There’s already some finger-pointing about the management of Brisbane’s dams in the weeks leading up to the flood. I don’t want to deal with that while the emergency continues, but I will make a couple of suggestions regarding future policy in Brisbane and elsewhere
* The historical statistics on the frequency of severe rainfall events (both droughts and floods) have proved to be of little value. Everywhere in Australia, we need to work on the assumption that extreme events will be more common in the future than they were in the (pre-2000) past.[1]
* As regards Wivenhoe Dam, we need a much more cautious approach to flood mitigation, going into wet seasons with a substantially larger reserve capacity. This in turn will reduce Wivenhoe’s usefulness as a water supply source, and buffer against drought.
* One response that is immediately available to us is to turn on the water recycling plant, built at great expense during the drought and never used. Current policy is to turn it on when average dam levels are at 40 per cent. This trigger should be raised significantly. As a very rough guide, it appears that when our dams are at 100 per cent of normal we currently have enough storage for four years supply. If instead we cut the maximum to three years supply (75 per cent0, we could (roughly) cancel the impact on supply by turning on the recycling plant at 65 per cent (40+25)
On a personal note, I’ve had some reports from home and it appears that cleanup will go faster than expected and that our biggest loss (the car) is covered by insurance.
Many others are not so lucky. And, as was discussed in comments to the previous thread, floods in poor countries regularly cause many deaths (400 in Brazil and 40 in the Phillipines in the time I’ve been following our own flood) and wipe out the few possessions of very poor people. If you are giving to flood relief appeals for Queensland (as I hope you are), please think about also sparing a few dollars for emergency relief or long-term development aid for the truly poor.
Update It’s been argued pretty convincingly in comments here and elsewhere that the additional mitigation capacity in the estimates above would only have had a marginal influence on the current floods, given the volume of rain that fell. Still, I think the basic premise is right. We will need more mitigation capacity and that means less water at higher cost.
Another point I meant to make earlire is that more reliable forecasts of El Nino and La Nina events would be of great value in managing water supply in a highly variable environment.
fn1. I will have a lot more to say about human-caused climate change in the near future. But until I’ve had my say, I request everyone to avoid any discussion of the AGW topic in this thread – failure to heed this warning will lead to deletion and may result in a permanent ban. Feel free to continue discussion in the agnotology thread.
John Quiggin 
@Nick
You’re just wrong as the data shows. The Bremer River at Walloon was only just above the minor flood level on Saturday until 11 am and didn’t get back up to minor flood level until about 10:30 pm on Sunday night. Warrill Creek at Amberley was in minor flood the whole weekend but was always at least 0.5 m below the moderate flood level and on the way down until Sunday night.
How much flooding? I can’t get the exact figures for Ipswich city at the moment but it would have been substantially below moderate flood level. (Do you know where to get them from?) I’d also be very surprised if it had been much above the minor flood mark on Sunday until Sunday night. I’m not denying that flow at the moderate flood mark at Savages would not produce a level around the moderate flood mark in the lower Bremer. It’s just that the Bremer was not yet at moderate flood level, especially on Sunday so there was an opportunity to make use of those facts.
Interesting you should say that. They were running the river past Savages at 14.5 m for several days after the flood (1.5 m below the moderate flood mark). At this time the river at Moggill was around 8 m, 2 m lower than minor flood level, and at Jindalee on Sunday it was around 5 m, 1 m lower than minor flood. So it appears that as long as they keep it below moderate flood level at Savages, then it’s not likely to get much above the minor flood mark in the lower Brisbane until you get to the tidal zone. At least this has proven that if they had run Savages at 14.5 m (220 Gl/day) on Saturday and Sunday, then that would not have caused any significant problem for Brisbane. Of course, it also would have meant Brisbane would have had 230 Gl less floodwater dumped on it on Wednesday and Thursday. And this doesn’t include doing anything different on Friday and Monday which are a little different from the weekend. The Bremer was a bit higher on Friday but I don’t think there was 220 Gl available to be released on Friday, perhaps 100 Gl. Monday was an entirely different situation and I don’t think there was much opportunity to release a lot more water. However, they should have known that there were deluges on the way (they only had to check their gauges at Peachester, Kilcoy Ck, and Linville on Sunday afternoon) and acted accordingly, i.e. release dam water as fast as possible.
Also, I guessed that they could allow just under the moderate flood flow rate, 267 Gl/day, at Savages but it’s clear that the dam management actually prefer 220 Gl/day which they released for several days after the flood, now that they’ve learned their lesson.
Here is a good explanation of why dams are damned
@Chris O’Neill
Chris,WaterGrid press releases on the days mention the concern about exacerbating water levels on the Lockyer and Bremer.This press release is real time reaction,not some post fact airbrushing. After the fact,indeed,it looks like the least of their concerns,but they were not operating ‘after the fact’.
I stand by what I claim the data say at the three above Ipswich gauges show for 7,8,9, Jan,and you may recall I did say that that period was not as severe as the next two days became in the Bremer system.
They did start to increase Wivenhoe releases on the afternoon of Sunday 9/1,as they had started releases on the 6th and increased them slowly from that time,all the time in the belief that they were managing what they were seeing. Flows at Savages increased by 30 GL from midday to midnight 9th,and as we all know really gathered pace hour by hour each of the next 24 and beyond.
My observation about the relatability of stream gauges per flood stage was a general observation.Of course it may be easier for a short sequence,but this is not at all a rule of thumb,if you look at a useful number of sites in various basins But as you note,below moderate at Savages can be below minor at tidal Mogill or Jindalee….it can also be above moderate at Mt Crosby,and the Bremer is always a concern.
If a better front curve had shaved a few hundred gig they would have had more headroom,but it seems that the Stanley system was yet to re-peak at midnight,they had an hour to contemplate the drop off in rain rate and wonder whether that would last, and maybe that few hundred cannot be directly translateable to shaving the peak outflow as much as you or I may speculate.
@Nick
Spare me the aspersions. I’m not denying they should avoid causing a moderate flood in Ipswich. I have argued that they could have allowed much more release without causing a moderate flood in Ipswich and presented data to support that. I haven’t been able to get the actual water level in Ipswich at the time but I don’t think it was much above the minor flood mark on Saturday and it was lower on Sunday than Saturday. I have the flow data for the time and when the next minor flood comes along I’ll be able to get the relationship between flow and Ipswich and Moggill levels. Then I’ll be able to settle this question once and for all. I could ask the BOM for the Ipswich level data but I don’t feel like bothering them yet.
I have provided detailed criticism of your claim and you have not provided any detail in response. That is disappointing. The critical issue is the level the Bremer in Ipswich actually got to for which the data has been removed from the BOM website. By the way, I managed to get a cached copy of this data that began just after midnight Monday the 10th. The level then was 7.85 m at 12.30 am Tuesday and rising at 0.25 m/hour. Even at that stage when there was FAR more flow than on Saturday (62 Gl/day rate at 11 pm Monday night compared with 19 Gl/day average on Saturday), the Bremer was still not at the moderate flood level (9 m). There is no evidence in any of the data I can get that the Bremer in Ipswich was anywhere near moderate flood level on Saturday.
They were following the book and reacting slowly as they always had.
And it should now be obvious that the acceleration was far too slow during the entire period, especially before Sunday night. I’ll also point out that on the previous week on Tuesday, Wednesday and Thursday morning, the release rate was very low, even though Wivenhoe levels were 102.1%, 102.4% and 103.2% at 6.30 am on those days and reached 106.3% on Friday morning. There is no apparent reason why they had such low release rates while the dam was over the 100% level. That 100 Gl excess they had by Friday morning would have made a big difference on its own if they had not allowed it to accumulate.
Indeed but that’s not such an important part of the river. They were actually running 16.5 m at Mt Crosby, 3.5 m over moderate, during the post flood releases, so they obviously don’t now think that it’s important to avoid moderate flooding of Mt Crosby.
Indeed that appears to be the most critical part of the whole system. I believe they could have done a lot better this time but that doesn’t mean the Bremer flow couldn’t have been, say four or five times as much on Saturday. That would have been very difficult to deal with. The Bremer is the key to getting the most value out of Wivenhoe flood mitigation.
I don’t see how the forecast on Tuesday morning had anything to do with missing the opportunity to release an additional 400-500 Gl before the flood. No matter what, that release would have reduced the risk of the flood peak going so high.
and the Bremer is always a concern.
Chris O’Neill. What aspersions? Pointing out that a press release was not air-brushed after the fact? Did you take that as a personal slight? I wasn’t aiming one at you.
You and I have looked at the same data for the Bremer system . We’ll just have to disagree.
“They were reacting slowly as they always had” What a remarkable statement!. And reactive management means starting behind. They had just emptied the buffer by the 4th. The buffer absorbs inflows.There is no way that it can be known beforehand whether they will permanently detain them. If the event had tailed off some hours earlier,or the rainfalls less intense,and the reactive management had proceeded as it did, it would have been a triumph of management..again in hindsight.
Running Mt Crosby at over moderate caused the delay in Ipswich dropping back below minor flood level. After the major flood had passed and the damage done,it was now a priority to drain the flood reserve,so minor flood levels at Ipswich were the lesser problem.
However,before the flood situation had developed,and as expressed in the press releases,while they had control they chose not to exacerbate immediate prospects of tributary flooding. In the wash,this appears pointless,but at the time,before all the rain had fallen,this may have seemed worthwhile. Is that unreasonable?
Let me explain my last paragraph better. What I thought I had expressed clearly was that,even with a [your] buffer gained a little earlier,how can one assume that the choices made subsequently would have exploited that buffer perfectly,given the data at around midnight on Tuesday?
Like food policy water policy should address population, increasing population hampers flood mitigation, increases the cost of water supply and hampers adaptation.
@Nick
Sorry, that’s the way I read it.
There was no mention of the Bremer until Monday and no mention of the Lockyer on Saturday or Sunday. They mentioned these on Monday morning when the Wivenhoe and Somerset together had reached 150%. We can’t tell from their press releases what they wanted to do about the Bremer on the weekend but they were releasing a lot more on Monday when they said they wanted to let the Bremer subside than on the weekend when they said nothing about the Bremer. If they thought their releases on Monday would let the Bremer subside, why didn’t they think the same releases on the weekend would also let the Bremer subside?
It’s a pity you’re not more specific about what you disagree with. For example, do you disagree that that the Bremer level on the weekend was not much more than minor flood level. Extrapolating back from the level at 12:30 am Tuesday the 11th (7.85 m), I calculate that the Bremer in Ipswich reached minor flood level (7 m) at about 8:40 pm on Monday. The flow rate at that time in Ipswich (58 Gl/day) was THREE TIMES the average flow rate on Saturday (19 Gl/day). I find it extremely difficult to believe that the Bremer in Ipswich was even at minor flood on Saturday. Prove me wrong if you can find the data.
Having the Bremer below minor flood level on Saturday and Sunday would have allowed them to add at least 2 metres at Moggill and still keep the Bremer below moderate flood. This would have allowed a huge increase in releases.
More later…
I’ve just checked the BOM for the Bremer at Ipswich and yesterday, Thursday the 20th, it was up a bit at 4.45 m but nowhere near minor flood. Also, the flow of the Bremer at Walloon just before that time was more than the combined total of the Bremer at Walloon, Warrill Creek and Purga Creek on Saturday the 8th so there was no way that I can see that this could have caused a minor flood in Ipswich on Saturday the 8th. Yesterday’s below minor flood caused the river at Ipswich to rise by 3 m from a low level. I don’t know how much such a flood would cause the river at Ipswich to rise if it was already much higher from a Brisbane River back-up but it would certainly be less because the river surface would be wider. I’m not sure how to go about finding out accurately how much but the below minor flood in Ipswich on the 8th would still have allowed a lot of room for backup from the Brisbane.
@Nick
Isn’t that the truth?
That’s just not true. As I pointed out earlier, Wivenhoe was at 102.1% and Somerset 102.9% on the morning of the 4th, so thats already 35 Gl that Brisbane (and Ipswich) had dumped on it later. And they demonstrably let it get worse on Tuesday, Wednesday and Thursday until they had 101 Gl on Friday morning that they dumped on Brisbane the next week. Their management wasn’t even reactive. Reactive management doesn’t mean starting behind. It just means reacting appropriately to the circumstances.
I cannot see why you place so much importance on avoiding minor floods in Ipswich. Minor floods do minor damage. The biggest effect is flooding a few roads. Minor floods are quite common in Ipswich in some decades, even with Wivenhoe. What you are implying is that we should reduce the frequency of minor floods at the expense of increasing the frequency of major floods. I just don’t think that’s a rational trade-off. The damage from one major flood is worth at least the cost of a thousand minor floods. I think it would have been a long time before Wivenhoe saved a thousand minor floods but I think their priorities are going to be different now.
Is that supposed to be an argument that it was pointless to gain an extra buffer? I’m shaking my head in disbelief.
“Is that supposed to be an argument that it was pointless to gain an extra buffer? I’m shaking my head in disbelief”
Chris, you are already trespassing on my goodwill, by continuing to post. Please stick to facts and state your views without deriding those of others.
@Chris O’Neill
The reason why Ipswich is below minor flood in yesterdays conditions is because the Brisbane River is now at its lowest since very early in January. If the past two days storms had delivered water through the Bremer system while the Wivenhoe drawdown of 300GL/day at the wall was continuing flooding would have likely briefly returned even to moderate levels at the Ipswich gauge as waters were backed up by high Brisbane river flows. You can see the clear relationship between levels in those circumstances by flicking between Mt Crosby and Ipswich gauges for the 17/18/19th.
I’ll repeat,when operators were expressing concern about Bremer and Lockyer levels, Bremer levels were high,and to rise to higher levels,and Lockyer was on the rise and dam releases were increasing. They did not know then what we all know now,and that simple reality has to be foremost in your mind when you judge their actions. Rainfall in the upper Brisbane catchment was scattered,and not particularly high even by the 24hrs to 9am on the 9/1,though inflow increased
From a peak of 123% on the 29/12/10,operators had lowered the buffer 250GL by the 4/1 and were easing releases.The late December event that had produced that rise in Wivenhoe had also produced flooding in the tributaries,and flooding downstream to Mt Crosby. The buffer was at its lowest on 4th Jan. That just IS true,Chris,whether they were 30GL over 100% is hair-splitting. Whatever you find difficult to believe about Bremer @ Ipswich on Sat 8th,concern was explicitly noted in press release by the 10th a.m. and I have no reason to believe that water operators had only just evaluated this relationship on that day.
I am not arguing that it was pointless to gain extra buffer.Surely that is one of the principles of operation. I simply cannot see the whole episode unfolding clearly in favor of dramatically reduced or eliminated flooding even if extra buffer had been gained given the realtime rain/inflow breakdown that operators had to face,or that it is reasonable to post fact optimise that buffer size to a best possible and proceed to judgement from there.
I’m a little baffled by your comments at times. There is nothing implicit in my views supporting scenarios that reduce the frequency of minor floods at the expense of increasing the frequency of major ones,and there is no evidence in the operation of the dam since 1985 that management policy has produced this outcome.It’s actually physically difficult to do.The volumes that produce major floods are not easy to come by. If one can contemplate a trade-off between minor and major it seems to be rare opportunity.
Hedley Thomas at The Oz has partially released one operators emails about the unfolding events. Despite the leading headline and unhelpful attitudinising,the emails reveal an insight into how real-time information was informing their decisions. The events of Tuesday morning are significant.
The peak flow values claimed to be communicated to the OZ by SEQwater raise some questions about how much of the Brisbane peak came from the catchment below Wivenhoe.If the 9,000 cumec figure is correct,it suggests perhaps one third of that came from tributaries and stormwater below the dam. The suggestion that peak release at Wivenhoe was 8000cumecs is contradicted by earlier press release figures of 7465cumecs.I hope this is not simply a case of rounding up for effect by the paper. As we definitely know,the transient peak at Savages was about 6900 cumecs at 1am on Wednesday 12. This includes Lockyer flows and smaller tributaries near the dam,an area which experienced the last most intense rainfalls of over 250mm in the last 18hrs of the event.
OTT but strangely not by much…
The notion of levies has been floated by our Prime Minister as a crucial means of funding the fix-up after the deluge. Levies upset some people, when they are of the financial kind.
My Mum wondered why the state and/or federal governments couldn’t set up a series of lotteries to raise money for the big fix-up: they have the dual advantage of being voluntary – levies on those who are still contemplating how to get their lives back together is not a good thought – and of being able to raise substantial amounts of cash. Plus, there is a winner!
Mum reckons they did this for the Opera House, so it isn’t without precedent. It has probably been mentioned (a few times) before and I haven’t noticed, but I thought there is no harm in floating the idea.
Unfortunately gambling is now a state run enterprise and any new lottery could take business away from the existing bunch.
The emails released in the Australian and previous work by Hedley Thomas indicate pretty clearly a number of things.
1. SEQ dam people were at work on the weekend before the floods.
2. They were concerned about causing minor to moderate floods by releasing more water on Saturday and Sunday, so they didn’t.
3. This was in line with their operating procedures
4. Their action was understandable given that they did not know for certain how much rain was going to come down on Monday and Tuesday, even though they knew from the bom forecasts that it could be a lot.
We now know that if they had released substantially more on Saturday, Sunday and Monday that this would have meant the release on Tuesday would have been much less, and major flooding in Brisbane would have been avoided.
The problem is we know this in hindsight.
But there is enough information from this unfortunate situation as to how the release operating procedures should be modified so that next time the damage will be less.
The new procedures should place greater weight on the bom rain forecasts, as these are much more accurate than they used to be.
And the procedures should better balance the risk of major flooding for a day against minor to moderate flooding for 3 to 6 days ie major flooding is much much worse than minor to moderate flooding, so the procedures should be modified in that light.
As you say, thing become clearer in hindsight. The big floods in the Hunter Valley were due to heavy rain falling in all the catchments of all the tributaries leading to a backing up in the system. The last big floods, in 2007, failed to break records due to there not being heavy rain in all areas. These catchments stretch for many hundreds of miles and are often in rough unpopulated terrain making it difficult to monitor. So it really is a wait and watch scenario.
@Nick
You only have to check the flows at Savages to know that you cannot make that assertion. For example, the flow on the 5th January was 6.4 Gl. The flow at Savages yesterday was 17.5 Gl, the lowest since before the flood. Thus my previous argument stands. Please check the facts before you make assertions that may contradict them. I feel like I’m doing 90% of the work here.
More later…
@Nick
OK, I guess you meant around the 5th of January. Sorry I didn’t realize that.
More later…
@Nick
Indeed the Brisbane at Moggill bottomed yesterday (20th) at 6:16 am (0.07 m). However, the flow at Savages the previous day was 41.6 Gl, more than it was on Friday the 7th, 27.6 Gl. The river at Moggill should still have been similarly low on Saturday the 8th, considering the flood peak took more than 24 hours to get from Savages to Moggill. So I can’t see any reason why the flash flood on the 8th was anything more than below minor in Ipswich on the 8th.
I’ve been looking through the river levels at Ipswich and Moggil from the 14th to the 16th when they were slowly decreasing at Moggill. The interesting thing is that the rate of drop at Ipswich was no more than the rate of drop at Moggill, perhaps a little bit less. This is in spite of the Bremer flow rate decreasing from 7.1 Gl/day to 3.5 Gl/day over a period of 36 hours. I would have expected this decreasing flow in the Bremer to cause it to drop more rapidly than the Brisbane at Moggil but this does not appear at all, thus that difference in flow rate, 3.6 Gl/day didn’t produce a measurable difference in Ipswich. I’m trying to estimate how much increase the Bremer flow on the 8th, 19 Gl, would have added to a Brisbane backup. So far it’s 4.4 x something too small to measure. If too small to measure is less than 0.2 m then the flash flood on the 8th would have produced less than a metre rise on top of a Brisbane backup. I’ll keep looking at the figures.
If I might be permitted another comment, I will add this.
The ability to point out an apparent best path in a complex system “after the event” (after many events actually) does not amount to an algorithm which can be used up front to always find the best path in the face of unknown future variables.
In fact, there can exist no algorithm (precise steps leading to a procise solution) for a problem with unknown future variables. Where events are governed by probablility (and are very complex in the sense of the multiple interactions between such events), the best that can be done is to develop a set of heuristics i.e. general rules (enhanced by modelling and calculations of outcome probabilities) to manage the problem.
Chris O’Neill’s last paragraph goes to the heart of the issue. I will first quote it in full.
“It’s very obvious to me that the initial problem is that the dam operators react slowly to events, e.g. on Thursday the 6th they stated the need for releases and even though the Lockyer at Rifle Range Road did not reach minor flood level until 11 pm that night, they said they were going to delay releases until after the Lockyer flood subsided. They had plenty of time to start releases before the Lockyer flood got to the junction with the Brisbane and they could have reduced releases by the time the Lockyer flood got to the Brisbane junction if they wanted to. Their decision meant it was going to be a long time before they got serious with releases, i.e. they were waiting for the Lockyer first to go up and then down. That was a very time-wasting response during which the Wivenhoe was going up by 50 Gl/day.”
Chris I respectfully submit to you that it is obvious in hindsight only. Please follow my reasoning. What appears as slow reaction might also be cautious reaction or in fact a “by the manual” reaction. The manual for flood comparatment operation amounts to a set of heuristics or general rules for action. At the same time the room for heuristic action is bounded by an ultimate imperative. The dam must not be pushed to the limit where it fails or even (if it can be humanly avoided) to the point where the flood fuse fails.
You say the delay to allow the Lockyer flood to subside was uneccesary. Perhaps one can say it was so after the event but could one have said it before the event? Before the event, there were at least two critical unknowns. It was unkown if the Lockyer peak would be prolonged or even exacerbated by further events in the Lockyer or lower catchment. It was also unknown that the next 2 or 3 days of inflows into the Wivenhoe would be so large as they turned out to be.
What you are actually advocating is that the operators should have taken a major gamble at that time; a gamble that they could make major releases and finesse them through before the Lockyer peak whilst depending on little lower catchment rain plus little follow up rain in the Lockyer catchement whilst also presciently forecasting extraordinary follow-up inflows into Wivenhoe.
At the time of the events, probability modelling no doubt showed that the best probability of getting a good outcome was to do exactly what they did. (I am not privy to how their probability modelling is done.) A flood was coming down one branch (from the Lockyer Valley) and they had nearly 100% of the Wivenhoe flood compartment available on the other branch. To make the decision to make major releases from Wivenhoe “branch” at that juncture (thus effectively not utilising the flood compartment) would have seemed almost perversely risky. If anything had gone wrong in that scenario then they WOULD have been in danger of negligence charges.
I hope people can understand that attempting to finesse the releases advocated in the face of the “known unknowns” would have been the more irresponsible path. Consistently managing in a high risk manner by finessing critical events with low tolerances for error will lead to a higher probability of greater disasters.
They had absolutely nothing to lose by starting releases on Thursday morning. Indeed, the dam was already too high two days earlier and continuously getting higher. At the very least they should have been releasing to stop it getting higher. This virtually cost-free strategy would have reduced delays in actions.
Their actions this time and in the past showed that they never believed the flood storage would ever be filled. As such they believed they could satisfy every little need for flood mitigation regardless of the effect on the storage. So if someone came to them and said “please can you mitigate our floods in the lower Lockyer” or “please can you not flood the low bridges on the Brisbane river”, they’d say “sure, no problem, you can have any flood mitigation you like”. All because they never believed the dam could fill up. If I’m guilty of hindsight then the only thing I think I’m guilty of is saying that the dam could fill up. We could go through the arguments they might have had about whether the dam could fill up or not but I think if you accept that the dam could fill up then that changes everything. It means priorities need to be established.
Once we’ve decided that priorities need to be established we can go about establishing them. The following is my opinion but the first things that fall off the list are the bridges in the middle Brisbane and the lower Lockyer valley. It’s interesting that the Lockyer is at the core of this disaster. First they wanted to mitigate a bit of flooding in the lower Lockyer and I’d say the delays associated with this probably added at least 150 Gl to the dams. Then on Tuesday night the Lockyer sent down a massive flood at about the same time as the peak release from the Wivenhoe. They could hardly have timed it better for maximum destructive effect.
Of course. They had at least twelve hours after they made their announcement plus the time it takes the water to come down the creek during which they could have been releasing dam water and then they could have decided to reduce releases if necessary. But they didn’t believe their dam would ever fill so they didn’t want to go to all that trouble.
They did already know that those inflows were large and going to be larger and they were planning to release them in the future.
They had enough time and they would have had more time if they had already been making releases which they should have been doing because the dam was already too full days earlier, and rising days earlier as well.
93.7% of Wivenhoe and Somerset actually. There have been two issues here. The first was the lack of action before the just moderate flood in the Lockyer. I believe I have established a clear violation of policy in that period. The second issue is the appropriate response to the Lockyer flood and whether the policy itself was appropriate.
The flood compatment was already being used and policy should have been to restore it for its number one priority purpose, i.e. to protect Brisbane. As I mentioned above, policy should not have given much priority to the lower lockyer. Wivenhoe can’t do very much for the lower Lockyer anyway, e.g. I guess if you’re upstream as far as Rifle Range Road then it’s too far upstream to cause much backup.
I don’t think there should be a fundamental right to claim negligence if flooding occurs with equal or less than what the natural flows would have been. I’m saying the dam should let through natural flows until near the moderate flood level. Only then should it start storing some of the natural flow. This would provide far better mitigation for Brisbane than it ended up with and it also provides mitigation against the Brisbane flooding the lower Lockyer. The value of trying to avoid flooding a relatively small area of the lower Lockyer is just not significant, I believe, compared with the risks imposed on Brisbane by reducing its flood reserve.
I’m sorry but I just can’t see how letting the lower Lockyer suffer a natural flood level or less comes anywhere near the disaster of flooding Brisbane.
BTW, some of the later mistakes have been leaked to the media.
@Donald Oats
I’m against these sorts of thing. Not only are lotteries almost always regressive in practice (the pool is skewed towards those on the wrong end of pareto distributions) but inevitably, today, they would be handed off to some well-established operator in the gaming industry (gosh who would be on the short list?) and of course they’d have to get fairly concessional treatment and a right to operate for a considerable length of time to recover sunk costs in the new venture. The Queensland government has been in the vanguard of flogging off public assets and it’s scarcely conceivable that the lottery would be run by either the state or the Feds (I don’t think the Feds even have such a power, constitutionally and certainly no tooled up department to do it).
Again, the least inequitable way of funding reconstruction is to choose the least inequitably raised funding pool available — the general taxation system, though Bob Brown’s idea that mining should have to pay a substantial resources rent tax is better. That was true pre-flood and it’s certainly not less true now.
I don’t know enough about the Brisbane set up to appreciate the quantitative constraints upon dam operation there, and I’m not about to try understanding something that is a career in itself.
It is not good to see that perhaps there were missed options which might have “saved” the day, so to speak, but I would be exceedingly cautious on assigning responsibility for a missed option to people who are operating under imperfect foresight (as in, what the Heck, who knew a deluge would hit precisely at that time across so many catchment areas), or under rules learnt harshly under the previous 1 in a 100/200/1000 year flood a few years earlier. A missed option discovered in hindsight might not have been an option at all, when gazing into an uncertain future set of events, as Ikonoclast points out quite clearly.
There is absolutely no harm in laying out a fine detail examination of the operation of what is a relatively new dam, as that is how previously missed options become part of the new playbook, if they meet all of the existing constraints upon the system, which is obvious; less obvious is that any change to rules to allow new options must not be at the expense of currently available options, unless those options are dominated by the proposed ones. Otherwise, all that will happen is new options will be allowed by new rules so that this flood event doesn’t happen again, all other things being equal; other options necessary for previously encountered situations might be excluded by rule changes. In fact, this paragraph is such a mess that I don’t ever want to be a dam operator 🙂
Let’s just say that as the environment changes the playbook must be adaptive. A peculiar set of pre-flood events may simply have exceeded the extent of knowledge as captured in the playbook. An enquiry that led to improvement of how we incorporate changing environments into the picture is probably a good enquiry. On the other hand, an enquiry that just puts the burden of what is ultimately a decision with terrible prospectiveconsequences for each branch of the decision (ie posssible choices confronting the decision maker) upon a few dam operators would in my book be unconscionable. Do I risk killing some people by a controlled flood in order to avoid an uncontrolled flood of indeterminate probability, or do I risk releasing only safe amounts of water at the possibility of killing some people through a bigger uncontrolled flood of indeterminate probability? Do I risk structural failure to get enough water out early to avoid an even bigger flood that might not even eventuate? Tough choices indeed.
In any case though, I’m willing to bet that the controllers (or whatever they call the staff inside the operations room) who have to actually push a button to activate various releases have agonised both over the decisions leading up to the floods, and the events post-flood. They didn’t ask for all the water though…or for the urban build-up over old flow areas. It wouldn’t surprise me either, if some of them, upon knock off time, had nowhere to go home to anyway.
@Chris O’Neill
“Their actions this time and in the past…” When in the past?”…showed that they never believed the flood storage would ever be filled” Ever? I doubt that the engineers would ever professionally hold a view that the buffer would never,ever be filled,either in the long term or per event. It really is not possible to hold that view when the operating manual does not. The established policy to empty the flood buffer within seven days of a peak certainly contradicts your view. Your next sentence,and the whole paragraph, is also completely contradicted by the manual’s strategies and scenarios based on full dam[s],in fact the simple presence of the manual speaks against your attitudinising.
Chris,if you were in Flood Ops on the 7th,how sure would you be of the rainfall quantities over the next four days?
“I’m sorry but I just can’t see how letting the lower Lockyer suffer a natural flood level…” Why do you forget the Bremer? “..or less comes anywhere near the disaster of flooding Brisbane.”
I don’t think anyone would disagree. Do we need this rhetoric?
You claim to have established a clear violation of policy. If its for the 5/1 and 6/1,I’d agree. On the sixth day,the 4/1,after the flood peak of 29/12,operators were within a 2.4% of reaching the level mandated within seven days.Instead of getting rid of that last 28GL,they let the dam rise on the 5/1 and 6/1 by a total of 1.1%/ nearly 13GL,then 3.1%/36 GL by the morning of Friday 7/1.That seems to be a violation of policy for the first two of the three days.They were releasing water on those days ,but only a very small amount- inflows were low.Then the 6/1 and 7/1 saw the ‘pre-flood’ flood enter from the Lockyer. At this time they were perhaps following Strategy W1A in the manual. But they certainly could have shed that 28GL in the first 48hours and only cut Twin Bridges and Savages and likely Colleges Crossings.
Any inquiry exploring negligence and culpability questions will have to establish the no-dam flood profile to the best possible accuracy once all rainfall and streamflow data is quality controlled and integrated. Then,whether operators followed procedure according to the manual,and given their more limited view according to their real-time sequential access to the uncontrolled data,whether their hourly response choices were reasonable. Then,establishing whether the manual needs revision is another matter.
As for Hedley Thomas,just stirring. Try and read those email snippets independently of Hedley or hindsight.
I’ve had a look at the dam operations manual and I must say I am completely mystified by how they got from a dam level that only allows the relatively low release rates on Saturday, Sunday and even Monday to a level that requires the dam-protecting release rates that began on Tuesday. Specifically:
Why did it take so long to get to Strategy W2 and then so fast to get to Strategy W4?:
Read the manual to go through the W1, W2, W3 and W4 strategies but briefly (and incompletely):
Strategy W1 deals with dam levels of 67.25 m to 68.5 m.
Strategy W2 deals with dam levels of 68.5 m to 74 m.
Strategy W3 deals with dam levels of 68.5 m to 74 m.
Strategy W4 deals with dam levels above 74 m.
You’ll note that the range for W2 is 4.4 times the range for W1, i.e. for the same rate of volume accumulation, it will take more than 4.4 times as long to pass through W2 as it does to pass through W1.
W2 should be reached long before the dam reaches the 148.4% level it reached at 9 am on the 10th. If you take the 74 m level as being 200%, then W2 should start below 100%+(100/5.5)% = 118.2%.
In spite of it appearing that they should have been well into W2 at 9 am on the 10th according the manual, their release rates at that time appear well short of what the manual allows, i.e. the manual allows 3,500 cum/sec flow downstream of the Lockyer (300,000 Gl/day rate) but the actual flow rate at Savages at 9 am on the 10th was 182 Gl/day and remember that W2 should have been entered long before this, I’d say by midnight the previous night at the latest.
At this point it, to me it looks like they were seriously violating the manual.
More later…
@Fran Barlow
My Dad agrees with you concerning the handing off to the gaming industry, of any lottery-based system. And indeed, the outcome is probably a function of human nature, a fine fiddle to play for the human situation. Still, a tax will be hard to sell and especially hard to get by the Liberals if Abbott’s recent musings are a guide. I hope the Gilliard spin-team are up to it; and, the strategists should surely be able to blunt an attack against a levy that is designed to help the flood-affected citizens of Australia. On recent form though, I’m pessimistic.
And now a statement from the prime minister on the not yet proposed but maybe being considered special inundation levy:
‘The special inundation levy process is an area where the transformative power of education requires a community consensus as reflected in a couple of forum group reports on the inaccurate things we believe about the bogan vote in Western Sydney. A citizens assembly would be a good way of proceeding. Um, we think. Maybe. Did I mention the transformative power of the education process yet?’
@Donald Oats
A levy would be a great wedge issue because the Libs are insisting on “living within our means” and insisting on reconstruction post-flood. The NBN is popular, especially in QLD.
And think of the puns in levee and levy one could exploit.
Oh dear. I had thought SEQ water had followed their manual, because they and some other experts were emphatic that the procedures had been followed, and the release of emails indicated they were conscious of the procedures. But it is now clear that they complied with the procedures only in the loosest sense ie they at various times followed the rules required under strategies W1, W2, W3 and W4, but the times at which they applied strategies W1, W2, W3 and W4 were not appropriate. What a terrible, terrible stuff-up. There may be a good explanation for what they did, but its not obvious what it is.
The dam levels used by the manual is “A prediction of the maximum storage levels in Wivenhoe and Somerset Dams”. In that case they should have known before 6 pm Sunday the 9th that the Wivenoe was going to go over 68.5 m, even if it wasn’t actually at the time. This is because gauges upstream of the dam showed that a massive deluge was on the way, e.g. Linville jumped to 263 Gl/day by 4 pm from 13 Gl/day earlier in the day, Kilcoy Ck was 70 Gl/day at that time from 2 Gl/day earlier in the day and the Stanley River at Peachester was 21 Gl/day at that time from 2 Gl/day earlier in the day. And this is without even paying attention to the weather forecasts which should have warned them even earlier. These flows should definitely have produced a forecast dam level high enough for Strategy W2. So Strategy W2 should have been operating as soon as possible after 4 pm Sunday, if not earlier.
However, the minimum flow the manual seems to imply for Strategy W2, 2,000 cum/sec or 173 Gl/day was not reached at Savages until nearly 8 am on Monday, at least 16 hours after they should have known it was needed.
Also, the release rates over the weekend vary in a way which is not specified in the manual, apart from being rates allowed by the W1D Strategy (Kholo Bridge and Mt Crosby Weir Bridge). I presume this strategy was allowed by a dam level of 67.75-68 m. A 68 m dam level is equivalent to 109.3% (dam surface is 10,820 ha at FSL of 67 m). Once they were in Strategy W2 and knew the huge rate of inflow, the manual says they should increase it to 3,500 cum/sec or 300 Gl/day. This flow rate was not reached until about 7.30 am on Tuesday the 11th. Why did they take so long to turn up the release rate to what I presume was the required level? Almost immediately after this, between 8 and 9 am, the operators caused the fastest rise in flow rate of the whole event. What is the point of allowing a 3,500 cum/sec flow if they are not going to let it happen until immediately before they switch to Strategy W4? It appears the operators had no interest in a rapid enough response until the dam fuses were in danger.
I also notice that the manual is not specific about required minimum release rates at each strategy level. It actually only specifies maximum release rates until Strategy W4 where it is unlimited. The actual release rate is left to the operators discretion with a maximum limit (except for W4) but no minimum. So the manual doesn’t actually require any particular minimum release rate at any level. I think this could be a serious shortcoming of the manual.
I have also calculated that the level at 9 am Monday morning, 148.4%, meant the dam was at 72.2 m, way above the level to switch to Strategy W2/W3 (68.5 m) and only 1.8 m below the level for switching to Strategy W4. What the hell had they been doing before that time?
By the way, they should have known at 11 am on Sunday morning that there was a big deluge on the way to Wivenhoe AND the dam was already at 68.55 m, above the level for switch to Strategy W2/W3 (68.5 m). This is looking very serious.
@Nick
“I’m sorry but I just can’t see how letting the lower Lockyer suffer a natural flood level…” Why do you forget the Bremer?
Looks like you can’t let go of the Bremer. All I’m going to say is that the dam operators said nothing about the Bremer on Friday, Saturday and Sunday. I’m no longer interested in dredging up yet more figures to win another argument.
A couple of figures:
Volume of water on Tuesday, Wednesday and the first hour of Thursday in excess of 3,500 cum/sec at Savages: 280 Gl
Additional volume of water the dam operators could have released according to the manual on Sunday, Monday and Tuesday: 303 Gl
By the way, I claimed that the operators acted as if the dam could never be filled. This appears to be the case under Strategy W1 which allows 1.5 m in the dam or 14% for minor flood mitigation. Of course, under the next higher strategy they are supposed to act completely differently.
If what Chris is saying is accurate – then I am appalled and angry right now.
@Chris O’Neill
Chris,the WaterGrid morning press release of the 10th of January notes,second paragraph:
“Although releases are being made,large quantities of water continue to flow into the dams and are being held back in order to minimise impacts downstream and allow for other inflows from urban runoff,and the Lockyer and Bremer Rivers to subside.”
Despite your claim “.. you are not interested in dredging up figures to win another argument”, this seems to be your sole interest..You have been,understandably,scrambling to get across the data for days after having clearly put your cards down claiming that SEQwater’s actions are essentially incompetent. You are now trying to define and judge actions by reference to a manual you have barely started to digest. I think you should start with a clear grasp of the data before you even venture a judgement.
@Nick
But later they did indeed release large quantities, and these releases did indeed coincide with the Lockyer peak.
Holding back to minimise impact is one thing, holding back so you limit your options and are essentially forced to release a whole lot more at the worst time is another thing altogether.
The inquiry should immediately establish an interim release date in the next one to two months to head off speculation – which is adding more pain to those badly effected.
@iain
Yes ,Iain,the strategy changes as the situation changes.The [scenario] process is about the shift of focus from mitigation to preservation of the dam walls.
I’d also like to see a quick but credible data crunch. With newspapers like the Australian around the place,we will see nothing but misinformation. But even with a quick-cred reconstruction, it still takes longer to establish whether actions were reasonable.
@Nick
I think we are being too apathetic towards accepting the timing given by the government to write this report.
Do we really need two wet seasons to play out before findings come through? And then, how much longer until procedures are updated?
It has already been shown that they didn’t update their procedures to account for La Nina last year. How much time is needed for these sorts of things to be written or updated?
@iain
I don’t think a good report can be rushed. There will be interim findings in August. There is so much cleaning and claiming to go through in the coming months that it would be good for insurers to get most of the way through their work,before any issues that may lead to counter claims and class actions get underway.There seems to be plenty of issues about flood cover adequacy and insurance company transparency already. We don’t need flood recovery paralysed by premature and poorly supported ‘findings’ about woulda/coulda/shoulda stuff
The 100% water supply maintenance philosophy was not a problem last wet season,which was by no means below average. In fact the good rains of late summer brought the dam up to 100% for the first time in many years and any freshes were dealt with without resort to buffer use. And the dam was below 93% in August 2010. A La Nina forecast is only a loose indicator of local likelihoods in a 13000km2 catchment,and the forecast is subject to constant revision.I do grant that if a major flood was to happen, it would be in a La Nina year and by October or November 2010 a major La Nina was entrenched.
I’ve asked this question before. How far would a decision to eat into WS percentages have gone if taken at the onset of La Nina? Then how much further to lower the reservoir as the intensity of the La Nina became well established? Yes,it could have been done,there seems to have been time to do something by start December,or even keep lowering the dam after the October event when the dam went up over 120%. The presence of the desal makes all this easier to do,as Dr Q has noted.
@iain
I hate to be so cynical but two wet seasons is a long time – is it time enough for people to lose the resources necessary – having lost resources already – to undertake to claim if there was any negligence on any instutions part (whether its private insurers or statutory authorities)?
Hat Tip: Deltoid:
Andrew Blot, fooled again by his own echo, shoots himself in foot over floods claim. Blot wrong again (this time on floods and Warming)
@Nick
And the 10th was Monday, i.e. not “Friday, Saturday and Sunday”. I don’t know why I have to go over this with you over and over again.
@Chris O’Neill
If Bremer and Lockyer flows were mentioned on Monday morning as possibly to subside [we know they rose further,yes],then operators were aware they were elevated [as in,in flood] over the previous days. Yes or no?
Your figures are rubbery,your flow timings are making big assumptions,and your estimates of storage percentage to dam level height are questionable [148.4% is equivalent to about 71.6m,not 72.2],and now you want to pass judgement on how the operating manual must be used?
@Nick
Another set of goalposts.
No, it does not mean they were in flood over at least Saturday and Sunday (until late) as I have shown previously and the figures for which you chose to ignore.
It’s a pity you rarely say anything specific about my figures and timings so that I can have the opportunity to correct them and the conclusions if necessary. It’s much easier to just sit on a high horse and make non-specific complaints about rubbery figures.
I based that figure on assuming the dam surface area does not increase as the dam rises so yes, that does make a big difference to the level and a big difference to my claim:
I’ll use your figure to give a corrected claim:
The level at 9 am Monday morning, 148.4%, meant the dam was at 72.2 m, way above the level to switch to Strategy W2/W3 (68.5 m) and only 1.8 m below the level for switching to Strategy W4. What the hell had they been doing before that time?
Try agin:
The level at 9 am Monday morning, 148.4%, meant the dam was at 71.6 m, way above the level to switch to Strategy W2/W3 (68.5 m) and only 2.4 m below the level for switching to Strategy W4. What the hell had they been doing before that time?
@Nick
My position at the beginning was that they don’t risk causing billions of dollars of damage without a very low risk of that happening. Thus they were either very unlucky or incompetent. By definition, very unlucky means they were probably incompetent. I don’t think it was unreasonable to imply that the most likely cause was incompetence.
You are now trying to define and judge actions by reference to a manual you have barely started to digest. I think you should start with a clear grasp of the data before you even venture a judgement.
@Nick
Is this clear enough for you:
At 11 am on Sunday morning there was a big deluge on the way to Wivenhoe AND the dam was already at 68.55 m, above the level for switch to Strategy W2/W3 (68.5 m). The switch to W2/W3 (> 2,000 cum/sec.) did not occur until at least around 7 am Monday morning or shortly before by the amount of time it takes water to flow from Wivenhoe to Savages.
@Chris O’Neill
It’s not clear enough for me at all. These are serious matters requiring a careful analysis in the light of full record, not back-of-envelope calculations based on partial information. I wouldn’t like to see either the apportioning of blame or the Wivenhoe flood management strategy rejigged on the basis of comments made here. This is a respectable blog not a tabloid newspaper.
You might be vindicated by the enquiry but I’ll wait and see.
@Chris O’Neill
Consideration of Lockyer and Bremer flows is not “another set of goalposts” If you have read the manual,then you should have a good idea of the goals,whether or not they are explicitly referred to in every press release.
As for my being on a “high-horse”,I reject that completely. Really,you are the one in the saddle. I am reluctant to pass judgement because for the umpteenth time,we cannot walk through the event in real time as bits of information arrive.We lack so much data,to consider,defer or reject,and while the manual gives recommended timings for gate movements,we have no idea how to ‘play’ the controls.There is some discretion granted to the senior controller. At the speed that events got to,no-one in control had prior experience either..so there is an unavoidable level of inexperience. They became unlucky,and their unavoidable inexperience became telling. Given that no engineer in Australia has ridden a major flood on the Brisbane with manual or without,I feel keenly for them when they are attacked as incompetent.No one else approaches their competence.
You think they should have been ahead of the curve into strategy W2 or 3 at least 24-even 48hours earlier than early Monday 10th,but we really do not know when W2/3 was activated just because flows were not optimised in the way you wish they had been. Maybe they were already in W2 then,which aims to keep natural flows at Lowood below 3,500 cusecs . Combined flows [Savages] were at around 1000 cusecs on mid 8/1,hitting 1700 by midnight 9/1 and 2850 cusecs late on the 10/1. Outflows at the wall were rising from 1340cusecs very early on the 9/1 to 1965cusecs by midnight,and by morning 10/1 were rising again to 2780 cusecs by late that day. (So you see there is always the discrepancy between dam measurements and Savages measurements,a function of different site hydrology.) On the 11th,the press releases also suggest some restraint in quickly raising releases further. I think they were working through W2,having decided at the time that they could handle what upstream data were telling them,and in the face of strongly conflicting projected rainfalls [courtesy of your linked Weatherzone thread].
I think that they still were thinking local minimisation- the lowest priorities need not be discarded under W2- on the 9/1,then by the 10/1 the Lockyer started its rapid rise and the focus shifted to avoiding the Lockyer peak,before,during and immediately after if possible.They wanted to avoid putting too much water into the river before the Lockyer arrived at a faster speed and climbed on the back of any elevated releases moving downstream. Unfortunately,upstream inflows[after slackening quickly p.m. 10/1 , second peaks developed rapidly] then started to increase,and buffer was being swallowed,and they were forced to increase releases just as the Lockyer rose rapidly during the 10/1 to record levels by late a.m 11/1 when the gauge could not record higher. The Lockyer put 30% of total flows past Savages over the 48Hours 11-12/1,and they were ultimately unable to avoid coincidence with them…or the Bremer which pushed in over 110GL for the 11th and more the next 24hours.
The Lockyer was pushing in 700+ cusecs at the junction early 11/1,as they were releasing 2,700 and climbing. They were through W2,then W3 to W4 by midnight because of that second peak and the intense rains overnight 11/1,which were well in excess of even the closest modelling forecasts. These explosive developments do make actions on the 8/1 look inadequate but they were outside forecast reach.
Factoring in the Lockyer and Bremer,plus over 500km2 of smaller tributary catchments [Splityard,England,Banks,Black Snake,Cabbage Tree,and Sandy-Watercress,etc] that saw the most intense falls in the last 36 hours to midnight 12th,up to 50% of the peak in the Brisbane at Moggill may have come from below the Wivenhoe.
Hazard a guess on the size/volume of a lower peak if more water was pushed out 8-9/1?
@Jim Birch
OK, which part isn’t clear enough? The dam level was 68.55 m before 11 am 9/1/11 with rapidly rising tributaries and heavy rain. Is it not clear that that was already above the level for switch to Strategy W2/W3 (68.5 m)?
Strategy W2 says:
“The intent of Strategy W2 is limit the flow in the Brisbane River to less than the naturally
occurring peaks at Lowood and Moggill, while remaining within the upper limit of non-
damaging floods at Lowood (3,500 m3/s).”
Is it not clear that that allowed the operators to release 3,500 m3/s (302 Gl/day) as soon as possible after W2 was required?
Is it not clear that the the operators took another 44 hours (until 7 am 11/01/11) before they increased releases to 3,500 m3/s when time was of the utmost essence?
I note that the manual doesn’t actually require the operators to release any water at all, as long as they protect the dam itself. It will be interesting to see if the operator puts up that disingenuous defense in court.
Yeah, let’s hang the dam operators out to dry…
An infeasibly large amount of water arrived in a sequence and at a rate that makes flooding impossible to avoid, as far as I can see. Sure, the extent of it might have differed somewhat under different operational events concerning dams and such, but really, as Nick and Jim Birch point out, understanding to what degree some flooding was avoidable is up to a detailed post-mortem, not the rush to tar-and-feather the dam operators of Wivenhoe dam. I think from here on, I’ll wait for the enquiry, or whatever vehicle they come up with, to complete its mission and releases its findings. Hopefully the dam operators and associated staff will be treated fairly in the interim.
And yes, Fran Barlow, it is too funny and too tragic to not mock, so mock I have. I shouldn’t, I know, but this guy has such a penchant for planting evidence to suit a conclusion, rather than concluding based upon the evidence. Hmm, the parallels…
@Chris O’Neill
Yes,Chris,it is clear that it would have been within operational policy to release more water earlier.
It is not clear that the operators decisions were not made to a reasonable standard in real time with real time data. Because we have not seen their data,advice or discussions.
As facetious as your last paragraph is,it’s pretty clear that the dam cannot be protected by not releasing water. Judging by the photos of the mass of rock debris in the spillway channel in the Courier Mail today,releasing water ain’t that good for it either,hence the advice on gate sequence and incremental timing in the manual. I wonder whether there is a sensor network in the spillway bucket and channel that lets operators know that severe erosion is taking place,and ,if so, whether at all this influenced decisions? What ‘s the condition of the substrate beneath the concrete apron under the spillway?