ONE way to help shift an accepted scientific paradigm (e.g. Anthropogenic Global Warming) is by coming up with a more practical and relevant way of describing and predicting real physical processes in the natural world (e.g. a better seasonal rainfall forecast).
Footer
About Me
Jennifer Marohasy BSc PhD is a critical thinker with expertise in the scientific method.
Read morePayPal
The quality of the data is key to the ability to recognize patterns.
Data which have been “homogenised” have a distorted signal and will injure the predictive quality of analysis based on artificial intelligence.
I’m impressed. Why isn’t the government funding this type of research instead of putting so much into global warming capers.
With the money Mr Cambell has saved in abolishing the Queensland Department of Climate Change, surely there will now be some support.
We all know how badly Anna Bligh and her team stuffed up with the rainfall forecasts and the management of Wivenhoe Dam in 2010-11.
Time to try something that works better and get some proper scientists on board rather than all the relatives.
“Anna Bligh and her team stuffed up with the rainfall forecasts” – tripe ! Every man and his dog was on a about a record wet for months.
This sounds very innovative. I wish you great success in this effort.
Just logging here that I have refuted significant aspects of the paper on the previous thread (given how top threads absorb the attention)
Dera Luke
I think I have hit a raw nerve with the truth, Luke!
Could it be that you are on the A-team, I wonder?
If they had not so completely stuffed up I would suggest that Brisbane suburbs near the river would have never been flooded at all
It was a total and complete farce – the weather forecasts were inaccurate and the management of the water release was totally incompetent.
I have signed up for the class action, as my property was flooded. Meanwhile Anna and her hubby are cozy at home counting their superannuation.
It was so good to see some of them fushed away by the recent election.
More rubbish to clean up.
LL
Luke,
If you really think you have something useful to say about this paper from a peer-reviewed journal, then I suggest you submit you own paper to the same journal.
Then, properly qualified scientists can evaluate your ideas and, if they make sense, your ideas can be published.
Otherwise, it all rather comes over as ranting from someone who is a little envious of those that can.
LL
Lester perhaps it’s you that have the raw nerve being flooded. In that you have my personal sympathies and I have assisted in flood cleanups so well know the mess.
The season leading up to the event was clearly a major La Nina. But that’s as far as you can go really IMO. Whether you would get a specific size flood in a specific system does not come from these forecasts. But the buzz on the street and newspapers were definitely for a flood year. It’s not like nobody had thought of it.
If you think the Bureau can forecast x in 1000 type events with great lead times above the catchment on a specific day well dream on.
The flood report was inconclusive – who knows what would have been plus or minus 0.5 m if everyone had twenty hindsight. Small group of people – lots of stress – extreme event. Good as you can do IMO.
So why live on flood plains – did anyone learn from 1974 – of course not. And whatever is done Brisbane will flood again so learn the lessons.
Blame Joh for the flood manual not Anna but Labor were over-reactive to drop the vast amount out of the dam for the next season. What a waste.
And good luck getting any engineers to look after dams in the future. You couldn’t pay them enough to put up with the crap. Engineers scapegoated as nobody wants to accept that shit happens.
Amazing how the AGW cultists can never be wrong.
The bad forecasts regarding the deluge ending the last drought are now on that list.
AGW true believers are expert practitioners of the science of epoxology: the science of sticking (permanently) to an untenable position.
An amenable view of truth, history and ethics seems to be an important part of epoxology.
At least Campbell Newman predicted that the flood was coming but the Q govt did SFA about it except look at their Q100 maps and tell themselves 1974 can’t happen again cos we got Wivenhoe now. But we got another ’74 even without a cyclone or a king tide.
The weekend when the floods arrived there was not only no one on watch, there was no one on board.
And 36 years after the ’76 step change in climate these great GCMs still can’t see patterns and cycles, only ACO2.
We can only hope, Hunter, that they eventually come down with a serious epoxy allergy.
“And 36 years after the ’76 step change in climate these great GCMs still can’t see patterns and cycles, only ACO2” – err nope – utter rot. POAMA which we are discussing does many things other than CO2 !
If you think IPO is a predictable cycle you’re deluded.
“The weekend when the floods arrived there was not only no one on watch,” – nope – all these dams have a set plan and engineers allocated.
I see Campbell in council and the Qld shire councils didn’t do anything either. Or the land developers. Isn’t it strange that there’s a group of people for decades that having been checking if properties they’re contemplating purchasing are in flood areas and those that don’t care. Group A and group B. Don’t buy flood land for homes – it’s that simple. Why do we have new homes at Emerald built on flood prone land?
So it is ironic that we’ve come back to Wivenhoe on a thread on seasonal forecasting. So you can run around and blame individual and governments but it’s really about our expectations of probability. So it’s a good case study in terms of ENSO, IPO, seasonal forecasts, weather forecasts, and people under pressure with little sleep in emergencies.
So what do we know:
(1) a big wet season was likely on La Nina indicators – despite BoM (on “an” outlook) fluffing it for a short period in early spring
(2) catchments very wet – antecedent conditions conducive to flooding
(3) very probability of exceedance rainfall events on a full dam (well full in terms of water supply not flood compartment)
(4) dam has dual function of water supply (had just been through a record drought) and the system precludes pre-emptively draining the storage down before the BIG hard to predict flood (eventually after the event the govt drained 25% – sheesh !)
The only way I can see that Brisbane would have been “saved” would have been to pre-emptively drain Wivenhoe down before the wet season.
Given all these forecast systems are probabilistic – not categorical – and are 6/10, 7/10 or maybe 8/10 at best – there’s a fair chance of it being wrong.
a scenario – So would we all be happy to drop the water level, the wet season is mild and no flood, autumn comes – system goes into El Nino and water restrictions are on in force.
Public would be up in arms – why did you drain that water !#&*#$
So hindsight is always a wonderful thing.
So does anyone understand this aspect. Probably not. You can get minority odds out of any forecast on first throw of the dice, middle or last. And you can really only say something sensible about a seasonal forecasts system after many years of trials. They’re not 100% and never will be (chaos!).
So to make a sensible judgement call her you need to know how often on a particular problem the forecast system “get it wrong” (which they all will at some time). And you need to accept that risk or not use it.
So do people think that we should use seasonal forecasts to pre-emptively prepare dams for wet seasons. As once you get into a major event the stress and panic is likely to not serve us well.
What do you think?
Luke, you dope, unpredictable natural cycles v ACO2 is what the AGW argument is all about.
At least Jen is sticking to what is predictable from past performance.
And when I and many others could see what was happening that weekend of the floods and tried to contact anyone on watch, we got nowhere.
I agree with you about the craziness of building on flood-prone land but those govt produced Q100 maps seemed to justify a lot of it. They were responsible for advising people that extremes that happened in 1974 wouldn’t happen again. Oh dear!
Luke,
we’re all wiser with hindsight.
Unfortunately, climate/weather/water management has been made into a political issue.
We have an insane obsession operating that believes we can come up with a magic set of rules based on long term averages and an ability to ‘predict’ the future. People stick to the ‘rules’ despite the fact that the rules don’t match the circumstances.
Wivenhoe was actually built with flexibility in mind. I’m not convinced it was capable of managing the extreme events that occured there but there was definitely an obsession with rules operating rather than a willingness to be flexible.
The sort of flexibilty you are sensibly advancing here is not supported by centralised bureaucracy and centralised policy/rules/regulations.
“The only way I can see that Brisbane would have been “saved” would have been to pre-emptively drain Wivenhoe down before the wet season.”
You’re right, we shouldn’t be rehashing this but you’re wrong about the solution.
In October 2010 after the dam filled and a La Nina was evident, the govt dumped a lot of water very quickly into a river bed that hadn’t had a decent flow for probably a decade, causing a lot of destruction and getting a lot of flack from the locals. If they’d released a continuous, lower flow from October when the flooding was already being predicted and even the usual wet season was not due until January it would have been just basic common sense, not rocket science.
Luke
“Given all these forecast systems are probabilistic – not categorical – and are 6/10, 7/10 or maybe 8/10 at best – there’s a fair chance of it being wrong.”
That’s what we tried to hammer into your head all this time, and yet you insist that the predictions of CAGW are correct!
If you can’t predict short term how can you expect sane people to believe your long term predictions?
Johnathon,
good question!
Some of the answers have been amazingly non sensical!
Trying to pretend that weather and climate are not related is the best one :-).
Somehow they’re only distant cousins several times removed with a bit of adulterous hanky panky when it involves AGW and CO2. . . . except of course when AGW celebs like Tim say differently 🙂
Then a few extra hot days in Summer in Sydney are very closely related.
But these floods were also the fault of the coal mining industry according to Bob Brown at the time.
Also Luke,
this study is demonstrating that it is capable of ‘less error’ than other methods and it appears that it’s the case because there are recognisable patterns in the good historical data which would be why it has more use in Eastern Australia (more access to reliable historical data).
It doesn’t seem to be as obsessed with specific start/stop ranges that are just calendar points as much of the AGW modelling is.
It also clearly demonstrates how it may be ‘useful’ for industries like Agriculture and Mining.
Looks like it’s worthwhile to pursue IMHO.
“If you can’t predict short term how can you expect sane people to believe your long term predictions?”
Well that’s not totally true for starters – weather forecasts do have some skill. As do seasonal forecasts.
BUT climate and weather – totally different problems and aims.
Weather is an initial conditions problem. Climate change is about boundary conditions. Chaos will make all weather predictions start to drift after a few weeks. Seasonal forecasts only “work” by exploiting aspects of climate that “lock-in” for a season.
climate change predictions do not try to predict the climate on June 24 2040
exploring a future climate is not about forecasting. Even within a climate change prediction there is a cloud of ensemble members – each a individual run done to evaluate chaos. Normally most people only see means (averages) of these ensemble members.
But the real world will only have one real world “ensemble” for real. So don’t expect a nice linear evolution of climate year after year as per the average often shown. Climate models make broad predictions about the future climate. They are evaluated on a vast range of ways they simulate aspects of the current climate. That’s climate not weather !
If you don’t understand the important difference between weather and climate prediction you’re not even at first base. All this is fertile ground for sceptics to spread nonsense. All I written above is just basic common knowledge.
If you’ve got weather, seasonal forecasts and climate simulations all muddled up in your head it’s gonna be a long conversation.
Debs if I told an RMSE error number what would you do with it – nothing !
Wouldn’t you like to know for a given forecast – a start time, lead period, forecast period (x months) – whether it will be wetter than median, median (sort of average), or drier than median (so called terciles).
So in May for Sept to October rainfall (for example) – will it be wetter or drier than median or in the middle?
Seasonal forecasters need to tell you that. AND how often if you picked the biggest probability slice of the pie – how often would you be wrong. i.e. the system says wettest tercile is high probability – you plant a crop – and it ends up being driest, crop dies, lose money, unhappy, dog bites you.
Isn’t that what you’d like to know. Not an RMSE value (which is science hacker stuff?)
SO in this land of droughts and flooding rains don’t we mug punters from Inigo Jones to now want to know whether it’s going to be wetter or drier in the months ahead? That basic !
In one line Debs ….
So gimme tercile forecasts, and gimme how often wrong/right on a validation test.
“Also Luke, this study is demonstrating that it is capable of ‘less error’ than other methods ”
– not really
against POAME 1.5 – better
against POAMA 2.4 with alleged 2x skill – untested
classic SOI phase system – untested
BoM statistical outlook system – untested
SPOTA experimental system – not as good
(Debs don’t see this as being overly critical nor a deterrent to keep exploring neural nets – it’s just that you lapse into imprecision )
Luke,
not any argument from me there.
Unfortunately however, the political agenda that has attached itself to (or more likely hijacked) this work does not demonstrate the sensible humility you have demonstrated here.
It’s settled remember?
Even worse, it does confuse climate with weather whenever whatever circumstances may be prevalent.
That remains my objection.
Luke
I understand perfectly the distinction you make between weather and climate.
I also understand the very close relationship of the two.
In a very benign climate you will never have extremes of anything.
If you deny that, than it’s you who is lacking the “basic” understanding.
Not to mention the fact that when convenient, the warmist brigade will announce with glee, how a once in a 50 year hot spell demonstrates climate change. Maybe you don’t, I have no idea, but many in the game certainly do.
It may just be a strategy but it works, young people and those with lack of observation skills fall for it.
Keep on topic Debs – we’re not talking AGW now. Let it go. Breathe…..
So tell us how as a landholder you’d like a seasonal forecast packaged. What would it say.
Luke
Did you write all this stuff last night too?
Someone is watching.
Nola
Anna Bligh’s husband, the former chief of her disbanded Queensland Office of Climate Change has resigned.
Its so good to hear that at least some of the government waste is stopping.
Pink batts, set top boxes, aboriginal programs, school halls, departments of climate change. Its time all the waste stopped
Luke,
I beg to differ.
Read the first sentence of this post.
I’m also delghted to note a shift in attitude on your part. Rather than dictating what we get you have now thought to ask what we might need.
Good for you.
As a landholder, we’re not that interested in packaging. We’d be far more interested in some shared goals re funding climate research.
Debbie,
Need the like button.
There will not be any funding for genuine climate research though, no revenue to be gained from the results. Can’t tax nebulae, only villains. And villains are easily constructed , with enough funding.
Crikey Debs “ONE way to help shift an accepted scientific paradigm (e.g. Anthropogenic Global Warming) is by coming up with a more practical and relevant way of describing and predicting real physical processes in the natural world (e.g. a better seasonal rainfall forecast)” WTF and huh? – and what do think has been going on for 25 years of seasonal forecasting research?! and what do we suppose climate models actually do? – sheesh ?
So you obviously don’t want to use any of this seasonal forecasting research “packaging!? is it” so what’s the point of shared goals.
Spanglers @ 9:43am
Oh yes I agree – but it’s all wonderful in hindsight isn’t it – not every La Nina has Brisbane in flood. Chances are higher.
The dam manual would have gone back to Joh dam construction days. No capacity to be flexible.
No serious floods in the intervening period. No action. No experience. No rehearsal. Years spent dealing with a mega-drought in the leadup.
So imagine if you dropped the dam in October as you now know with 20:20 marvellous hindsight – but in a different universe no flood occurs and next season goes El Nino.
Would you be prepared to take the rap for this? You’ve dropped billions of dollars of water into the ocean. Off to the CMC with ya ! Off with your head !
I think we should use seasonal forecasts but public needs to know there is a downside. And our response needs to be progressive and adaptive.
So given there are 3 dam engineers referred to the CMC – you’d like future engineers to be
“more flexible” and adaptive would you?
Hah !
pay me $500,000 p.a., state legislated protection, and a massive liability insurance for good measure and maybe I’ll think about it. LOL. No not even then.
Of course we could not build on flood prone land (wow !) and build levees around towns where possible (and someone will whinge about their scenic amenity).
Luke?
Some of what has been done in the last 25 years has been very useful.
When have I ever said otherwise?
Most of your criticism re Wivenhoe probably comes from a lack of shared goals don’t you think?
I find it a bit amusing that you’re attributing responsibilty to the Joh govt. A lot of water has gone under the bridge since then (so to speak).
‘So imagine if you dropped the dam in October as you now know with 20:20 marvellous hindsight – but in a different universe no flood occurs and next season goes El Nino.’
They were hanging on to every drop because they honestly believed in the AGW theory, that was their folly.
If they had looked back they would have had a better idea of knowing what was coming, flood follows drought on a regular basis.
Good luck with your paper, Jen.
El Gordo – you simply ducked my question. Arm chair bulldust El Gordo.
As for hanging onto every drop and AGW – how preposterous. They were operating under guidelines which don’t allow for preemptive releases weeks before and ended up on a major event which put them into a fatigue zone.
If you want preemptive releases on La Nina summers – need a rewrite of the rules.
Debs – simply that in drought period management tends not to look at floods (no excuse but human nature). I suspect any review wouldn’t have considered a rare high intensity event on top of a full storage (well full to water supply levels).
Jennifer,
Is it true that the technique you tested has been used and validated in other regions?
Hi Hunter
In Asia, particularly China and Iran, the use of artificial neural networks for short, medium, and long-term rainfall forecasting is an active area of research within the meteorology, mathematics and computing communities. We review this research in the beginning of our paper – I can send you a copy of our paper?
It is difficult to compare and contrast the different types of neural networks used, and their skill relative to the other systems used in these countries, because researchers generally only provide a comparison of their output with observed values for the particular region and time period.
No-one much is benchmarking. We hope that our paper gets people thinking about how to benchmark and the need to benchmark including against General Circulation Models.
Jennifer,
Yes please, I would love a copy- thank you very much.
From what you are saying about the models, it seems the implication is that this might be something actually useful- a way to understand regional weather better. One problem I see in the GCM world is that claiming climate is separate from weather is problematic. As is the case with carbon black finally, possibly, getting moved up the priority list, perhaps we can also see regional forecasting become more useful. Hopefully your work could lead closer to that worthy goal. You do a great interview, by the way.
Luke, with your ideas I wouldn’t let you near a dam in a drought let alone in a flood. Fancy having the hide to blame Joh! [as you said to Deb, it’s not about AGW, {or science} it’s about ideology].
They let too much go too quickly in October which caused the expected reaction after years of drought but all they needed to do was allow a reasonable, continuous, non-destructive amount through, knowing that the wet season was coming and floods were predicted.
You don’t need a manual to get that. You learn that playing mud pies as a kid.
Wivenhoe was within inches of self-destructing that weekend because no-one was there to see what was happening and when they finally woke up they had only two options left: destroy Brisbane or destroy Wivenhoe AND Brisbane.
Yep,
And in flood periods they tend not to look at droughts. Which is just a tad silly perhaps because that is one of the main reasons we build water storages. . . to even out the hugely variable rainfall/infow in our land ‘of of drought and flooding rains’.
That’s probably another good reason why this work of Jen’s and Aboot’s may help us all to look at the whole picture with different eyes.
From the perspective of someone who works in Agriculture, I certainly can’t see any harm in what they’re doing. I would suspect the mining industry would be the same.
I’m not sure why that makes BoM so defensive. They don’t believe they have all the answers do they? It also doesn’t require anyone to negate the excellent work BoM and other organisations have done.
A bit more ‘sharing’ of information and some ‘shared goals’ about good outcomes re projective climate modelling would be a better idea (IMHO).
And more drivel from Hunter – “One problem I see in the GCM world is that claiming climate is separate from weather is problematic. ”
the ignorance level is amazing and folks you had to read it here – POAMA in fact goes from weather to climate change. Same with UK Met. The drivel just flows and flows.
Spanglers – “Wivenhoe was within inches of self-destructing that weekend because no-one was there to see what was happening ” – just rot. Engineers were on duty as they would be.
And I’m sure you can predict a 1 in 5000 rainfall event just as the storage is filling.
And it would not self-destruct – what nonsense – the fuse plus would have blown – it was close to doing so – the dam structure would be intact but a huge pulse of water would have gone down the system if the fuse plugs had failed (as they are designed to).
I didn’t blame Joh – I suspect the manual was from that era. Yes Labor had 20 years to review it. Not sure a review would have changed anything.
The only way you can bullet proof this issue is to preemptively drain the dam early. And woe betide you if it ends up in not flooding after all.
As I said – nobody would want the job running that dam.
Just love those arm chair 20:20 hindsight critics Spanglers. Ironic on a thread about forecasting – wot !
Yes Luke,
Hindsight is a wonderful thing.
It’s also wonderful when we learn valuable lessons that hindsight can give us.
The real point about water management (IMHO) is that it is lacking some very necessary flexibilty and therefore too focused on centralised rules that don’t always match the circumstances. It isn’t only the management of Wivenhoe that has seen this problem emerging. We’ve seen similar in our part of the world.
I thought the paper was about seasonal forecasting. My prediction is that Jens gambit claims like “shift an accepted scientific paradigm” (AGW) and “natural rainfall cycles have not been perturbed by the current elevated levels of carbon dioxide” sound like a tangled web of denialist spin and weave – nothing to do with the actual content of an evidence-based paper that is hopefully non-ideological. All potential users in Queensland would have wanted to know is does it do better than the SOI? . Pitching it against a GCM is a contrived battle of straw men. Users don’t care that neural network analyses make no contribution to knowledge – models do that.
Given the backgrounds of the researchers and the funder, and their lack of any track record in seasonal forecasting, you would have to wonder.? Are you being having a lend of.? Real scientists are usually humble until there work has been shown to be repeatable, refutable, and comes from a reputable stable.
Bazza,
I hope it’s non ideological too.
That gives it more chance of being useful.
Your typical MO of ‘shooting the messenger’ hasn’t changed I noticed.
If Jen and Abbot are on to something that could be helpful to Agriculture and Mining what does their backgrounds or their source of funding matter?
‘No serious floods in the intervening period. No action. No experience. No rehearsal. Years spent dealing with a mega-drought in the leadup.’
It was not a ‘mega drought’, just a typical long drought which is automatically replaced by flood on a regular basis.
The inexperienced engineers responsible need to be debriefed and brought up to speed. AGW is a myth, plan for the future by looking at the past and don’t take your eye off the ball.
El Gordo – Nope worst drought on record in terms of Wivenhoe catchment inflows. Worse than the Federation Drought. Figure 1 – http://www.longpaddock.qld.gov.au/about/publications/pdf/seq_drought_2007.pdf
“floods on a regular basis” – more rot- look at the record – floods are regular are they – this is your take on “regular” http://www.bom.gov.au/hydro/flood/qld/fld_history/brisbane_history.shtml
Hilarious Debbie – “too focused on centralised rules” – that will be of great comfort for 3 engineers now fronting the Crime and Misconduct Commission for a poompteenth of adaptive management.
And even more humourous Debbie – you say ” how would you know if Jen’s work helps agriculture” – well it indeed might…. How would you use this information in your area – oh sorry – I forgot you’re not interested…
Luke,
please follow your own advice and . . . Breathe. . .
I agree there is too much focus on litigation.
I also didn’t claim I wasn’t interested, I merely pointed out that us landholders aren’t that impressed with packaging, we’re more interested in shared goals re the funding.
I can’t see why this is degenerating into a ‘them and us’ discussion.
Jen and Abbot have not tried to do something counter productive here.
It could very well be useful.
The Resident Dipstick is at it again:
” … 3 engineers now fronting the Crime and Misconduct Commission for a poompteenth of adaptive management.”
The allegation is that they didn’t follow the dam Manual and then tried to cover this up
It is the allegation of cover-up, not “adaptive management”, that they are facing the Commission for
The Dipstick truly is pathetic – malignant mendacity is his true forte
Okay, maybe irregular, its a big island.
Thanx for the links, I can see why they they may have been disposed to think it would continue dry. So yeah, point conceded.
The 19th century appears to have been wetter and we may have already entered another cool/wet phase, which will require a rewrite of the manual (sic).
“malignant mendacity” – sounds terminal. Unlike your decrepit drongoism which goes on forever. Tongue in cheek clearly out of fashion among the humourless
Drought will come again. Sometime or somewhere in Qld, some future dry spell will break “records”. But we need to assume that that the floods that occurred in 1841 and 1893, which were probably called “unprecedented” in their time, will come again. In fact, the 1974 and 2011 floods were a pretty good re-run, though all those disasters were, of course, different to one another.
Just before the Qld flood of 1890 (not quite so bad as 1893), there was a drought so severe it killed off native forest.
Preparedness is more important than anything. In the midst of flooding, we should be thinking about drought – and vice versa. Part of being prepared is to welcome the efforts of those who strive to know a little about climate.
As for those who have claimed to know a lot: I suppose we could go on giving them Nobels, Oscars, Australian of the Year awards, Logies, Maze-Master medals…anything at all. So long as we never believe them again.
What I want to know is if this dab of software on tab can do better than my brain or yours for that matter, given all that weather info we can see on the TV. Add the BoM radar on the net and it’s an armchair observer’s paradise.
80% chance of rain in Canberra tomorrow so I won’t pack the wagon for a half day out in the open. It’s still dark for some of that time and we can’t follow the clouds overhead. Forecasting hour by hour the next shower demands some skill in imaging based on the day before and a few other experiences.
It will be quite a while before I hand over to something on a lap top.
Great stuff, go get them Jen.
‘…there was a drought so severe it killed off native forest.’
That was Luke’s mega drought.
Luke, do you have any idea what destruction would have been caused had those “fuse plugs” blown?
Those “fuse plugs” are an earth fill spillway and the scouring would have been horrendous.
That famous aerial shot of the 200%+ full dam within inches of destruction told a story of managerial neglect of a potentially catastrophic situation by those in charge. Worse than Christine Nixon in charge of the bushfires.
“Nope worst drought on record in terms of Wivenhoe catchment inflows. Worse than the Federation Drought.”
Luke, BoM have no idea how the inflows in those catchments compared with the Fed drought. Inflows have no correlation with annual average rainfall as any farmer will tell you. Also none of those dams existed during the Fed drought.
More BoM GCM GIGO modelling?
And my rainfall nearby showed average or above for 4 out of 7 of those so-called 21st C drought years.
Wasn’t that convenient that Luke’s link to the BoM didn’t show the 2011 floods?
Maybe Luke can refresh our memories; there were some emails floating around in, I think the Australian, to the effect that the duty engineers were screaming for permission to begin releasing on the weekend because they could see historic inflows to an already full dam?
As a farmer dependent on rain I would like some certainity in the 5-10 day range of forecast. I am sick of the forecast telling me it is going to be wet/dry that far out only to get within 1-2 days of the planned event, spraying, harvest, planting whatever and seeing the opposite. When I check the “forecast” it has changed to match the obvious, what’s the point! My climate/weather “model” may just as well be a piece of string nailed to a post!
“Inflows have no correlation with annual average rainfall as any farmer will tell you” – well of course not – in fact they are based on very well calibrated rainfall-runoff models by engineers (not BoM) based on daily data and catchment soil water balance.
And of course the dams were not there in 1900. Sheesh. But it may surprise you that rainfall-runoff models exist and work. Used every day in fact. But you know maths is a funny old thing – little rainfall for years on end just seems to produce NO RUNOFF. Ain’t that amazing !
ANYWAY forget that – try rainfall deficit – Figure 15 on page 25 http://www.derm.qld.gov.au/climatechange/centre/pdf/climate-change-in-queensland-2010.pdf which shows an even longer period on rainfall deficit – the Millennium drought went way beyond the Federation drought in that region – which is why folks were worried.
Yes of course fuse plugs would cause a massive release of water, downstream damage and scouring – but the dam would have stood (somewhat more important eh?).
NOT what you previously claimed. Don’t bait and switch eh? The plugs are to prevent overtopping and catastrophic failure.
On 2011 floods and graph. Hardly convenient Spanglers – it’s a dated web graphic that has been there for years. Do go on. Why don’t you get your little texta pen and draw the 2011 event on your screen.
Runoffs in below average rainfall years can exceed average years and unless you are there to observe, all the modelling of century-old evidence isn’t worth a cup of water.
And with a Dept of Climate Change trying to justify itself, I have to hang extra sceptical.
And Luke, have you ever stood at the top of those “fuse plugs” and considered the obvious destruction if they over-topped in a big flood?
You wouldn’t have a functional dam again until well after the La Nina. Possibly years!
And I’m sure if you really wanted to be honest you could dig up a more up-to-date graph.
“Runoffs in below average rainfall years can exceed average years and unless you are there to observe, all the modelling of century-old evidence isn’t worth a cup of water.”
Whew. Somebody’s actually observing and thinking.
He may be spangled, but he’s certainly no drongo.
Sigh – but we weren’t talking average years were we. Drought years. Double sigh. Is the realtive comparison between droughts enough. Did the Millennium drought go for longer. Triple sigh.
And in terms of runoff in average years a lot of that has to do with antecedent soil water conditions (wet or dry catchments) which is why they have soil water models attached. It’s a bit more than runoff = a + b (Rainfall).
A more up to date graph – why not use your texta ? Does adding 2011 SUDDENLY make it “regular” does it. No. Sheesh.
“Climbed on top of fuse plugs?” – well did you hop the fence eh? Norty. I’ve only looked from the fence as you’re supposed to. 3 sections are part of the auxiliary spillway – in a concrete walled channel.
http://cdn9.wn.com/pd/2a/92/9eb2fd13c646997e6d9f099e19e4_grande.jpg
What a picture ! http://bushtelegraph.files.wordpress.com/2011/02/clip_image002.jpg from
this in the dry http://www.springhillvoice.com/images/wivenhoe.jpg
And Spanglers what’s this rot about not having a functional dam again for years if the plugs subsided. The dam wall would be intact. Most of the water storage capacity would still be there. Some temporary lack of flood mitigation yes. But the fuse plugs in their concrete channel could be rebuilt??
And it wasn’t a dept of Climate Change graph – they simply quoted it.
Exactly QF!
Observe:
http://www.yr.no/place/Australia/New_South_Wales/Coleambally_Creek/
http://www.eldersweather.com.au/nsw/riverina/coleambally
http://www.bom.gov.au/nsw/forecasts/riverina.shtml
A week ago they all said with a 90% + probability that we would have at least 25ml today and tomorrow.
Today they all contradict each other.
Often it’s the other way around…..and still not correct.
So if Jen and Abbot are on to something that is capable of ‘less error’ and focused on being ‘more useful’ to agriculture and mining….then more power to them (and probably more funding too) IMHO.
BoM and the other sites have started to collect and categorize impressive amounts of reliable historical data….how it is being used and what research is/isnot being funded is the big question.
Us landholders are just a tad over being dictated to by some of these bodies…..especially when we are fully aware that much of what they deliver is way too speculative…..and therefore not that much help as a decision making tool.
Don’t get me wrong….I hope they do get better at this stuff….there will be no one happier than us farmers when this work is capable of ‘less error’.
I wish you the best of luck Jen and I sincerely hope this delivers on some of its promising early indications.
“(3) very probability of exceedance rainfall events on a full dam (well full in terms of water supply not flood compartment)”
The only partially true thing luke has said; Wivenhoe was an AGW dam; full to cater for the endless drought ‘predicted’ by the BOM; it could not handle the cyclical 2010 storm which was almost identical to the 1974 event; and it was full because the pro-AGW modelling was adopted by government instrumentalities.
Anyway the upcoming litigation may shake a few dead birds from the AGW tree in respect of AGW predictions and public policy based on those predictions as described here:
http://www.onlineopinion.com.au/view.asp?article=13303
We’re actually very naive how POAMA would be used for seasonal forecasting. In fact a country mile off the pace.
BoM are not necessarily looking at downscaling the rainfall out of the GCM – moreover they are using POAMA to forecast a bunch of indices – Nino regions, EMI, DMI, 140 Blocking etc.
Then forecasting rainfall from those indices.
Importantly they’re also considering hybrid systems of POAMA and normal indicators which mix various predictands optimally by region based on best fit. And that varys at time of year.
They’re considered covariance amon the indicators. New indicators we haven’t considered – El Nino Modoki, 140 Blocking
Also important comparions with persistence and climatology. AND sometimes climatology is the best !! (no skill)
We’re a mile off the pace in terms of our discussion. Totally immature.
So doing a comparison of any new system versus POAMA is not even what you’re up against !!
But I’m sure you all knew about this work (didn’t you?)
http://www.cawcr.gov.au/publications/technicalreports/CTR_031.pdf
http://www.clw.csiro.au/conferences/wirada-symposium/documents/presentations/Wed/Session6b/1100_Hendon.pdf
http://www.cawcr.gov.au/publications/technicalreports/CTR_031.pdf
Cohenite will love this !
What sort of AMAZING drivel is this – Cohenite has soiled himself ! “The only partially true thing luke has said; Wivenhoe was an AGW dam; full to cater for the endless drought ‘predicted’ by the BOM”
And this absolute PORKER – “cyclical 2010 storm” – hey Spanglers – they’re now cyclical. OMIGOD.
As if ANY sceptic forecast the Millenium drought or knew years before when it would end. As if the IPO can be forecast !
Excuse me – Wivenhoe was built in 1984 and management rules date back to those times. GOOD GRIEF !
“AGW Dam” – ROFL and have sprayed coffee all over my monitor (just after having punched the last one out).
And more to Cohenite’s disgraceful try-on – the regional bureau was advising all through the leadup to the flood – months before – about a strong La Nina. Everyone on the street was saying that it looked like the drought was ending. Cohenite well illustrates why sceptics are fast and loose with the truth.
Luke,
Stating the bleeding obvious eg:
But you know maths is a funny old thing – little rainfall for years on end just seems to produce NO RUNOFF. Ain’t that amazing !
Is not doing anything to help your case.
You do have a point about what has happened to those engineers….but you’re missing the point that SD has made and therefore what we’re all trying to say about water management and the insane obsession with inflexible, impractical and myopic rules that are incapable of the adaptability that is required when (quite typically) the circumstances alter from one extreme to another.
The managers are actually safe from litigation as long as they followed the rules…..it’s the inflexible rules and the mindset behind those rules that is the problem (IMHO)
Our rainfall, inflows from varying valleys, sudden spectacular storm events, timing between one rainfall event and the next and numerous other variables including the high probability of low inflow periods are all reasons why we built water storages in the first place.
Those simple truths about our climate mean that this obsession to mathematically formulate a magic set of concrete centralised rules is doomed to fail.
The climate/rainfall patterns is simply not interested in conforming to yearly water accounting formulas or even in forward borrowing, pay back environmental flows and numerous other rule changes that have been introduced long after the exit of Joh’s Govt and interfered in the actual original purpose of building water storgae.
“It’s a bit more than runoff = a + b (Rainfall). ”
He got that right, between sighs.
Let’s hope our models are supermodels if we get another 1893-style flood in Brissie. It was actually three distinct floods in February, followed by another in June! Wipeout.
The 1890s certainly will do for “extreme” until the real thing comes along. In my region we copped a major flood three years in a row. In 1899, there was a freaky Big Wet in normally dry August, followed by our lowest-ever measured fall for January. You can probably chuck in a major drought too, though our driest year since measurement was ’02. (No, silly, NOT 2002!)
Now droughts are longer! Can’t win sometimes, can you? Everything’s worse than we thought.
“As if ANY sceptic forecast the Millenium drought or knew years before when it would end. As if the IPO can be forecast !”
Luke, stop trying to justify your mate Tim. Wivenhoe was built to guard against droughts as well as floods. At least sceptics look at both sides of the coin.
And my main point about the “fuse plugs” which you don’t understand, is that this destruction that came within a whisker of happening right under the nose of all that great “supervision”, just proves how poorly managed the dam was at that critical time when it was plain to many, by just looking at the radar for the previous days leading up to this, what was really happening.
Even a drongo like me could see it. I was ringing every phone no. for SEQWater and politician I could find and no one wanted to know. They were all out to lunch, like Christine Nixon.
To simply alibi this as “shit happens” is so typical of what’s going on in lefty govts everywhere.
If they had given the management to the local farmers co-op they would have got a much better result. It only takes common sense. Wivenhoe has a 100% safety margin and panic stations don’t arrive suddenly.
Oh come on – engineers were there – Lockyer happening – locals complaining about releases – and you get a 1 in 5000 event in the catchment
“Everyone on the street was saying that it looked like the drought was ending.”
A couple of things; drought finishing is not AGW; and in respect of AGW Bolt has compiled an interesting list which includes Flannery but mentions many more people on the street ofAGW:
http://blogs.news.com.au/heraldsun/andrewbolt/index.php/heraldsun/comments/column_the_alarmists_have_taxed_our_creduility_once_too_often/
It is scurrilous to suggest the 2010 flood was predicted and that adequate precautions had been put in place for reasons I have listed above.
Having had a look at the technical report on POAMA 1.5 which luke kindly supplied I would say that over the years I have been grateful for luke’s links which invariably involve luke shooting himself in the feet to such an extent that he must now have mere stubbles where his limbs were once and capable of transporting him down the corridors of the AGW citadel.
Jennifer has applied an Artificial neural network to real data based on three climate indices (Southern Oscillation Index, Pacific Decadal Oscillation and Nino 3.4) to establish patterns and cyclical behaviour from which predictions can be made about rainfall.
The technical analysis of POAMA 1.5 says this:
“This study has assessed the possibility of using predictions of standard climate indices as possible predictors of Australian rainfall. These results suggest that there could be a significant benefit of using the predictions of the climate indices to provide better rainfall forecasts than are directly available from POAMA, especially at longer lead times. At longer lead times (4-9 months) the most skilfully predicted indices will be those associated with ENSO. Prediction of the IOD indices will be primarily useful only to a lead of ~ 4-5 months. The indices associated with atmospheric variability (140oE Blocking Index and SAM) will not be useful at lead times longer than 1-2 months.”
That reads like a vindication of Jennifer.
Well said Spangled,
The purpose and construction of Wivenhoe is rather impressive.
It did indeed have that 100% saftey margin in mind when in was designed.
We all know that ‘shit happens’….that’s why it was built that way. Plenty of ‘shit’ has happened in that particular catchment in the last 100 years or so….as has happened in all the other major catchments in Australia.
I’m rather amused that Luke is appearing to attribute blame back to 1984…..if the same mindset was in operation as was in operation at construction…..the management of the dam would have been different as the seasonal indications started to quickly turn around.
There’s little question that the intervening drought and the public panic and changed public expectations created by AGW and the AGW celebs like Flannery et al, influenced the management decisions and the rules and regulations….and Luke…..THEY HAVE BEEN FUNDAMENTALLY CHANGED!!!!!!
It has now become more about bureaucratic accountability and ‘cost recovery’ and what I personally believe is an insane idea that water can be managed inside a financial accounting year from July 1st to June 30th.
Inflows and rainfall are NOT INTERESTED in human invented calendar points…..they won’t co operate that way and that can’t be managed that way (IMHO)
Sorry for the shouting….but please pay attention to the actual point that is being made here.
As a further observation, the work and the focus of the research that Jen and Abbot are conducting could definitely help to focus the management of dams back into the realm of some much needed common sense…..that looks at seasonal indicators in a more practical manner.
Debbie, isn’t it wonderful how the Lukes can blame the Johs and yet alibi the current managers?
If La Ninas average about one every four years and big floods only once a decade or so, then wouldnt letting water go every La Nina empty the dam? Researchers having been using the SOI to look at water resource management for close to 3 decades. They have not missed anything. it would be nice if you could run the world on opinions, beliefs, fears, but you can do better by actually checking the facts.
To those who have been using secret blog language to urge that water be let go every La Nina, consequently emptying our dams:
Identify yourselves!
You know who you are, fairy commenters!
‘Everyone on the street was saying that it looked like the drought was ending.’
I don’t doubt you Luke, but a link to back that statement up would be appreciated.
No Bazza,
It would be nice if we could run water storages with some common sense and pay attention to seasonal conditions not inflexible and impractical calendar based rules.
They have missed plenty because their administrative focus/process has changed and they have altered the rules and redefined the purpose behind building the storages in the first place.
It is not just about the extreme events like droughts and floods…..they’re always difficult….as Luke says….’shit happens’. Dams like Wivenhoe do however have the capability of smoothing out the worst of those extremes if they’re managed with that in mind.
So while your so called facts have some basis in truth….you have forgotten to factor in the other approx 6/10 benign years that would usually happen in those approx 30 years since the construction of that particular dam was completed…..there are lag times….particularly after floods….especially in relation to excess inflows that are available in later seasons and not necessarily related to the size of rainfall events…..or did you miss that part?
They happen because the catchments have been given a bloody good soaking by the flood events.
So the answer to your rhetorical question is…..in all probability…..no…because in all probability there will be plenty of inflows available to top up the storages later in the season/s…..but that will not necessarily happen conveniently inside predetermined, water accounting calendar points.
Can’t win, can they Deb?
I didn’t blame Joh at all. Simply the manual needs revising in 20:20 hindsight. IMO a review before the 2011 event would still not have seen the special extreme circumstances of 2011 – off the record events being what they are. IMO
Drought periods in 90s and 00s lulled us in false sense of security.
People can rant and rave about Flannery and AGW as much as they like – Lake Wivenhoe was ta rock bottom – worst on record drought – no end in sight – and Brisbane 1 year off no water. I think if you were on watch you might have been getting nervous. Another drought out there – more people – need more water – water grid will not be wasted.
So in a debate locally obsessed about following manuals (DID THEY FOLLOW THE MANUAL !! – ANSWER THE QUESTION !! http://www.youtube.com/watch?v=hopNAI8Pefg ) yo’all seem to want some latitude to do some adaptive management using both probabilistic weather and climate forecasts. Sounds like fertile ground for class actions and parasitic lawyers. But using forecasts should happen. And there are downside risks (which everyone will ignore).
You may wished to discuss instead of hoiking gibbers.
Cohenite – you need to read more.
You’ll find they’re now at 2x skill POAMA 2.4
Proposing POAMA predict indices not directly downscale
Proposing hybrid mixture of POAMA + indicators + climatology (where appropriate)
Analysing covariance between indicators
i.e. they’re a light year off in another galaxy. The debate we’re pseudo having here has been bypassed. We’re not even close to the current reality (SOI Phase for Qld) or we’re BoM are headed.
Add me to the dummy bin too. We’re simply not up to date. All over most people’s heads of course.
So instead of hoiking rocks – do some reading.
“i.e. they’re a light year off in another galaxy.”
Are BOM still adding CO2 to the mix? If so, they certainly are “in another galaxy”.
It’s fascinating how this CO2 contextual red herring has dominated the discussion. Talk about obsessed. Although I suppose serious seasonal forecasters would be worried about their baselines.
Luckily the Walker circulation weakens in response to global warming. The SOI does not decline in response to global warming. http://journals.ametsoc.org/doi/abs/10.1175/2011JCLI4101.1?journalCode=clim and http://www.cawcr.gov.au/meetings/fd/SPCZ_workshop_Apia_presentations/Power_SPCZ_wshop_Apia_2010.pdf but let’s not get distracted.
SST devotees may not be as lucky.
“Luckily the Walker circulation weakens in response to global warming.”
You keep repeating this luke and you know it is problematic:
http://landshape.org/enm/walker-circulation-and-enso/
http://landshape.org/enm/files/2011/01/walkerarticle.pdf
Why do you put all your faith in POAMA Luke when you know it is only one of a number of models contiunally run by the BoM? You seem to hold it as some holy grail. Much of the work on POAMA consits of trying to validate the out puts against real world. POAMA still has considerable work to go and it will still only be a model amongst many.
C02 contextual red herring?
Talk about obsessed?
Oh the irony!
ROFL!
Bronson – I didn’t say I was in love with POAMA – it happens to be what we’re discussing here. (or trying to).
As usual Cohenite you lot have missed the whole story with your unpublished shady stats dalliance and not knowing what you’re even doing.
“Global warming weakens Walker circulation but (surprisingly) increases SOI. The SOI is not a good guide to changes in Walker circulation forced by global warming.”
What Caused the Observed Twentieth-Century Weakening of the Walker Circulation?
Power and Kociuba
Journal of Climate, December 2011, Vol. 24, No. 24 : pp. 6501-6514
Who watched this program last night?
“How To Grow A Planet – Life From Light”
http://www.abc.net.au/tv/guide/abc1/201206/programs/ZX9029A001D2012-06-03T193200.htm
We are well down this path now, but Deb; something is out of whack
http://en.wikipedia.org/wiki/Gaia_hypothesis
The ABC is definitely out of whack. It now beats the commercial media at retailing the nastiest and silliest mental trash.
Robert; are you by any chance, just a creationist?
What’s out of whack Gavin?
Little Lukey rambles on with little contact with reality:
““Global warming weakens Walker circulation but (surprisingly) increases SOI. The SOI is not a good guide to changes in Walker circulation forced by global warming.””
Lukey, remember when we Deniers started hammering y’all cause the FINGERPRINT of Gorebull Warming was missing?? Remember the silly response y’all gave?? Even though the charts in AR5 were quite clear y’all tried to tell us that ANY warming would cause the hot spot.
Try and guess what my response to “Global warming weakens Walker circulation …” is!!!
HAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHA
“Robert; are you by any chance, just a creationist?”
Gav, we come come from all schools of thought, those of us who have realised that the ABC is a greater insult to viewer intelligence than a whole team of Willesees and Hinches. You might just as well be getting your current affairs from Tracey Grimshaw and your doco knowledge from The Flintstones.
Yes Luke @ 6.22pm
They followed the manual. And if they followed the manual they probably won’t be in legal trouble.
The problem however is what’s in the manual.
It is rigid and unable to adjust to prevailing circumstances because of slow and cumbersome bureaucratic processes that become even more difficult on weekends.
Something remarkably similar happened here in the first week of March.
By the time all the green and red tape was ticked off, the situation was already out of control.
As Quiet Farmer said:
the duty engineers were screaming for permission to begin releasing on the weekend because they could see historic inflows to an already full dam.
They knew trouble was brewing….but they were unable to do what little could have been done because they had to wait for permission.
However,
The fact that Wivenhoe was also over full and too close to its outer safety limits before this situation got out of control also needs to be questioned.
There were already sufficient inflows to run the dam at a safer level before the disastrous circumstances eventuated…..that catchment was already wet and the creeks etc were already delivering good inflows.
Why was it so overfull when the catchment was already wet?
What was the prevailing mindset behind those decisions?
Debbie you need to get some background on what “full” or 100% means wrt Wivehoe. Full to water supply level of full – or the flood compartment level full.
http://www.seqwater.com.au/public/sites/default/files/userfiles/file/pdfs/Factsheet-SomersetandWivenhoeDams.pdf note the diagrams and stats
http://www.australiangeographic.com.au/journal/brisbane-floods-did-the-dams-work.htm
So the system was working, engineers on duty. But this was very rapid extreme high end event situation.
The issue is the speed and proactiveness of the release sequence. My understanding is that a one in 5000 rainfall event in the catchment mid-event was a major gotcha.
So it was going to flood to some level – how much if any could have been saved is not clear even after the inquiry. http://www.qt.com.au/story/2012/03/17/show-me-answers-demands-paul-pisasale-ispwich/
Engineers now referred to the CMC for allegedly making up the sequence story post hoc. However this was an extremely stressful period with people on little sleep for a week. Equipment failed. So it’s not totally obvious IMO.
The only great solution would be to drop the storage below water supply level full before a La Nina. Now that might be brilliant. But if it doesn’t flood you’ve also wasted millions of dollars of water into the ocean.
So I still have not heard here any analysis of the up and downside of this modification to the dam manual. Certainly the response needs to be progressive and adaptive. Everyone just seems to want to kick the engineers.
And if your home was flooded and you’re in a class action how can you be calm and dispassionate about the issue?
So Debs – ain’t all about “packaging” – you have to discuss the forecast systems with users 🙂
And for a touch of AGW (and not implying anything in 2011 necessarily) – the mid 00s CSIRO modelling work indicates more extreme supercells like 2011 in a globally warmed SEQ.
It is true Wivenhoe has 2 capacities; a water supply capacity, which was at 100%, and a further flood capacity allowing for a > 100% overall capacity.
The 2 issues which arise is why the water supply capacity was at 100% and why water releases were at the rate they occurred. The law case will establish that. In the meantime this provides a reasonable overview:
http://www.theaustralian.com.au/in-depth/queensland-floods/engineer-warned-of-danger-from-wivenhoe-dam/story-fn7iwx3v-1225991369355
Gav, gaia and veggies, eh? I say that was nuts but then they are plants too.
As the neural net (NN) paper concludes for Qld stations “ the error values suggest its skill at forecasting rainfall with a 3-month lead at least as good as POAMA..”. And Table 4 confirms that. But I am curious as to how come there is an extraordinary correlation between the errors for POAMA and NN when they appear to be totally different approaches. In fact you could predict NN station errors with 95% accuracy using POAMA and vice versa which makes one of them redundant. Maybe it just reflects the skill of seasonal forecasts – a lot better than chance but a lot short of perfect. Maybe it is something to do with seasonality introducing spurious results. Maybe both approaches simply owe a lot to various ENSO measures. I give up.
Supercells on the way! I hate those things. We still haven’t got over ’49-’50 around here. Or that triple-whammy in the 1890s I was talking about. And don’t even mention those supercellular seventies.
Supercells in CSIRO modelling! Lots of worse-than-we-thought, complex-but-simple, extremer-than-once-upon-a-time “records”, no doubt.
However, since the Arctic death spiral and the evacuation plan for the NSW coast, it’s getting to be like when you hear that awful crash from the lounge room. Turns out it was just someone playing Tom and Jerry or Die Hard too loud on the TV.
Luke, I haven’t read the manual, have you? but I’ll bet it covers a situation like Quiet Farmer describes, such as 190% full dam and huge inflows still occurring and an acknowledged huge wet season with 6 months to run.
I’ll bet it even covers a similar situation with only a 150% full dam and I’ll also bet there is room for precautionary discretion in those situations.
I also don’t know who was responsible for opening the taps but whoever it was, was not paying attention.
As I said before, decision-wise, there was not only no one on watch, there was no one on board.
It seems that with lefty regulators, the more red tape and regulation they draft, the more they are are strangulated and constipated by it.
And it isn’t rocket science to know that when there has been a long dry spell, that if you get a La Nina plus not having had a flood in thirty five years, the chances of a good wet and a possible flood are on the cards, as Campbell Newman said in October 2011.
Luke,
In essence we agree (amazingly)….except for your tendency to over protect what was clearly a cumbersome bureaucratic process.
Cohenite asks the pertinent question which is similar to mine.
Of course that rainfall event was a doozy…..no argument there…..BUT…..and this is a big BUT….the catchment was already wet and that dam was at full supply storage capacity and it was kept that way despite the fact that there were already suficient inflows happening to leave a bit of extra safety room…..in high inflow conditions and flood conditions millions of dollars of water will always go to the ocean….can’t prevent that unless we build many, many more storages.
IMHO the mindset that hung onto the income producing portion of the available water is questionable as it didn’t seem to recognise the prevailing conditions and the inherent danger….that doesn’t mean that they should have emptied the dam…..but it does mean they could have considered the high probability of sufficient inflows from that wet catchment and adjusted the management accordingly.
I do understand the capacity formulas of Wivenhoe….I think it is quite impressive and showed some real foresight in the engineering design…..much better concept than our dams here.
BTW….I’m not kicking the engineers Luke….they are not the culprits (IMHO)….they rarely are.
I hope they have the courage to highlight the inadequacy of the process….but to protect themselves and their jobs….they probably won’t and I can’t really blame them.
Also….We were completely flooded here in March….there is also a class action happening (I am not personally involved)…and you’re right….being calm and dispassionate is not usually a feature…..but interestingly….the focus is on the process and the rules that had been adjusted during the drought years….not on individuals who could not do anything until they were given permission….which of course was too late.
Like the Wivenhoe situation….the problem had started to develop in the preceding weeks…..but nothing was done…and then we got whacked with a doozy rain event.
So….if Jen and Abbots work can be added to what we already know about these things…..then it may help to inform earlier, smarter management decisions.
But please get your head out of that ’empty the dam’ argument. That was not ever going to happen as that would be just as silly if not sillier….no one has argued that should have happened.
It isn’t just about rainfall….it’s also about the state of inflows, how wet the catchment is and the prevailing seasonal conditions.
Under the right conditions, even 100ml events can cause a fair bit of angst.
At least Jen is showing us her methods.
More than some:
http://bishophill.squarespace.com/blog/2012/6/3/who-are-you-calling-a-charlatan.html
Cohers – full to overtopping is 225% – water supply full is 100% – they got to 190% I think.
SO think about it Debs – if dam is 100% – big wet is forecast – you drop the dam to 75% – doesn’t flood. Next season turns to El Nino – you’ll be asked why you dropped the level – cannot you see the potential dilemma.
Anyway – focus on Bazza – he’s on task !
Nope,
You’ve missed the point Luke.
The catchment was wet and the inflows were already up…..dropping the dam a bit would not have created that sort of problem….there were already usable inflows available to refill if the big wet did not eventuate….a normal wet was more than plenty to alleviate that problem…..there is only a finite amount of storage available there…..would be different if there were more dams I guess.
But of course….there’s always potential dilemna…..it’s never going to be perfect…..especially if you try and make it fit neatly inside a water accounting year.
Bazza said he’d given up?
Another example of mad modelling by the wacko green regulating classes that has the opposite effect of what’s intended.
The precautionary principle gone mad [again] much to Australia’s disadvantage:
http://aefweb.info/data/Australias%20Unappreciated%20and%20Maligned%20Fisheries%20May%202012.pdf
Just announced by Senator Farrell. “$4 million to support a national hydrologic modeling platform”. Could have saved his dough and asked some of the resident intuitive hydrologists right here.
Deb, I had not given up, only on offering explanations so over to you and any one else.
Climate and water policy in Oz are starting to resemble Kruschev’s agricultural policies: compartmentalised thinking, over-riding dogma, self-multiplying bureaucracy, and just enough knowledge and bright ideas to get into a world of trouble.
Well, at least we don’t have Lysenko advising us. Although…
So how would you like complex science to be done Robert?
bazza; A guy who came to my place this weekend started chatting about tools and and other things we used then got on to risks in the job. He is a tree doctor now but had been trained for power line maintenance. He left that authority after a couple of co workers were knocked off some high voltage lines. Next topic was induction and what a boot that can be.
All forces unseen are a threat, especially those encountered at heights, up poles and round towers. Many routine operations are so stressful that long term individual exposure is not desirable. So, when do we quit? Who do we train next?
Construction work after design still requires a level head and all work with a disadvantage for the individual needs constant monitoring. How experienced are the consultants?
Mighty projects do go wrong despite our best endeavors and people do loose lives. Very few have such foresight that management has no problems. One of my early mentors said in the end, perfection is only an illusion. We modified workplaces after industrial accidents.
Deb; hindsight is a wonderful attribute.
bazza
“Could have saved his dough and asked some of the resident intuitive hydrologists right here.”
Perfect, he should have.
The accumulated wisdom of man on the land is invaluable. Problem is nobody ever asks or if they do it’s only so as to be polite and then ignore it. After all book learning is sooooo much more accurate than actual observation. And don’t think I’m knocking learning, I’m an engineer with master’s.
Not saying every landholder is a wise individual, just as getting old doesn’t impart wisdom, but politicians and academics refuse to listen, they think it’s beneath them or they think they know better or some other ulterior reason, I don’t know.
Robert; creationism stunts growth.
JW; I’m worn out cleaning up after clever bods
Gav, I’ll pass on your warning to some Seventh Day Adventists in our valley. You know how everybody just loves being lectured and tut-tutted by posh Canberra leftists.
“So how would you like complex science to be done Robert?”
Essentially, by people who have an inkling of the enormity of what they don’t know and can’t do and can’t predict. I think they’re called scientists. They’re so infernally bright that they work unscripted. Imagine that!
gav,
For the last time, I don’t care if you post or not! Your ramblings confuse me.
I once told you, you resemble the proverbial older tradesman in the movies who puts the young inexperienced engineer in his place. It wasn’t meant as a compliment although you took it as one.
If you go back a few years and look at your own posts, you started out as an instrument technician, over the years your qualifications have grown to that of, I don’t know what?
Pity I can’t put you on an ignore list here, I’m trying to be polite here.
Please do not reply to my posts.
Have a nice life.
So Robert – would you like them to work on unscripted on the tax payers dollar as mavericks on increasingly complex problems with multiple criteria for success for those involved.? I’m not sure myself anymore…..
Totally eclipsed you on that one Gavin.
Go back and reread the earlier posts.
Deb; help, what have I forgotten? You write so much.
JW; It’s so bad you can’t read me cause I’m busiest when sharpening other peoples tools.
Robert; still in the closet?
Luke, I’ll be specific. By “unscripted” I meant that a pure researcher should feel free to come up with any findings at all – or none at all – regardless of intellectual fashion or what the patron thinks or wants. Failure is fine when it is recognised as such. Expense is fine even if it yields no results, provided you learn. A best-possible conclusion can be a disaster if you don’t see its inadequacy, regardless of how “high quality” it is. Better to waste millions learning than waste trillions pretending. (Think of all those “best available” piles of junk called alternative energy, while Australia still has no nukes and still burns its coal in ancient clunkers. Alarm bells?)
Of course, so-called believers in CAGW will claim that impartial science has delivered a harsh truth to a selfish world. The harsh truth is that nobody truly believes that human generated CO2 is a danger. This is easily demonstrated by pointing at the hilariously bad solutions proposed, the stupendous financial waste that feeds the major climate scams (GS,GIM), the frittering of resources (including coal!), and the private behaviour of all “believers”, great and small. CAGW is about satisfying the collectivist urge so dear to elites. This new form of collectivism works more by paralysis than by force, but the authoritarian element is unmistakeable.
In spite of its messy appearance, prefer freedom, always.
Alas Robert – increasingly difficult in this project planned, MBA driven, performance managed world of modern science dominated by boards and steering committees. But people want accountability you see.
Gavin’s contributions here are invaluable in keeping us grounded.-
And what appalling condemnation Robert when AGW research has revealed many of the mysterious climate drivers around Australia in the last 20 years. What an amazing comment. But then again you don’t read any journals on the subject do you?
“many of the mysterious climate drivers around Australia in the last 20 years.”
You referring to Greg Combet or ACO2?
But Juliar’s Carbon Tax is the solution and its cost to measure and collect will safely see us all, jobless and foodless, in the third world, toot sweet.
Betcha it doesn’t
Luke; News today, Top at our Nobel famous research institution where I had a job for a mo was asked by Vice to go find his own dough. Free wheeling may be a thing of the past for all science given the current gloom.
SD; I’ve anticipated a big fall in aluminium use here post Carbon Tax, stored a cross section of domestic Al products for the record but an industry guru on RN a few days back said they could cut electricity consumption by 25% after a push from governments. That means taxpayers here are already on a ride with age old subsidies.
On the dam topic; twice experienced operators who were confronted with catastrophic system failure. Both abruptly dumped their pressure vessel load to atmosphere without consultation. Being a witness at any inquiry is easier if some evasive action is taken during the crisis.
Jen,
despite all the yabber here from people who think they have special and secret knowledge of climate and weather that is just too complex for us mere mortals, I congratulate you and Abbot and I sincerely hope that your work does deliver on its early, promising indications.
I particularly congratulate you both for focusing on how this work may be useful for industries like mine which rely heavily on seasonal weather.
It looks like a very useful way to use the increasingly impressive amount of historical data that has been collected by various organisations.
Can someone translate what gav said?
But Debbie you’re not interested in packaging – so how will you be using it?
And how will you choose between this system and the number of others?
No catches – really basic questions.
What am I to do?
http://www.bom.gov.au/nsw/warnings/severe.shtml
Luke,
It is extremely disappointing that you have asked that question.
You obviously don’t listen or you are unable to countenance what producers actually need. After many discussions at this blog you are still criticising and dictating.
It is a disease that far too many bureaucrats and academics suffer from.(IMHO)
I have already answered that question earlier.
Rather than making a fuss about packaging, think about the ‘value’ you might be offering the paying customers and also the taxpayers for their huge investment in climate research.
Sorry, but most of the time us mere mortals do have a legitimate complaint. We see a fiscally ravenous monster that is suffering from a massive superiority complex.
Too much emphasis on engaging in what Starck called ‘an academic pissing contest’ , too much emphasis on ideologies and not nearly enough on sharing community goals or achieving practical/useful outcomes.
Who are you trying to serve? A political ideology and myopic academics? If that’s the case then there is nothing wrong with your packaging.
I put it to you that you have no use for seasonal forecasting at all or you would be able to offer some consideration of what it is you would like to see? Or if you are a customer you’re not serious – the old “I only use their forecast when its right?!?” go figure. Whatever that evens means.
Who cares about the value proposition – you’re not a customer. All hat – no cattle.
Zero to do with politics.
There you go again Luke,
Merrily dictating away.
Of course I have a use for seasonal forecasting, I am a farmer. Of course I am therefore a customer.
When have I ever said otherwise?
It is disappointing that you don’t think ‘value’ is important.
Also amusing that you think politics has zero influence.
I wish that was true.
That southern low is now centered on Captains Flat
http://www.bom.gov.au/products/IDR402.loop.shtml#skip
Luke, the whole country’s businesses [and farms, soon] will be trying to measure an unmeasurable gas in order to stay out of gaol. The bureaucrats will be trying to prove them wrong and fine them or lock them up.
They can make unfounded accusations that will be impossible to refute.
Look what the greens are currently accusing the CSG industry of with rotten-egg-gas smells along the Condamine that have occurred there naturally for years.
What a fantastically stupid occupation! How incredibly unproductive! What a great way to distract us all, in every way, from doing things that are necessary and worthwhile.
And gav, seems like you might need an aluminium boat.
What’s that you say? No aluminium? Well get out and cut down some trees and hurry up and build one.
But you can’t tell us how Debs – or string a sentence together.
SD; My seriously tall trees were cut down last week. Today we could be grinding the stumps except for the wind.
Last month was the coldest ever and this could be the windiest but how nice it is to just sit inside and watch the radar as this intense cell moves on.
It’s hard to guess the degree of CO2 enrichment of BOM’s predictions lately but back a couple of decades or so ago it would probably lie somewhere between the gospel according to St. Hansen and a weighted piece of string.
Good work Jennifer, and also looking good on the telly. 😉 🙂
Luke,
While I agree that sometimes producers are not necessarily articulate when attempting to explain what they need, I put it to you that bureaucrats and academics have lost the motivation to listen and, unfortunately and even more importantly, the motivation to even consider trying to ask the right questions.
They appear to have decided that they don’t even need to bother and it isn’t important anyway.
Your flippant comment about value is a case in point (IMHO)
I offered you a genuine comment about the behaviour and the focus and a possible reason for the disconnect….and your reply was ‘who cares’.
At least Jen and Abbot have understood the importance of clearly outlining the purpose and the possible uses of their work.
That is a huge improvement on the general attitude in curent climate research (IMHO)
Maybe you could learn something from that?
There’s nothing much wrong with my ability to construct sentences BTW…It’s actually an ability that I’m fairly comfortable with.
gav, I just hope you sequestered all that timber.
That’s what boats are for. My sheds are littered with wooden boats.
And as for grinding stumps! What a carbon footprint!
Here I was, thinking you were a greenie.
Don’t tell me I could possibly have the high moral real estate on you.
Gav, this “extreme” and “unprecedented” cold is everywhere, it seems. Shell Alaska have had to postpone their naughty drilling operations in the Beaufort Sea. Seems there’s lots of unusually robust sea-ice along Alaska’s north coast, and it’s even choking the Chukchi and the Bering further south. The drills aren’t spinning, but NWS spokespeople are spinning like crazy, of course.
As a bush-dwelling creationist I naturally blame gay marriage and Sputnik, but those educated city folk who drink frothy coffee tell me it’s all due to global warming.
Which globe would that be?
Debbie – what is striking – never a mention of http://www.managingclimate.gov.au/research/
I asked you simply would like forecasts in terms of terciles or absolute amounts or ?? and how would you like risk of being wrong expressed. How would you use in a cropping decision. I am really stunned that I can get no takeup.
In fact mega-disappointed on the whole lack of discussion on forecasting on the whole thread – we just seem to be locked in an endless cycle of warmist bashing and CO2 fixation and carbon tax – Jen hasn’t engaged except to say I don’t know anything about neural nets which really isn’t relevant. I honestly thought I’d made some good points. As had Bazza.
I learnt something – science is not communicated despite decades of effort. And most of us are clueless really. I suppose we do have a clue about is $23 tonne. Hate Labor. Hate greenies.
So despondent really. Anyway last time I mention.
Back to mindless but skillful sledging I guess.
Our little cold cell has just intensified for Sydney according to Terry Hills radar and my point is we can all take cover in time.
In my opinion the BoM radar network is the best weather tool to enhance this digital age. Cell by cell observation in this manner beats all other devices available to your average punter
Luke,
I have never said that climate research is of no use….quite the opposite in fact.
I have questioned and I still seriously question the use of the projective modelling especially in relation to AGW and the obsession with C02 etc.
In the meantime however, there has been a very impressive improvement in the collection of reliable historical data…..that is a good thing.
There will be no one happier than us farmers when projective work becomes more reliable.
If Jen and Abbot can contribute to that they have my congratulations.
You may want to ask yourself why there is that endless cycle of warmist bashing, C02 fixation and the carbon tax.
Gavin,
I love the BoM radar as well….but it isn’t relevant to the seasonal projective research that Jen and Abbot are conducting.
Deb; given the info in Luke’s link I must confess to being more excited by the extremes such as we have now as this nasty little low moves up the SE Aus coastline. Gippsland, Alps, Canberra, Sydney then Newcastle are most likely in the records again from today.
Luke; the odd behavior needs to be seen and I say it’s all driven by much larger paterns that push these elements around now. I watched this lot earlier.
http://www.abc.net.au/news/2012-06-05/thousands-without-power-rivers-rising/4053706
If the impossible were ever to happen, and actual weather conformed to a theoretical average, our Green Betters would insist that the lack of extremes was due to human-generated CO2. Gav could watch the BOM radar by the hour and report to us on the sinister monotony of the climate.
No reactions to my concerns on the neural net forecasts and this could be the last post on this topic. As a genuine sceptic, I don’t share the excitement of some of the faithful, and I have actually read the paper. I had (June 2 11:41 am) predicted that “Jens gambit claims like “shift an accepted scientific paradigm” (AGW) and “natural rainfall cycles have not been perturbed by the current elevated levels of carbon dioxide” sound like a tangled web of denialist spin and weave – nothing to do with the actual content of an evidence-based paper that is hopefully non-ideological.” My prediction was right. Turning now to what was actually in the paper, as I said there is an extraordinary correlation between the errors for POAMA and Neural Net. But they appear to be totally different approaches. It is like having two sets of scales that are inaccurate in their own way and discovering their errors are highly correlated. My guess is that the results should be deseasonalised. The paper claims that showing the error terms for each station is the best way to display, and not coloured maps. Users happen to like coloured maps which show for example the big variation in accuracy by month. Despite the hype, the paper states it has “skill at forecasting rainfall with a 3-month lead at least as good as POAMA”. But as Luke pointed out, it was an outdated POAMA. And it is not as good as SPOTA. Other claims include it “can improve the synthesis of knowledge” even disputing the experts (Zwiers and Von Storch,). But it didn’t. It used various measures of the Southern Oscillation which has been known about for over a century. Finally, the neural net was trained on data to 1993 and then validated on data to 2009. The comparison with POAMA was over the validation period. I naively expect the validation exercise to be independent of the training period. But the paper states there was “continuous monitoring of the errors in both the training set and the validation set during training”. That begs the question of how good are the forecasts in the training period compared with a simple benchmark based say on SOI. I will wait and see what the experts say.
Well baz, this expert sez it’s not the future forecasts that are the problem. It’s the past forecasts that keep changing that have us all worried.
“The future is certain, it is only the past that is in doubt.”
Even on the news today, the ABC was sure that we had never had past weather forecasts like today’s.
But is that because the GCMs are better at embroidery whereas the old forecasts were ho hum?
http://www.abc.net.au/worldtoday/content/2012/s3518426.htm
Gavin,
what’s wrong with your memory?
This event, although wild, is not that unusual.
And Bazza and Luke,
what’s wrong with adding NN work to what we already know?
Between them, it’s possible we may get more answers or at the least, as Jen says, capable of less error.
The data base gets better every year. Hopefully it will help to unlock more answers.
Keeping the politics out of it would seem to assist in that endeavour as well.
Some genuine humilty re the shotcomings of projective modelling would also help.
Not if it doesn’t improve things – do you think it hasn’t already been tried? But why do you care – you won’t be using anyway
Luke,
I believe it has been tried but not a great deal has been done here?
The focus is somewhat different and we now also have a more reliable historical data base, especially in the eastern states.
I do care and I do refer to forecasts, several of them, but particularly BoM and YR.
Maybe you and I have different definitions of ‘use’?
I see projective modelling as a useful tool, not only for weather/climate forecasting. I don’t see it as a magic crystal ball or as ‘settled’ or as a good mechanism to dictate policy. It is essentially just statistics.
Debs points included 1 what’s wrong with adding Neural Net work to what we already know?.2 Keeping the politics out of it would seem to assist in that endeavour as well.3 Some genuine humilty re the shotcomings (sic) of projective modelling would also help.
As for adding, so far it is perhaps subtracting if you read what Jen said it is at least as good as POAMA ( but an inferior version) and not as good as SPOTA. As a true sceptic, I said I would wait and see.
Keeping politics out was presumably you having a tangential shot at Jen for linking a paper on seasonal forecasting to AGW?.
As for humility re projective modelling (also aimed at Jen?), humility is how our brains work when we can stop intuition running riot. (Intuition is often useful of course but it has scary biases). Have you ever given a non-stereotypical serious example of the failure of projective modelling or how you could risk manage without it. I happen to think “half a loaf is better than none”. If you want to hang around and wait for certainty, you would not be a farmer. Why do you bother with insuring anything if you cant work out the odds.? My premiums have gone up because reinsurers think AGW has increased uncertainty about the risks they face. Have yours?
Bazza,
which part of ‘I think projective modelling is a useful tool’ did you miss?
It is not only used for weather/climate forecasting in case you didn’t know. Buf you must know because you tagged the fact that insurance companies use it too.
I also see nothing wrong with looking outside the box which I think Jen and Abbot have done to a certain extent.
If I didn’t think having an open mind was a good idea, I certainly wouldn’t be farming.
And yes, no question, I was having a shot at the political agenda that has hijacked much of the projective work and used it inappropriately. I thought that was quite obvious, but thanks for re iterating it.
It was not all that long ago that you claimed taxing carbon was ‘the only card to play’.
Let’s be totally clear here: I fundamentally disagree. Risk management can and should be focused on much more sensible goals.
two bob each way Deb. Some facts on seasonal forecasting best practice. I found the following from the BOM AR 10/11 informative. “Rainfall forecasts assessed alone over the whole year attained their highest skill level in 2010–11 since seasonal outlooks commenced in the 1990s, with a ‘per cent correct’ statistic over Australia of 63%;
The seasonal outlook for rainfall issued for November-January 2010–11 was the most skilful individual seasonal outlook issued by the Bureau of Meteorology since at least 2000 as measured by the ‘per cent correct’ statistic at 90%.
So when skill is about 2 out of 3 on the money ( ie based on 63% “correct”) using a squared error term is next to useless if you want to meaningfully compare two seasonal forecasts ( might be OK for weather where actual amounts get forecast). The good bits as in 2010/11 get diluted. You need to compare probabilities and show the complex non-intuitive seasonal and spatial pattern as has been standard in Australia for a couple of decades. I read where some old timer said “those El Niño forecasts only work half the time half the year in half the country”. His mate said “sounds like my son!”
luke’s cri de coeur ;
“In fact mega-disappointed on the whole lack of discussion on forecasting on the whole thread –”
Duscuss these:
http://www.bom.gov.au/climate/ahead/archive/rainfall/20100824.shtml
http://www.bom.gov.au/climate/ahead/archive/rainfall/20100923.shtml
This is interesting:
http://reg.bom.gov.au/cgi-bin/climate/change/extremes/timeseries.cgi?graph=R_30&ave_yr=A
On another thread I linked to a Warwqick Hughes analysis of BoM’s failed ‘forecasts’ and luke advised I should read comment 19, so I did; it says:
“19.David Brewer Says:
May 19th, 2011 at 5:49 am
Apparently there is a strand of probability theory (Richard von Mises?) that says that probabilities are meaningless for single events. After looking at these comparisons, and considering the BoM’s qualification on their outlooks, as related by AS, I can see the sense in this approach. Sure enough, a BoM estimated probability is not disproven by the eventual result having been given only a 25% chance. However, that also means that the probabilities are never right and never wrong. Even a 99% estimated probability of hot weather is not “wrong” if it turns out to be cold – it’s just that the 1% happened to get up this time.
If we leave theory aside, and think only of practicalities, the situation looks different. There is only going to be one reality to test the BoM’s outlook against. If the outlook says more than 50% per cent chance of warmer than usual, then the only practical decision one can expect from those sturdy individuals, our farmers and graziers, is to plan on it being warmer than usual. If it turns out to be cooler than usual, then the outlook was a dud tip, and said cockies have a legitimate gripe. If they keep on having legitimate gripes close to 50% of the time, as seems to be the case, then the BoM should give up publishing this stuff until their strike rate improves.”
Discuss away luke.
“My premiums have gone up because reinsurers think AGW has increased uncertainty about the risks they face. Have yours?”
Yes they have bazza, but only a fool would conclude that it is because reinsurers think AGW has increased; a more probable interpretataion is that the reinsurers think that the community believes AGW is happening and can therefore use AGW as an excuse to increase insurance.
Jen should go for it. The reason we need competing methods and alternatives is because of the grave doubts in the tech elites engendered by CAGW. As I said, nobody actually believes in the danger of human-generated CO2, and that is easy to demonstrate. CAGW is a pet theory of neo-authoritarians and an emotional/neurotic response – but hardly a belief.
The problem lies not in the alarmists’ tested intelligence or ability to calculate. It is their ability to think that is in question. (A minor example: people who think also value their thoughts enough to express them with clarity.)
Deb “whats wrong” etc.
I waited a while before commenting, then drifted on while thinking why this NN thing was conceived. After the Wivenhoe Dan issue was introduced I considered that was fair game but Bazza’s comments have returned us to the NN versus BoM what ever.
After calibrating many systems for this and that, I suggest it’s important to not only know the original goal but how to test a scheme’s success in achieving that end. Are we trying to clobber a climate model or are we trying to find justification for a vendetta against water managers? Either way we must know the extremes of weather probabilities to find the right sensitivity for analysis. This is why I hammered events yesterday.
Initial setup requires a middle point then probe the extremes to give the thing a useful range, Dam builders and dam operators need both short and long term views, so do designers.
I spent a long time living on tank systems too, overflowing gutters aren’t much use there either. Gouging out springs and soaks under hillsides in order to catch enough to pump up during the inevitable dry season helps a lot with climate considerations beyond the current weather pattern. Water running above a mudstone full of fossils was about the best I got for a permanent supply.
Cohers, back on message and further to “If they keep on having legitimate gripes close to 50% of the time, as seems to be the case, then the BoM should give up publishing this stuff until their strike rate improves.” would not that apply to Jens neural net approach given BOMs statistical forecasts are probably similar to POAMA.? Maybe you haven’t got around to reading her paper.?
Given lawyers earn a crust exploiting doubt, I am suprised you think forecast users cant handle the downside. Even mug punters bet each way sometimes or lay off or make a book against the bookie, or dont put all the eggs in one basket. It is just nonsense to think when someone is confronted with odds of say 7 out of 10 that they ignore the impact of the other 3 in 10. What is pointless is looking at 7 out of 10 on its own with out considering gains and losses of the two outcomes.
I commented on Wivenhoe because it is an example of how policy based on ‘forecasting’ can impact on communities.
“Back on message” 2 issues are being conflated here; one is the probability connected to any forecast, which is unavoidable given the stochastic nature of the weather and climate. The other is the parameters which are included in the modelling which is used to generated those forecasts.
This is the true difference between Jennifer’s method and the BOM’s modelling; BOM incorporates CO2/AGW; Jennifer’s neural network extrapolates from patterns found in the rainfall observations within non-CO2 parameters. Since you’ve read Jennifer’s paper you know her modelling has a much better success than POAMA 1.5. The comparitive results showed that “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.”
What have I missed?
Well Cohenite – if you did a course on use of seasonal forecasting and you were told that 70% of the years were greater than median . Well 30% won’t be. And minority odds can occur first time you throw the dice. If you think the 70% number will get up all the time stop looking at the forecasts and never look again. So the issue is can you make any money/advantage – certainly if it was a casino and the dice were loaded like that you would over many throws make money. But a “throw” in the seasonal forecast sense is a winter or summer season.
The way agriculturalists impatient to know whether any “forecast system” – just treat it as a black box is any use is to have a crop model, pasture model, stream flow model and see what would happen for the 120 years or more of rainfall data we have if you had used the forecast to make decisions throughout history – how many wins, loses and mega-wipeouts. As good as you can do (as droughts and floods occur whatever) is “perfect knowledge” where you make the right decision knowing what history foretold.
So you have to use the forecast system with a farming model to really know.
Now one more issue – just because the forecast expressed probabilities like 70% doesn’t mean it has skill in a mathematical sense. If the system of choice produces a set of analogue years as the outcome and there are a small number of years in that group you need to know if the forecast has skill or is statistically significant (AS WELL). Skill scores are LEPS, ROCS – but stats test like Kolmogorov–Smirnov is a non-parametric test to see any probability distribution is different from another or Kruskal-Wallis to see if any of pdfs are different. Ask Stockwell – he’ll sort ya. http://www.adv-geosci.net/6/211/2006/adgeo-6-211-2006.pdf
AND – I think producers need more than probability of exceeding median. They need probability of being in terciles (quintiles if stats is good enough) – terciles being wettest, middle, or lowest range.
You would also expect all this to vary with lead time, geographic location, and time of year. And you need to know that. Many systems will struggle in autumn as ENSO fades, flips, or restarts.
If the forecast has no skill – probably climatology (average) is about all you can do.
So how many do you reckon have an inkling about this?
Hughes is just having fun with cheap shots on an easy target.
So I’d ask anyone with a system (stats, NN, solar, lunar, arthritis, GCM) – give me your run against the historical record forecasting terciles. And tell me how many you get right and wrong (well spot on, out by one class, or total wipeout). The user can take it from there.
Now isn’t it funny how often this simple information is missing….
Discuss Cohenite.
http://www.naa.gov.au/collection/publications/papers-and-podcasts/prominent-people/inigo-jones.aspx
I wanted to say something like that Cohenite, but it got so circular I gave up!
Did you get that bit about insurance Bazza?
I also liked that comment 19.
It calls for some much needed humility and a recognition that climate research, although getting better, is not as an important decision making tool that some seem to keep saying.
“Ask Stockwell – he’ll sort ya.”
David is too busy making money.
In respect of your linked paper and cumulative distribution functions (CDF), such methods are meaningless unless you factor in the weighting effect on climate variables which occurs in natural oscillations. This has been Stewart Franks’ point; it is meaningless talking about 100 year floods and such extreme events unless you appreciate that in PDO oscillation such things as floods cannot be measured by such terms or CDFs, because it is the oscillation itself which affects the probability of a flood or extreme event occurring. That is, floods and extreme rainfall events are more likely during a -ve PDO phase.
Conversely, if your model and approach is based on the assumed climate sensitivity of CO2/AGW, your model will have increasingly reality-divergent forecast results.
Moso bamboo being what it is, with a mature grove I’m still dependent on rainfall, but no longer dependent on rainfall prediction. So I’m more curious than engaged these days. Nonethless, as well as radar, satellite etc, two indicators that have helped and impressed me are PDO and ENSO. (Before anyone asks me if I’m grateful to the various researchers and fisheries experts for these advances, the answer is yes.)
What has amazed me is the fundamental change in our eastern weather since 2007, particularly as regards moisture and winds in winter/spring. I’ll go so far as to call it climatic change. The amazement stems not from the change itself, but the lack of interest in something so defined and dramatic. There was an El Nino in 2009, but it was so different to those of the eighties and nineties that it really accelerated my interest in PDO. They say it “flipped” in 2007, before a minor and temporary reversal, and, while I wouldn’t pretend to know about something so vast, the winds and climate around here would seem to agree.
I don’t regard such indicators as exact and/or predictable mechanisms, nor do I believe in neat phases. The Big Wet of the late eighties, the drought and heat around 1960, for example, weren’t typical of the prevailing conditions before and after. Nonetheless, there are vaguely observable patterns for people standing in paddocks, and if science can sharpen the image, let’s go for it. (I should add that it is the observed PDO for the past century that impresses me mightily. The reconstructed PDO for the last millennium leaves me bewildered.)
The most important thing is to recognise the enormous volume of stored carbon – aka regrowth – which will again become combustible when the pattern changes and dry westerlies again hammer the bush in August-September. As for other regions, I’ve never lived in crown-fire country, but heaven help the south of our continent if we forget – as even farmers and hydrologists did in the 1970s – that it’s just when you get used to floods that the fire returns.
And if people like Abbott and Marohasy want to have a low-cost go at predicting, all I can say is that IBM never felt threatened by spotty teenagers working out of West Coast garages.
Not at all in the slightest – this is all basic maths. The mechanisms don’t count.
The forecast can be your grannies arthritis prediction method. So if the PDO counts adding it to an ENSO alone system should give an improvement in the mathematics of evaluation – e.g. SPOTA.
Four issues –
snake oil – being aware of systems with false skill, overfitting, lack of independent validation, bad logic, sounding too good to be true;
how to evaluate a forecast’s worth;
understanding of probability vs categorical; and
how a user needs to have forecast results conveyed to be useful.
Cohers re at 1:05pm your “Conversely, if your model and approach is based on the assumed climate sensitivity of CO2/AGW, your model will have increasingly reality-divergent forecast results.” I thought sensitivity was an emergent property of a coupled model. And anyway the models simply show recent divergence is not unusual in the context of previous noise around the trend related to ENSO etc. So what are you on about? What is your estimate of sensitivity and where did it emerge from and why?
Funny too that the neural net approach used Sydney temperature which has I think negligible trend compared with much of Qld. In any case, given seasonal forecasts are low skill, they would be the last place you would look to detect an AGW trend in skill. Not because it is not there – it would just be a low powered test.
“I thought sensitivity was an emergent property of a coupled model.”
Are you talking about AGW or luke and his girlfriend?
Actually, and luke may be able to help here, the AGW spruikers don’t even know whether CS is an intensive or extensive property of the climate system.
Well Cohers is doing an inverse Debs – (Small text volume but trying his best not to engage on topic).
CS is an emergent property of both extensive and intensive variables.
ANYWAY – I have give Mr Bob Fernley-Jones a science lesson so excuse me for a minute.
Well Bobs – your graphs of sophistry have been published here only about 200 times but not as much as Dorothea Mackellar.
The answer is you need to understand what the AGW science is saying with respect to the MDB (not Tim Flannery or Andrew Bolt). The MDB is a big place. You might think it’s a long way from Dubbo to Goondiwindi but the MDB is much bigger.
So averaging over that distance and range of climate systems is a bit silly.
http://www.seaci.org/ has the ducks guts on this anomaly (Figs 1 and 4 here) – http://www.bom.gov.au/climate/current/statements/scs16.pdf
The research does not say it will never rain again. It does however point to a long term trend to changes in atmospheric circulation – which changes rainfall tracks. So the science says (and its complicated but worth reading) that this persistent change will still be there post La Nina.
To do with a centennial change in the intensity of the subtropical ridge and also southern annular mode. It’s southern Australia not the whole MDB.
Now there are also interesting things in SW WA, coastal water temperatures, ENSO and the Walker circulation but let’s not get excited.
When you understand the SEACI work report back.
Now back to forecasting PLEASE
“Now back to forecasting PLEASE”
Sure and since you raised David Stockwell before:
http://landshape.org/enm/files/2010/10/Critique-of-DECR-EE.pdf
Cohers, you keep doing a bolt from one courtroom to the next. I don’t know whether you are closer to “artful dodger or great pretender”. I did raise a few queries about the neural net forecasts and all you do is dodge off on tangents. I respect your being across some of this stuff enough to assume at least you would respond to some of the points of concern I raised about Jens paper. Some were internal to the paper, some were external. My prediction was correct. For example “I thought the paper was about seasonal forecasting. My prediction is that Jens gambit claims like “shift an accepted scientific paradigm” (AGW) and “natural rainfall cycles have not been perturbed by the current elevated levels of carbon dioxide” sound like a tangled web of denialist spin and weave – nothing to do with the actual content of an evidence-based paper that is hopefully non-ideological “ I was right I learnt when I read the paper as you have apparently. Your comment?.
And another. “But I am curious as to how come there is an extraordinary correlation between the errors for POAMA and Neural Net when they appear to be totally different approaches. In fact you could predict NN station errors with 95% accuracy using POAMA and vice versa which makes one of them redundant”. My hypothesis was that the method used was swamped and confounded by seasonality. Over to you for a considered comment.?
As usual I Googled a few key words above including “insurance, global warming,” etc and came across this doc.
http://www.nwc.gov.au/__data/assets/pdf_file/0016/22129/Water-policy-and-climate-change-in-Australia-full.pdf
While reading, a chance phone call re Rio and work within then something unexpected from my next search, an introduction to SEEA.
http://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/4628.0.55.001main+features10May+2012
Btw; “Forecast the Facts” had links on dropouts at Heartland during May.
We do talk a lot about Dorothea and her poem, and the Fed drought, don’t we? So, how were things before all that?
“Captain Matthew Flinders reports drought and bush fires from 1782 to 1792. There was a great drought in 1797 for 100 miles round where Melbourne now stands ; 1799 to 1806 were very wet years, and in 1806 the floods culminated by a rise of 101 ft. at Windsor, on the Hawkesbury River….
The excessive rain kept on till 1810, but 1811 cut it short, and was so dry that water was worth 8d. per bucketful in Sydney. This drought was sharp but short, and there was plenty of increasing rain for years afterwards, till in 1820 the Hunter River rose 37ft.
Ordinary weather followed till 1837, but 1838 and 1839 saw the champion drought of the century. Stock were all but exterminated. The Murrumbidgee is a great river, 150ft. wide, 60ft. deep, and overflows its banks, like the Nile, when the head snows melt, for five miles on each side to a depth of 3ft. This gives a volume of water equal to a river of 1450 ft. wide and 120 ft. deep, and besides this it fills a group of lakes each from seven to eighteen miles in diameter.
Yet this great river dried up so thoroughly in 1839 that the fish died and putrefied at the bottom of it.”
The Brisbane Courier 1889.
Er, did he say 101 ft at Windsor?
Of course, one could go on. About the parching of the inland rivers, leading up to Black Thursday of 1851, a couple of years before Gundagai was swept away, with mass drownings. One could go on and on.
So I have no trouble in believing in “a long term trend to changes in atmospheric circulation – which changes rainfall tracks”. I assure you, I have no trouble whatsoever in believing that.
Bazza – Cohenite is pretty good and very well read. But he won’t engage if he doesn’t understand to keep the pretenses of invincibility. Pity as a useful exchange would always improve understanding. Of course unless you’re not really interested in things climatic and simply want to join a “revolution”. Vis a vis some of us I suppose figuring this stuff is actually important to those in the field.
Part of that engagement of course would be actually spending some time with BoM and CSIRO instead of assuming they’re some sort of evil empire.
And really the level that BoM are planning with their forecasting technology is light years ahead of the discussion here.
What is very disappointing is our inability to get Debs to give one example of how she would use a seasonal forecast and how she would like that information conveyed. Do you think she feel we’re trying to trap her or that she’s not really interested.
So Robert given that excellent account of our climate history how do you think science could attribute changes between natural variability and anthropogenic influences. Just asking.
A book of interest http://www.amazon.com/Change-Weather-Climate-Culture-Australia/dp/1876944285
I said I was busy but just to show Tweedledee and Dum that I’m still engaged, a couple of queries:
““But I am curious as to how come there is an extraordinary correlation between the errors for POAMA and Neural Net when they appear to be totally different approaches. In fact you could predict NN station errors with 95% accuracy using POAMA and vice versa which makes one of them redundant”.”
Which system had the most errors?
“My hypothesis was that the method used was swamped and confounded by seasonality.”
As luke knows, PDO is the sum of ENSO, a seasonal indice; so what I said about PDO.
“…how do you think science could attribute changes between natural variability and anthropogenic influences. Just asking.”
Luke, it’s of no great concern to me. Both are obviously real. The one is enormous, the other trivial. But I would say that, wouldn’t I?
The nineteenth century climate in Oz seemed to be pretty hectic without our contributions. Interestingly, the local aborigines said that the disastrous 1841 flood was a pussycat compared to a previous deluge in their living memory. Sounds grim. Me, I’d have hated to be around the last two years of that Fed drought. Everyone but Qld got little breaks from the Fed drought…till 1901-02, probably the driest warm season ever recorded. The prob with the Fed was that it was preceded by excellent rains, but followed – in spite of the usual Aussie flood disaster – by inadequate rain in eastern Oz till after WW2! Like you say, Luke, “a long term trend to changes in atmospheric circulation – which changes rainfall tracks”. Amen!
Bugger of a country, don’t you think?
Robert; although my last post hopefully had something for Deb in particular, have a read cause we are entering a new phase of accountability across a spectrum of issues that will include the impacts of natural events such as the current floods in Gippsland.
Have a look at the organisations involved too.
Have already read it Gavin,
what was your point?
The NWC is yet another bureaucratic organisation that regularly and spectacularly fails to adequately consult with and listen to the people who actually pay the bills.
When they do come and speak to us they actually only ‘present’ and ‘self justify’.
Victorian floods, Gav?
In winter 1909, after the Federation Drought and before several other droughts, rivers in Western Vic rose, many to their highest recorded levels, causing all the the usual disasters…then they promptly receded so drought could resume, in time for the warm season. (As Luke put it so well: “a long term trend to changes in atmospheric circulation – which changes rainfall tracks”.)
Fortunately, that state has lots of brown coal for powering an energy-hungry desalination plant, needed because, so they say, the climate has changed because of…well, the burning of brown coal, among other things!
Never mind. In this age of accountability we have found a way to accountably fund our remedies to this climate change. We will replace coal-gobbling facilities by letting them get so old and inefficient that pixies will build something else to generate power. We fund it all by carbon taxes and exporting over 75% over our massive production of….guess what!
There are rumours of a new pixie bureaucracy in Vic which has a plan to hire thousands of worker-pixies to lift an entire desalination plant and dump it in the Mitchell River. In the interests of accountability and pixies, this will not be called a dam but a storage.
“The NWC is yet another bureaucratic organisation that regularly and spectacularly fails to adequately consult with and listen to the people who actually pay the bills.”. Fair go Deb, what about the headworks? You could at least say thanks.
Bazza,
No one says that good work isn’t done.
I think however that the whole system of water management is in some serious need of streamlining.
On last count there were 8 State and 6 Fed bureaucracies involved in water management in my patch.
Often they get in each other’s road and often the policies are completely contradictory.
No one is ever accountable for the stuff ups.
This NWC report was commisioned by the politics and was aimed at justification, not at assisting the end users or at achieving some much needed common sense or even identifying goals that we could help them achieve.
It is completely out of whack at an administrative level.
We need clear accountabilities and strong feedback loops and less administrative justifications.
Administrative costs are spiralling out of control for little or no return benefit or value socially, economically or environmentally.
Goes back to my comment to Luke re thinking about ‘value’.
The rhetoric has no matching policy and no matching performance.
Taxpayers and customers have not been provided with a clear mechanism to judge performance.
Instead we’re told we should just ‘trust’ the ‘science’.
So Bazza, that tends to overwhelmingly negate some of the good work that has been done.
The obvious superiority complex that you often display is not helping to improve this problem.
You may be awesome at statistics and modelling, but they are NOT the answer to everything and they are NOT proving to be a good method to dictate social policy.
Quite clearly they are far too vulnerable to abuse, just as they have always been.
The added power from technology doesn’t change the fact that it’s still just statistics.
It’s also not just used by science.
Listen to one session of parliament (if you can tolerate the childish behaviour) and observe how often exactly the same sets of data are tortured by statistical modelling in just about every contentious issue you can think of.
Deb; when I give you a link, you should know by now to look much deeper and stay with in step with the band on the march forward.
Meanwhile still no discussion of the science ….. as the thread grinds to a halt.
“Meanwhile still no discussion of the science”
I asked 2 questions above in response to bazza’s complaint about one of the 2 approaches being redundant; the point is not that POAMA and the NN can predict each others errors to a 95% certainty but which one has the most errors; for instance being able to predict a 5% error rate to a 95% certainty is different to being able to predict an error rate of 50% to a 95% certainty.
Take that one before you look at his other complaint about Jennifer’s conclusions being “swamped” by seasonality.
Well this is where you can tie yourself up in knots. Instead of RMSE which I’d agree is a measure of how well does the actual value compare with the model estimate – does this inform you on useful real world evaluation. So to this end I think forecasters should predict terciles of rainfall (wettest, middle, lowest) or maybe quintiles. (wettest, wetter, average, drier, driest)
If an RMSE is a bit different in comparison does this mean much in terms of your decision making with the forecast – I can’t tell really.
Then tell us how often and the time series of prediction vs actual rainfall terciles for a set of validation data. So this would give you a real world sense of predicting wet, dry, average
On seasonality Bazza has a lurking concern that part of the RMSE skill is tuning to a seasonal summer dominated rainfall cycle not variability. Which is a known unknown.
So again my theme is forecast evaluation and how users need information communicated.
Not getting the shoulding comment Gavin.
Marching forward to where?
Luke,
considering there is a large error component, there isn’t much more that you can do for users until there is a capability of ‘less error’. If that’s what Jen &Abbot are working towards, then no argument from me.
The obsession with C02 is another issue.. . which I recognise you’re trying to point out.
Unfortunately the exponential increase of reports is not recognising that at all. Last report commisioned from our CCs being a classic case in point.
The NWC report linked by Gavin is another.
The most recent CSIRO report commisioned by the MDBA is another.
There are many, many, many more of them.
Maybe they should start listening to you? You appear amenable to seperating the issues at present.
It would perhaps help to allocate more funding towards the ‘useful’ stuff and we could all work together on some achievable risk management goals.
Debbie – given your avoidance of any discussion on actual use of forecasting – you won’t be any wiser. Would you like to buy this bridge I have in stock?
Perhaps useful stuff may already have been considered (ignored by raging sceptics of course) http://www.managingclimate.gov.au/research/ – hmmmm I wonder what the portfolio looks like.
Bazza – I was just looking up what “an overly emphatic forecast” was. Is it a problem?
and “ensemble hit rate”
“The frequency distribution of forecast probabilities (shown inset in the diagrams) shows
whether the forecasts have a tendency to be emphatic (indicated by large numbers of very high or very low probabilities) or equivocal (large numbers of near 50% forecasts). There is no
’correct’ frequency distribution; instead this information aids the interpretation of the reliability
diagram.” http://www.cawcr.gov.au/publications/technicalreports/CTR_040.pdf
and found this in the into – LOL – The poor reliability of seasonal forecasts based on dynamical coupled models is a barrier to their adoption as official outlooks by the Bureau of Meteorology.
So your strawman issue is spot on. Which is why they’ve been beavering away on 2.4
LOL !
Cohers , you disappoint. Your points on %s this and that are pointless. I thought you would be perplexed at the extraordinary high correlation between two apparently different ways of doinng the same thing. Anyway it is an old POAMA and SPOTA is 20% better. Try this one in case you dodged it. “The neural net was trained on data to 1993 and then validated on data to 2009. The comparison with POAMA was over the validation period. I naively expect the validation exercise to be independent of the training period. But the paper states there was “continuous monitoring of the errors in both the training set and the validation set during training”?
“the extraordinary high correlation between two apparently different ways of doinng the same thing.”
The point of my %s was to indicate that the correlation you thought was present was not in fact present. What is it about the Neural Network, to be precise Time Delay Recurrent Neural Network, having a lower RMSE than POAMA that you don’t understand?
Why and how is SPOTA 20% better, and compared with what?
The validation period is independent of the training period; but there has to be a certain time overlap to minimise “overtraining”; read the Method portion of the paper.
Will you answer why you think the NN was “swamped” by seasonality?
Well, we always knew they did this [and not just noaa]:
http://wattsupwiththat.com/2012/06/06/noaas-national-climatic-data-center-caught-cooling-the-past-modern-processed-records-dont-match-paper-records/
Who cares – we’re in a serious discussion.
“Who cares – we’re in a serious discussion.”
You mean as in your intentions as opposed to what really happens.
“You mean as in your intentions as opposed to what really happens.”
I’ve been wondering the same thing SD.
As I said Luke,
rhetoric is not matching performance.
It’s my fault Debbie, I disappointed bazza.
I rest my case Cohers. It needs a good old conspiracy theory beloved of the right to explain the decline in useful contributors but not contributions/contributor, as we drift to a double century with a little heat and a fading light. Remember when Douglas Macarthur got sacked. The story is roughly this: he told the US senate “old soldiers never die, they just fade away. “ HL Mencken reported “ there was not a dry eye on the republican side, nor a dry seat on the Democrats.” Red of eye and red of neck, I say. Back to conspiracy message. Deb is clearly a double agent – an infiltrator driving away potential contributors in droves. Well done Deb, pat yourself on the back. What can be said of Luke. Maybe he is moonlighting (now there is a driver) employed by the operator to add some scientific respectability to the debates, if they could be called that. Also serves as a bit of a dunk tank – nothing like a right wing think tank. Score a hit on him (fluke the Luke) even from a cheap shot and you get to dunk him. Such gratitude for injecting a bit of hard evidence from mainstream science into the non-debates. Then there is Cohers, bless him, all ready fire aim, hit and run injecting a bit of random sidestream science from up the creek. Not to mention the rent a crowds, maybe Robert and Neville, perhaps they are alter egos tossing in random fizzer grenades . And the spangled drongo, ever the empiricist has undoubtedly gone north for the winter not realising with his limited context in time and space etc that it’s a warming world. And of course Gavin from left field , gotta love him underpinning the empiricists and the trajectories of the projectile modellers with his well grounded understanding of measurement errors, a long way from the biases of the right whingers. Flashback to spangles. He also has the drum on sea level rise (albeit a sample of n=1, but that is oK here ) and he quotes Lawson so things aint all bad, “Whether the sea or the bush it be, The heart of a man prevails”. So heads up from that. I should fade away, the joke is not on me.
6 wasted comments. It’s amazing to think that some the bloggians here couldn’t themselves do a daily drive by of Watts and Novas to check the latest disinformation ruse. They’re incapable. Meanwhile Cohenite ducks and weaves and Debs still can’t think of anything.
Just read Bazza’s final analysis on the lack of analysis. Well my contract has expired too so that’s it folks.
There you have it folks; amphigory of the lowest order from as self-satisfied a pair of nitwits as is available outside of Federal parliament.
Let me finish by once again congratulating Jennifer on her paper, which amongst other things, reminds us of the parlous position official climate science, that is from the BOM and CSIRO, is in this country.
Good night and good luck.
But we would just hate you to miss any tiny bit of the agony.
Maybe all this extra CO2 is making you too sensitive?
Or just a little decoupled?
http://wattsupwiththat.com/2012/06/07/claim-todays-climate-is-more-sensitive-than-that-of-the-past/
Red-neck laughter in the bar isn’t always directed at you.
“I should fade away…”
“Well my contract has expired too so that’s it folks.”
Never mind. To paraphrase A. A. Schwarzenegger, they’ll be back!
Tracking back I found this @ IEEEexplore 2008
“Using neural network emulations of model physics in numerical model ensembles” Abstract-
In this paper the use of the neural network emulation technique, developed earlier by the authors, is investigated in application to ensembles of general circulation models used for the weather prediction and climate simulation. It is shown that the neural network emulation technique allows us: (1) to introduce fast versions of model physics (or components of model physics) that can speed up calculations of any type of ensemble up to 2 -3 times; (2) to conveniently an naturally introduce perturbations in the model physics (or a component of model physics) and to develop a fast versions of perturbed model physics (or fast perturbed components of model physics), and (3) to make the computation time for the entire ensemble (in the case of short term perturbed physics ensemble introduced in this paper) comparable with the computation time that is needed for a single model run.”
http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=4625775
and this –
“A simplified recurrent neural network for solving nonlinear variational inequalities” Abstract-
“A recurrent neural network is proposed to deal with the nonlinear variational inequalities with linear equality and nonlinear inequality constraints. By exploiting the equality constraints, the original variational inequality problem can be transformed into a simplified one with only inequality constraints. Therefore, by solving this simplified problem, the neural network architecture complexity is reduced dramatically. In addition, the proposed neural network can also be applied to the constrained optimization problems, and it is proved that the convex condition on the objective function of the optimization problem can be relaxed. Finally, the satisfactory performance of the proposed approach is demonstrated by simulation examples”
Please carry on
Cheers
Gee whiz Bazza,
it didn’t take long for you to pick up that gun again.
You definitely seem to miss the fact that my complaint is not about the science or scientists per se.
Unlike you apparently, I am a results orientated person and I judge policy by results not by political ideologies.
Climate science has been hijacked and the projective work in particular has been used inappropriately.
That has very little to do with right or left.
I should also point out that right and left are socialist terms and therefore almost irrelevant to this discussion.
Labelling ‘groups’ and claiming there are ‘sides’ is totally counter productive.
Taxpayers actually want ‘value’ and practical social policy.
Out in the real world where we are responsible for producing, people like me are tired of the political rhetorical spin and thoroughly appalled at the huge amount of money being wasted on ‘marketting’ or ‘packaging’ something that obviously is not yet capable of delivering on its promises.
The results of current ideologies that use international treaties and trip the ‘precautionary principle’ are stifling and overwhelmingly negative.
I have no problem at all with continuing work on climate/weather forecasting. If Jen and Abbot can make a positive contribution to that, then more power and more funding to them.
The increasingly impressive data banks should assist some real progress if we were actually focusing on shared goals.
I don’t know what I’ve enjoyed most…
Shallow and snobbish Lysenkoists posturing as serious researchers, flattering one another and exchanging their Dorothy Dixers? Or an arbitrary cut-and-paste from any old tech-talk that contained the words “neural network”?
No, I think it was the bits about leaving, even if – like all else they project – it doesn’t really happen.
Yep,
gotta love those Dorothy Dixers.
Burke pulled out a classic ‘dixer’ last week re the MDBP.
Soooo obviously political and soooo obviously competetive 🙂
This DD was intended for AR5 but hopefully it’s been sufficiently torpedoed.
Ya wouldn’t wanna bet though:
http://bishophill.squarespace.com/blog/2012/6/7/another-hockey-stick-broken.html
Bazza
Of course the training was independent of the forecast.
If there was anything so basic wrong with our methodology we would never have got the paper published by Advances in Atmospheric Science.
There is nothing wrong with our methodology. You are just way out of your depth and unable to acknowledge an important contribution because you believe too hard in nonsense.
It looks that way to me too,
‘unable to acknowledge’.
Why is that?
The comments from Bazza were tainted with comments that questioned the validity of the methodology and also the qualifications of Jen and Abbot.
Professional jealousy perhaps? Or maybe Jen has nailed it: believe too hard in nonsense?
Aren’t you able to give credit to other published studies Bazza?
Or perhaps recognise that maybe others have the ability to make positive contributions?
Just because people like me like the approach this work has taken, doesn’t automatically mean that we think the other research is instantly negated.
If it is useful and can be added to what we already know, then it will have support from ‘end users’ like me.
And Luke,
that was the answer to ‘how would you use it’.. . . which I answered the first time you asked.
I imagine it will be an added use, not unlike we use numerous sites together already, which are all packaged differently.
Jen, forgot to add there are a couple of errors I spotted from the shallow end – data common to tables 3, 4 and 5 don’t agree where they should.
As a friendly climate scientist told Eli when the Bunny asked about neural nets, they tell you what will happen, not why, and if anything changes, neural nets are not very useful. GCMs tend to be a lot better on the why.
Eli Rabett June 17th, 2012 at 8:27 am wrote:
“As a friendly climate scientist told Eli when the Bunny asked about neural nets, they tell you what will happen, not why, and if anything changes, neural nets are not very useful. GCMs tend to be a lot better on the why.”
Here’s the Larry Aphorism of the day:
GCMs are incontrovertible proof of the existence of computer programmers.
Dodgy assumptions are fed into the GCMs. Then the GCMs make predictions, which thus far have turned out to be total garbage. Then Warmies take the existence of the GCMs as evidence for their little secular religions. Somehow that is supposed to make GCMs “a lot better on the why.”
Eli, can you say, “confirmation bias?” Or “circular reasoning?”
Scientific Literacy 101: In the full-fledged theoretical sciences, prediction is the coin of the realm.
Oops, that’s two Larry Aphorisms in one day. Bad Larry! *slaps own wrist*