensland and tune in to WIN TV News tonight at 6.30pm I should be on talking about artificial intelligence and its application for rainfall forecasting.
http://www.wintv.com.au/central_queensland/news
Jen, you are obviously not making enough assumptions [as in the GI in order to get the GO]
Artificial intelligence simply supplements native intelligence. Neville Nicholls has demonstrated how extreme rainfall in eastern Australia has been driven by strong SOI during the last hundred years.
http://rogerpielkejr.blogspot.com/2011/04/neville-nicholls-on-australias-extreme.html
Continuing the theme that artificial intelligence simply aids native intelligence….. On this site 3 years ago Stewart Franks related Australian weather to the ebb and flow of El Nino – La Nina cycles, and correctly predicted increased rains with more La Ninas. He was correct; the global warming activists who predicted droughts have been wrong.
http://jennifermarohasy.com/2009/02/better-planning-for-extreme-floods-possible/
Hi,
Do you know what type of methodology is used by the Norwegian yr-no people ? I suspect that it is something along this line, with additional astral input. They have a strong following in the bush,as like you, they differ successfully from Met.
You really are a busy young lady , Jennifer. You must make it very difficult for Media Watch to keep track of who you are corrupting.
Don B
Thanks for your comment. We use neural network technology – one area of AI. Artificial neural networks are massive, parallel distributed, information-processing systems with characteristics resembling the biological neural networks of the human brain.
We reference the important work of Stewart Franks and also Neville Nicholls in our new paper.
Ian Thomson
Thanks for your comment. The software that we use is out of Sweden (Synapse, Peltarion). Our paper reviews the work to date on the use of neural networks for rainfall forecasting. So far the research effort has been almost exclusively by scientists in Asia – including China, India, Malaysia, Iran.
We appear to be the first to input Australian climate indices (e.g. SOI, Nina 3.4) into a neural network system to forecast Queensland rainfall. We reference the limited work done for the Parramatta catchment done some time ago.
We use a dynamic stand-along recurrent, time-delay neural network. Our prototype spits out actual numbers for total monthly rainfall as a forecast.
PS There has also been some comment here:
http://www.itwire.com/science-news/climate/55007-notebook-neural-network-beats-supercomputer-model-for-predicting-rain
Well done Jennifer.
What influence does the IOD have in causing rainfall/drought? Both in 1974 and recently a negative IOD and a La Nina converged to cause extreme rainfall (but seemingly lower rainfall in SW Aust).
Hi Ian George
The IOD, Indian Ocean Dipole, is measured through the Dipole Mode Index (DMI) as a difference in sea surface temperatures between the tropical western Indian Ocean and tropical southeastern Indian Ocean. We used the DMI based on HadISST1 (1870 to present).
When we ran our prototype neural network with the DMI it did not significantly improve output/the forecast. It could be that for those runs the potential influence of this climate indices had already been incorporated through other input variables/classes of data. We used four classes of data in a limited number of combinations: monthly rainfall, climate indices, atmospheric temperature and solar (sunspot number and total solar irradiance).
There is much work still to be done.
It wouldn’t require much of any intelligence to be ahead of supercomputer models.
Just a little study of history would do:
http://www.dailymail.co.uk/sciencetech/article-2152004/Lost-photos-prove-Greenlands-ice-melting-FASTER-80-years-ago-today.html
Well congratulations for having a go and please continue in the holy grail of seasonal rainfall prediction. We salute Inigo Jones. And good to see sceptics also prepared to torture meteorological data.
But alas an interesting paper spoilt by press releases claiming far far too much and taking a shot at BoM.
We need a Version 2 of this paper. Version 1 is not convincing.
Why – well POAMA 1.5 is two generations back. POAMA 2.4 introduced October last year has twice the skill !!
There are also some fundamental flaws in the paper – including the seasonal cycle in the predictor and predictand says more about neural networks reproducing the seasonal cycle than seasonal variability. And given the variability is the problem that’s what everyone wants to know. Given lagged rainfall as a predictor the neural network should do this well.
The forecasts themselves are not from POAMA – they are the authors’ construction. They are a construction of anomalies with linear interpolations across grid boxes. This is a rough as guts way to downscale given the high orographic and distance from the coast related rainfall gradients. Could introduce mega errors.
Additionally how important is RMSE when it’s the interaction of the forecast with decision making that’s the issue. More on that later.
Also you’ve only covered one PDO cycle. (would be good to see what the neural network predicted for 2010/11 and 2011/12)
CO2 is not correlated you but you have temperature confounded in the analysis. So unconvincing. What would the mean, maximum, minimum temperature series for Queensland add? Why use Sydney? And what would a warming Coral Sea SST grid box or northern Australian SST grid box time series add? So the CO2 comment is trifling really.
Additionally POAMA spans weather through intra-seasonal (MJO) – seasonal (ENSO, IOD, SAM) through to climate change scales over the whole of Australia. Not just seasonal forecasting for Queensland. So it’s very specific to claim victory on a small sample of overall capability.
Although this is a preliminary paper your 3 months lead is a small snapshot – POAMA regularly does 1 month incrementally to 9 months I believe.
How does the neural network predictive vary temporally – as you know autumn is a “predictability barrier” for most systems – where La Nina or El Nino events can reform, go neutral or even change to the opposite phase. Any skill through the predictability barrier would be well regarded especially for the animal industries entering winter and wondering about next summer rainfall.
But POAMA is not the Queensland benchmark either – the classic Roger Stone SOI phase system would be the 3 month system to beat and what many use.
Additionally the experimental (when will it ever be published) SPOTA system has two validation periods – one in recent times and one with early records. SPOTA is a long lead system for the animal industries. Readers would note that you assessed the neural network as not as good as SPOTA on an initial comparison. But SPOTA is long lead so a different beast.
Given warmists KNOW that their indice base lines are changing – SPOTA is built on gradients across SST regions that mimic ENSO processes and IPO processes. This is an “attempt” to “climate change” proof the indices.
Have you checked the trend and variance stability of your predictor baseline components. E.g. e.g. http://journals.ametsoc.org/doi/abs/10.1175/2011JCLI4101.1?journalCode=clim Janice Lough’s classic study of rainfall in reef catchments shows increasing variance over the 20th century. Wetter wets and drier drys.
Bazza of course would be concerned about statistical assumptions if your training period and baseline period variances were different.
But to forecast evaluation – RMSE – grizzabits and beezebulb ….
As a farmer I wouldn’t pay you on RMSE. You can’t bank RMSE. And I know winters are cold and summers warm. And summer rainfall tends to fall in summer. A duh. But tell me what I don’t know – the variation !
What would be more impressive if you can tell us for different lead times, for different starting periods what your ability to forecast the right rainfall tercile is. (Or if Koala Bear has a really hairy chest – quintiles).
This is what users need for real world decisions:
e.g. do I sell or buy animals – on a wet or dry forecast
do I drought feed knowing it will break (I could ask Stewart Franks I guess LOL !)
What if I burn pasture on a wet season prediction and it goes El Nino (how often does your system get it wrong is something I need)
Do I up the crop nitrogen rate on the forecast of a big wet or cut it back (Or not even bother planting) on an dry forecast.
Do I stockpile coal at great expense on the prediction of a big wet
And if you close in at intra-seasonal – will I have a crop planting opportunity (soil moisture from surface to depth) in the following x weeks
If it’s going to be an El Nino year do I plant low yielding but frost resistant wheat.
So to make it useful for real world users you need to be able to put the forecast into a decision analytic. Terciles ! Sod the RMSE ! And I need to know how many times you’re right and wrong and when. Downside risk.
Nitpick: Your thread declaration of “Contrary to the claims of most warmists: these natural rainfall cycles have not been perturbed by the current elevated levels of carbon dioxide. “ are not scientific given the significant number of papers on Australian climate change that illustrate the contrary in SW WA and SEA, 2011 La Nina and SST strength, SST trends around Australia, ENSO changes to Modoki – and related changes in southern hemisphere circulation. If the thread is reviewing a scientific paper it would be appropriate to say why these previous papers are incorrect. But lets not go there or we’ll be here for weeks.
Other predictors especially if you’re going national – sub-tropical ridge (STR) latitude and STR intensity, the 6 major SST grid boxes around the nation, SAM, MJO.
And phew – have you thought about doing streamflow?
Anyway don’t mind me – just thought you’d like some serious feedback. Not convinced on the NNs (yet).
Little Lukey,
you got a laugh out of me.
“CO2 is not correlated you but you have temperature confounded in the analysis. So unconvincing. What would the mean, maximum, minimum temperature series for Queensland add? Why use Sydney? And what would a warming Coral Sea SST grid box or northern Australian SST grid box time series add? So the CO2 comment is trifling really.”
Excuse me, but, would you like to actually compute for us the difference in temperatue that 10ppm of CO2 alledgedly drives by YOUR estimation?!?!?! Then try and figure out how much this tiny number could possibly have to do with changes in rainfall on a yearly basis!!
HAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHA
The IPCC as a max gives about 6c over a period of 100 years. With natural variations that means so little to the rainfall over a 10 year period any inclusion of CO2 would be for sh!ts and giggles!!!
For short term forecasting CO2 is meaningless. Haven’t the modelers and the IPCC taught you nothing boy?? Their big deal is centenial!!
Seriously, the long term projection by the IPCC is about .03c per year. This is meaningless to even decadal forecasting man!!!
Not relevant Kookers wasn’t talking about the prediction – thanks for playing. Yes the big deal is centennial.
Luke,
What do you mean you weren’t talking about prediction?
Here:
What would be more impressive if you can tell us for different lead times, for different starting periods what your ability to forecast the right rainfall tercile is.
e.g. do I sell or buy animals – on a wet or dry forecast
do I drought feed knowing it will break (I could ask Stewart Franks I guess LOL !)
What if I burn pasture on a wet season prediction and it goes El Nino (how often does your system get it wrong is something I need)
Do I up the crop nitrogen rate on the forecast of a big wet or cut it back (Or not even bother planting) on an dry forecast.
Do I stockpile coal at great expense on the prediction of a big wet
And if you close in at intra-seasonal – will I have a crop planting opportunity (soil moisture from surface to depth) in the following x weeks
If it’s going to be an El Nino year do I plant low yielding but frost resistant wheat.
So to make it useful for real world users you need to be able to put the forecast into a decision analytic.
And here:
Other predictors especially if you’re going national – sub-tropical ridge (STR) latitude and STR intensity, the 6 major SST grid boxes around the nation, SAM, MJO.
And why is the big deal centennial?
Do you mean centennial prediction or historical data?
Abbot certainly seems to mean historical data:
Here:
Dr Abbot, who honed his skills with neural networks through share market trading, considers the prototype design for rainfall forecasting very preliminary, with potential for significant improvement and application for anywhere in eastern Australia with at least 100 years of high quality historical rainfall data.
And your ‘not scientific’ claim is assuming that projective modelling is ‘science’.
I disagree….projective modelling is a tool used by scientists….. it is also used by other professions…..but it is still projective modelling….not science or economics or marketting or numerous other places it is used.
Debbie for heavens sake – I’m agreeing with Kookers and he has misunderstood what I’m saying. – I meant test for applicability in formulation of the analysis with the 85% of the 110 year rainfall data set (or whatever it was). So you test various factors – Nino regions, solar, CO2 – whatever makes you excited. Some of these factors won’t add any predictive skill (perhaps) – so you exclude those from your final predictive model.
So you make your predictive model by tuning/training/regressing/genetic algorithm optimising/Gauss-Marquardt-Levenberg method optimisng it on the training data set (here 85% of the rainfall series) and then test it on a sub-set not used in the tuning process – so called independent validation data.
So I was talking about the tuning process.
And it was a VERY small point anyway. And this is so tedious.
Try reading the paper Debs. It’s a start. And kookers is a ninny – he has no clue on this stuff.
Hi Luke
Lots of words, but neural networks are obviously not your area of expertise.
We are keen to benchmark against POAMA 2.4. We were given output for POAMA 1.5.
We have no problems forecasting out 9 months, again keen to benchmark for this.
We consider SPOTA to be the best equivalent for Queensland and as you have our paper you know we also benchmark against SPOTA (run by the Queensland government). We believe SPOTA is using better Climate Indices than the Bureau of Meteorology and better than what we have used and we would be keen to experiment with the SPOTA indices.
As detailed in our paper, we use 85 per cent of the data to train and then make a forecast for the remaining 15 per cent. Not sure what you are quibbling about here? This division of available data is standard in the practice of neural networks.
After some discussions the guys at the Bureau agreed that RMSE was the fairest and best method for comparing errors between the two very different systems/outputs. What is the realistic and practical alternative?
The guys at the bureau also agreed with the methodology employed for comparison of outputs for each site… recognising they forecast for a grid area and our prototype for a specific site. Indeed it could be argued that the method used gives them a slight advantage.
The Bureau was provided with our conclusions, output and more before we wrote up the results and once the paper was finished and we acknowledge their assistant in the paper.
Jen – really that is just a rubbish response – who cares what the forecast technique is – it can be a black box. I think you don’t know very much about seasonal forecasting. Fancy predicting the seasonal pattern ! wow
I was hoping for something more considered. What a waste of my time in review.
2.4 has double the skill of 1.5 so the press releases are pretentious.
For what it’s worth this is the latest on POAMA 2.4 http://www.wmo.int/pages/prog/www/DPFS/Meetings/ET-ELRF_Geneva2012/Presentations/GPC-Melbourne-March2012.ppt
Nola – your stupid drivel doesn’t even dignify a response. I wrote most of it last night actually. So onya bike sport if you have nothing of substance to add.
Hi Luke,
It’s a weather forecast. Is it wrong or right?
Is it better than the ones produced by taxpayer funded organisations ? Yes or no.
Is it close to what humans know already ? Yes or no.
Is it a little bit in conflict with computer climate models ? Yes or no.
IT IS A WEATHER FORECAST AND IT WORKS.
Refuted ???
Luke,
You are a very rude and arrogant person.
If you think that you have something valid to contribute to rainfall forecasting, then why do you not write a paper and submit it to a recognised climate science journal? This is what Jennifer and John Abbot have done.
Let us all know when you have written your scientific paper and have it accepted in a peer-reviewed journal.
If you were able to accomplish somthing like this, even once in your life, then you might feel better and not be so offensive.
Nola
It’s not really a weather forecast – it’s a seasonal forecast for a time period of months. Weather is the next few days.
It’s skill against existing systems is unclear – read what I have written in my long response. ONnSPOTA the authors on limited testing say neural net is not better. On SOI Phase – untested. On the latest version of POAMA – untested. Whether you would peg it against POAMA even is a moot point. Indeed the Bureau use statistical systems for their outlook (not tested against)….
My cited Powerpoint at 6:11pm says “Current seasonal outlooks for Australia based on a statistical model
Probability of rainfall/temperature in Tercile/Above Below Median categories
Trial of dynamical model forecasts.”
POAMA is in trial and used for much more than rainfall forecasts
Terciles are what we need to have for many wet /dry outdoor and farming decisions at intra-seasonal and seasonal scale.
You need to know how often you’d be in the right slice of the tercile pie. i.e. wettest, middle, driest. Quintiles maybe for extremes but stats are harder if you go finer.
I said specifically “refuted significant aspects” – I have congratulated the authors for their work. So are we unable to discuss it?
If they had blown POAMA and other systems into the weeds great stuff. And I have tried to be helpful suggesting how the analysis could be tuned for users. I could have instead of generating “a lot of words” to bore Jen, I could have simply put the boot in on the weak points and gone for broke.
So in summary Ian – it’s an interesting bit of forecasting work – but quite there yet (or unproven). The authors could do well to harvest and discuss the comments, and keep going.
All IMO of course.
Nola – true. But we’re discussing this most interesting paper. Do you have a contribution.
Luke
I am sure we can look forward to reading your next contribution to the journal Advances in Atmospheric Sciences.
Nola
Know-it-all-Luke writes lots of meaningless words. Luke will say anything except admit he is wrong.
Publish your silly ideas in a peer reviewed journal Luke.
Gosh you have written a lot on this topic in a short period – somebody has hit a nerve!!!
Do you have a real job? Or do you just haunt Jennifer for some sin committed in a past life?
Poor Nola and sp sauce – nothing to say except abuse. Nothing to contribute. It’s sad guys – what do you gain – except for being cheer squad zombies.
sp – know-it-all meaningless words tell me you’re a drongo with no interest in the topic – so why are you here – to say “rah rah rah”. Outstanding.
And it’s fascinating all this SUDDEN talk about peer reviewed journals when for years it’s been argued that they’re corrupt and can’t be trusted. Now you’re recommending this are you?
And think of all those papers in so many threads that have been published in journals that have been attacked here by bloggians. So that was fair game?
So why don’t you both crawl back into trailers, read the paper (which you haven’t done) and try to utter something bordering on useful.
Don’t worry sp, Jennifer is not haunted by Luke, He’s the resident symbiotic troll . A sort of brainless leach but necessary all the same.
Thanks Mack – Another missed opportunity to make a science comment.
This has actually been an interesting exercise – virtually no engagement on any science discussion by the punters. A sort of blissful sweet night mist has descended. So here we have a key paper and all the punters can say is good on ya Jen and how dare anyone make a critical comment.
I bet if this paper was written by someone from BoM or CSIRO you’d all be at it hammer and tong.
Pathetic guys.
Luke – Jennifer’s paper challenges the the orthodoxy, it presents an alternative view to the myopic mind-set of alarmists like you.
And thats what you dont like. Because you are comfortable with your world view. And there cant be an alternative worldview because you are always right.
Following pople like you means repeating the same old mantra over and over again.
Following the work of Jennifer means exploring new ground, new ideas – a concept alien to your atrophied mind.
Jennifer has challenged your worldview and presented an alternative that does not require CO2. She has also presented a method that relies on identifying the trends inherent in the data, and not identify a “trend” that conforms to an (incorrect) theory.
Jennifer has gone outside the box and applied some new and innovative thinking – that is science.
What you propound is orthodoxy – and that is not science. It is clear you have a political agenda.
Your world view bubble has burst in the face of facts. Your only response to this is bluster and insult. This is obvious from the quantity of vitriol you have produced in a short time.
And the amount you have written in a short time is testament to the impact Jens work has had you.
She has shown a way out of the morass of misinformation presented by your kind.
She has preesented a paper that will (I think) contribute to you becoming becoming irrelevant.
And the more you write on this topic the more you demonstrate your obsession with suppoorting the political orthodoxy.
Present your own paper for peer revew publication. Put up or shut up.
sp – you’re just an ignorant boofhead. How does a technique previously used before in a known field with well known indicators challenge the orthodoxy. You’re just writing drivel my friend. Just drivel.
I have made specific comments which are obviously way above your head. You have made ZERO science comments coz you can’t.
Mate this is a blog about “evidence based” environmental science as Jen has often espoused. It has a comment section. Ideas are contestable. It’s not an echo chamber for backslappers IMO. A paper has been tabled for discussion. There is none apart from me! A pity as the field and application is most interesting. And relevant to agriculture and open cut mining.
Debbie won’t even speculate how she’d use a forecast. You can’t get a whimper out of the usual cheer squad. Frankly because they’re clueless to what she’s done. You have no idea do you?
So do you have a science contribution to make or are you just a stupid spectator?
But Luke, you have already REFUTED Jens paper (in record quick time).
The science is settled according to St Luke the Blovator. What is there left to discuss?
The current “models” consistently fail when compared to real world measurements. The relationship between CO2 and temperature is weak to descriibe real world events.
Why not explore alternative methods to examine the real world relationships? Perhaps there are other ways?
But we cant explore those alternatives or examine real world events in different ways because you have already REFUTED the paper.
Have you not?
Luke?
Debbie won’t even speculate how she’d use a forecast?
I speculate all the time Luke.
Weather forecasts are more important in my industry than most.
If Jen and Abbot have opened the doors to something more useful then I’m happy for it to be added to what we already know.
I hope it does prove to be more useful.
Why are you being so defensive and critical?
I can’t see anything inherently suspicious or counter productive here. They’re not pretending that they have invented something entirely new. They are just looking at it in a different manner and also taking advantage of the ever growing source of reliable historical data that is available paricularly on the eastern seaboard.
0.01% attempt at a contribution. It’s an improvement.
I never said the science was settled. That’s why it’s interesting.
Alas you’re way off the mark – as you’ve confused AGW stuff with seasonal forecasting. Not even warm sp. It’s a diversionary issue here.
If you noticed sp – I have suggested that the authors could (1) compare to other well known used every day systems (2) improve how farm users might receive the information (3) suggested a few other indices to try too – especially if they’re going for all eastern Australia
Have you thought how you would use the system?
You should read what I wrote – you don’t see these suggestions as a positive contribution?
Easy Luke use it for stream flow prediction for water yield and fire season potential – currently the statisical forecastes and POAMA 2.4 aren’t all that great at providing data for either.
No Luke,
I haven’t confused it at all.
I’m commenting on potential use/usefulness which is the way Jen and Abbot seem to be looking at it too.
That ‘usefulness’ judgement does tend to draw a comparison with the exponentially increasing number of contradictory AGW modelling…..from my humble landholding perspective….especially when it’s clear that huge amounts of political attention and hence climate research funding has gone in that direction.
How would I use the system?
I suspect in a similar manner to the way I use the ones available now.
I tend to usually go to to BoM and YR and compare those 2….sometimes go further afield if those 2 are completely contradictory….which happens more often than we would expect.
Both of them help but I don’t find either particularly reliable until about 36 hours beforehand.
They don’t pretend they are either….they speculate on known factors as well…and give probability figures.
I also commit what you and Bazza would likely consider a grave sin and factor in what I have learnt from the ‘old timers’ (including indigenous) who seem to know how to read signs in the behaviour of the natural environment (plant and animal)
Those are very often (but not always) a better earlier indication of impending changes but not necessarily any better indication of amounts of rain or duration of rain events.
If Jen and Abbot’s work help to make the speculative stuff capable of ‘less error’ then it has no argument from me.
And your last question.
I saw most of what you wrote as ‘defensive’ rather than ‘postive’….but maybe that’s my fault.
However….the suggested tests against indices and other well known systems was not defensive but was rather obvious advice (but nonetheless good advice).
So you have not REFUTED the paper? Merely suggested improvements?
Do you agree with the central tenet that it is worth investigating alternative methods without an obessive focus on CO2?
The scientific method of review is to provide the paper to expert reviewers who, over a period of time, usually months, will consider the paper, seek supplementary information and clarification, discuss contentious issues, and seek changes, clear it for publication, and further scientific discourse.
The political method of review is for automatic gainsay by taking a contrary position, and attempting to declare the discussion closed, and that any further discussion after this arbitary point in time is useless. The poposition is thus REFUTED and any further discussion diluted by reference to the spurious “science” approved by St Luke.
Its not hard to see which method you apply.
I mean it took you what? A few hours to REFUTE the paper. Amazing!
The bottom line is it is worth investing alternative methods outside of the CO2 or any other current “approved” paradigm.
You use “science” as a smokescreen to hide you are pushing a political wheelbarrow.
Jen has challenged your orthodoxy from first principles, and you dont like that. You just dont like it because it may get other to people to think outside the box you want them to think within.
St Luke of the Bulging Intellect, who forsake common sense and manners, for the love of his cause.
Well done Jen; I haven’t time right now to read your paper in detail, or luke’s detailed bloviations, but will do so soon.
BOM has no standing at all for making predictions, about anything, over any time scale:
http://www.warwickhughes.com/blog/?p=931
So, perhaps it is not doing your paper justice to use BOM as the benchmark; POAMA or ACORN, they are all duds.
Dear dear Cohenite could do well to read comments 8 and 19 from Saint Warwick’s most excellent informative site.
Thanks for the link Cohenite,
And Luke…..comment 19 at least demonstrated some much needed humility about the veracity and reliability of this work.
Seriously….us humble landholders are more than just a tad over the dictating and would prefer some serious input into where climate research funding goes and for what useful purpose.
I particularly liked your comment Cohenite about CBA.
But Debbie why would want any input – you won’t even speculate how you’d like to use the information or what would be useful. I tried to engage you on terciles after I said how would you use Jen’s RMSEs or indeed her forecast (and wasn’t a shot at Jen – just a question). . You’ve spurned attempts by Bazza and myself to engage with you. “Not interested in packaging” as you phrased it. Which is strangely typical – lots of opinions – then try to engage – and find the audience wasn’t really interested after all.
Luke-I am a stockowner dependent on the weather and the BOM with all the radar gear is improving marginally at predicting up to a week ahead. You say-
I bet if this paper was written by someone from BoM or CSIRO you’d all be at it hammer and tong. …………
These people pick up their pay and never have to be right and most of the time they aren’t! The CO2 scam has devalued science so badly that I will never believe another thing to come out of the csiro.
Roger Stone has had a blank cheque and a wonderful career but I would sooner toss a coin and read my stars than put money out on his predictions. An improvement @16 of 17 spots for beginners on a tiny budget must be spooking the likes of you Luke -I assume that you have your nose in the taxpayers trough. hence all the time trolling on here.