Saturday, March 31, 2007

Considering Collignon’s Concerns and Comments

This week, the ACT Government launched a three-month period of community consultation to consider the merits -or otherwise- of developing a planned indirect potable water recycling scheme to supplement Canberra's drinking water.

The “Water2WATER” project would involve upgrading effluents from the Lower Molonglo Water Quality Control Centre to potable quality so that they may be then be placed back up into the catchment of the Cotter reservoir. The ACT water authority, ACTEW is currently investigating which combination of technological barriers would be the best option but the decision has not yet been announced. From the Cotter reservoir, the water would be treated again at the Mt Stromlo drinking water treatment plant before being distributed to customers.

The proposal has received tri-partisan in-principle support from the governing ACT Labor Party, ACT Liberals and ACT Greens. However, like all planned indirect potable recycling schemes, it is sure to have its sceptics and detractors.

The most high profile ACT opponent to voice concerns so far is Professor Peter Collignon. Prof. Collignon is the director of infectious diseases and microbiology at the Canberra Hospital, and a Professor at the School of Clinical Medicine, Australian National University. As an eminent and highly respected expert in his field, Prof. Collignon’s concerns must be taken extremely seriously.

I expect that Prof. Collignon has or will formally present his position as a submission to the community consultation process. However, in a commendable effort to raise awareness and stimulate community debate, Prof. Collignon this week detailed his views in an accessible format as an opinion piece titled ‘Recycled Water Not Worth Risks’ in the Canberra Times.

Prof. Collignon’s concerns are numerous. He states that the proposal does not appear to make sense in environmental or economic terms. However, given his position and expertise, I assume that his primary concerns regard risks to public health. He wrote:

“A number of methods are [proposed] to make this recycled sewage "safe" but how many systems work perfectly all the time? If membrane technology is used, how can we be sure that these membranes will be able to accommodate the planned 24 million litres of recycled water that they need to filter each day? How will we know when there are small tears in parts of the membranes? Bacteria are very small and unless the pore size of these membranes is 0.2 microns it is unlikely that all bacteria will be removed.”

The first question, regarding system reliability, is –I acknowledge- the most difficult to answer. I concur that no engineered system should be assumed to work perfectly all the time. However, there are well established means for ensuring that we anticipate ‘unexpected’ malfunctions and design water management schemes such that malfunctions do not present elevated risks to public health. We do this by implementing a high degree of ‘redundancy’ in the system. This means that we have multiple barriers to contaminants that we are concerned about and that the barriers operate independently of each other. Even if a number of the independent barriers malfunctioned completely, the water would still be fully protective of human health. If an indirect potable recycling scheme does not provide this degree of surety, it simply has not been designed according to world’s best practice and should never be built in this country.

The question regarding the ability of water treatment membranes to accommodate 24 million litres of water each day is similarly a question of treatment plant design. A Californian indirect potable reuse scheme reliably treated more than 100 million litres per day by reverse osmosis and is currently being upgraded to a capacity of 266 million litres per day. A real advantage of membrane treatment systems is that they are modular and can be expanded relatively easily to accommodate any necessary capacity. Spare capacity can be built into a system relatively easily.

The ability to detect “small tears in parts of the membranes” is extremely important. Numerous approaches are used at different plants. The integrity of membrane treatment processes is monitored and checked by a number of direct and indirect measurements. In most advanced water recycling schemes, water quality is monitored by ‘Supervisory Control and Data Acquisition’ (SCADA) systems with sensors placed at strategic locations within the treatment process and at the final point of dispatch from the plant. The most useful online parameters are typically conductivity, turbidity and total organic carbon (TOC). These may be used to indicate the quality of water leaving the plant as well as for closer monitoring of individual treatment modules or ‘sections’ of plants to identify maintenance requirements. Significant ruptures of reverse osmosis membranes (which are exceedingly rare) can be observed by sudden drops of pressure across the membrane. These can also be continuously monitored by pressure sensors.

Prof. Collignon notes that bacteria are very small and that they would not be effectively removed unless the pore size of these membranes is 0.2 microns. He will be relieved to know that most commercially available microfiltration membranes do a fine job or removing organisms of this size. Ultrafiltration membranes typically have pore sizes between 0.1 – 0.01 microns. Nanofiltration membranes have pore sizes 0.01 – 0.001 microns. Reverse osmosis membranes remove contaminants of 0.001 – 0.0001 microns which is numerous orders of magnitude smaller than any known bacteria or other living organism. Removal of both bacteria and viruses is assured and membrane processes are only one or two of the multiple barriers that a full treatment system would incorporate.

Prof. Collignon wrote:

“However, if the pore size is so small, I find it difficult to see how these membranes can satisfactorily work without being frequently blocked by larger waste material. Even if such small pore sizes are used, this will still not remove viruses, which are much smaller. Membranes will also not remove drugs passed in urine and faeces that are not broken down (such as oestrogens)”.

Prevention of membranes being ‘blocked’ (or ‘fouled’, as we tend to say) is indeed a key challenge for the operation of advanced water treatment plants. However, it is achieved in Australia and overseas by careful pre-treatment. This typically involves placing a microfiltration or ultrafiltration process in front of the reverse osmosis process. Nonetheless, membranes do have a lifespan which is typically dictated by gradual fouling and thus need to be periodically replaced. This is part of the anticipated cost of operating a membrane treatment process.

I am a little bit shy about telling an eminent professor that he is wrong about the inability of membranes to remove species as small viruses, drugs or oestrogen hormones. But frankly, on this point he simply is uninformed. More than a decade of peer reviewed literature has demonstrated the ability of membranes to remove these contaminants. I suggest he check some of the references that I have previously listed here and here. Another peer-reviewed paper that demonstrates this is:

Drewes, J. E., Bellona, C., Oedekoven, M., Xu, P., Kim, T. U. and Amy, G. (2005) “Rejection of wastewater-derived micropollutants in high-pressure membrane applications leading to indirect potable reuse”. Environmental Progress, Vol 24, No. 4, 400-409.

Prof Collignon wrote:

“Ultraviolet light will also be used as an additional sterilising agent. However, this is far from an ideal disinfectant. There are many issues such as time of exposure, susceptibly of different microbes, and so on, for it to work. How can we be sure that this can handle 24 million litres of waste-water per day?”

Indeed, UV disinfection must be operated at sufficient doses, but assuming that the system is designed appropriately, there is no reason to assume that it wont be operated appropriately. Given pre-treatment by membrane filtration, UV disinfection is extremely effective, -more so than chlorine for many microbes. We can be sure that it can handle 24 million litres of water per day by designing the system for such a capacity. Much larger capacity plants are successfully operated in many places such as California and Singapore.

A point with which I strongly concur with Prof Collignon is the following:

“Safety monitoring is planned, presumably by culturing the water and looking at coliform counts. If coliforms (eg E. coli) are present in the treated water this implies faecal contamination (and thus a failure of the system). However, this type of monitoring has problems. Around the world numerous outbreaks with water contaminated with viruses and Cryptosporidiosis have occurred despite low or zero coliform counts. In addition these indicator bacteria take one or two days to grow and identify.”

I absolutely agree that coliform monitoring is not sufficient for ensuring the absence of pathogens such as viruses and Cryptosporidium. Numerous studies have demonstrated a very poor correlation between the presence of coliforms and these pathogens. Accordingly, coliform monitoring alone is not an acceptable means of monitoring pathogens in indirect potable water recycling schemes. Furthermore, no method of monitoring any pathogens just at the end of the treatment process is really sufficient. Instead, modern methods of quantitative microbial risk assessment (QMRA) are essential to ensure that a scheme is designed for reliable removal of pathogens as advocated by the Word Health Organization. One of my more senior colleagues, Prof. Nicholas Ashbolt, has been instrumental in introducing such approaches in Australia. This is a topic worthy of much more detailed discussion than is appropriate for this post, but one that I do want to devote some serious discussion to in a subsequent posting.

Prof. Collignon’s comments regarding the sensibleness of recycling water that already has a value as environmental flows are, in my opinion, perfectly reasonable. His preference for retaining water for Canberra that would otherwise be used by downstream communities and irrigators is also worthy of national debate.

I would suggest that Prof. Collignon’s comments are among the most valuable contributions that we have had to the ongoing debate about potable water recycling in Australia during the last couple of years. His expertise and experience are far too valuable to ignore. I was pleased to hear the ACT chief health officer Paul Dugdale acknowledge Collignon’s concerns. His response was appropriate:

“There are always risks in any complex system but they can be managed...That's where my role really comes in, is making sure that the management systems and the design is such so that if those risks of failure of one part of a system actually happens, that you pick it up and it doesn't cause health problems.”

The debate is exceedingly healthy and will ensure that all schemes and proposals are rigorously scrutinised in Australia. I’m sure many regular readers (and maybe some new ones) will want to discuss some of Prof. Collignon’s comments in further detail. I, too, am keen to do so, so please let me know what you reckon.


rwindsor said...

Dear Stuart
I feel this debate has gone off half cocked, or, rather the decision to have the debate has gone off half cocked.
While Prof Peter Collignon's concerns are valid for the technology currently employed by ACTEW (which when it was designed and commissioned, was leading edge) they are probably well addressed by current technology (especially RO) and by quality assurance programs such as ISO 9001. However, the political will to institute such programs is suspect.
In order for the process of public comment, called for by ACTEW and the Territory Government, to have validity, it must start from an informed base.
All of the publicly available documentation regarding the deliberations of ACTEW and Government can be found here
Nowhere is there information on the rationale of the decision making process. It is inferred that there are cost/benefit analyses that underlie the decisions. Where are they? Certainly not there for the asking of mere mortals.
If one reads the documents closely, one finds that the preferred option for long term security of Canberra's water supply rests easily with the Large Tennant Dam option, on all grounds disclosed. There are factors, not previously considered, which now reinforce the wisdom of this choice.
Recycling, at the stage these documents were researched and written, was so far out of the picture, it did not rate a mention in the front running contenders (and some of them were pretty far fetched).
Prof Terry Dwyer raised a number of issues regarding the nature of the deliberations and
These were done at the Independent Competition and Regulatory Commission enquiring into Canberra’s water pricing structure. Dwyer made the point that, to that stage, rational debate on the economics of water supply had become the first victim. (Interestingly, the ICRC is currently holding another enquiry into Canberra’s water pricing structure as the debate on recycling goes on).
Thus Collignon’s concerns regarding the safety of the water are, in reality, far less important than the economic debate Underlying, are the assumptions that we in fact need a new water source. Also the values we attribute to a crazy concept of “Environmental flows” concocted by someone with apparently no sense of statistics or ecological principles.
All that aside, the release of learned discussion regarding “Global Warming”, climate change and the southward movement of climatic patterns means that the whole debate needs to be reassessed in the light of greatly reduced runoff into the existing storages. Given, that with the current infrastructure, ACTEW can already extract from the Cotter catchment, all of the water in excess of the nominated “environmental flows” with room to spare, it seems economic suicide to enlarge the Cotter dam. This would cost some $350 million and pump reclaimed water from the Lower Molonglo treatment plant (at a cost of $11 million annually) in order to reclaim 9 gigalitres per year. (disregarding depreciation, amortisation, maintenance and opportunity cost). The alternative of the large Tennant option has some 10 times the yield for possibly little more than half the capital cost and a fraction of the operating cost, to supply water of extremely high quality.
In any case, all of the effluent from Lower Molonglo is currently recycled into the downstream supply chain via Burrinjuck Dam. Any shortfall in this ACT contribution to the currently agreed environmental flow would have to be substituted by releases from the existing dams in the Cotter catchment, negating any value of the whole exercise.

rwindsor said...

I should have started my comment with this link

Jim said...

Using recycled water for drinking is the cheap and easy option for governments!

Stuart Khan said...

Thanks for your detailed comment RWindsor,

Your argument underlines the need for a transparent triple-bottom-line approach to the evaluation of competing water management options.

I also tend to agree with your concerns regarding environmental flows. Its important to recognise the fact that treated effluent released downstream of inland towns and cities already has a significant value in terms of the water flow it provides. I think the loss of such water was major shortcoming of the Toowoomba proposal.

Thanks too Jim, for your comment.

Mark said...

Hi Stuart,

It makes me nervous that the technology is so new that we are making important judgments based on limited and pilot-scale data. There’s still a lot that we still don’t know about the technology. For example, I don’t think anyone knows how such a process will respond to a large unexpected chemical discharge into the sewers by accident, or (as we must consider post 911) deliberately by any irresponsible, ignorant or malicious whacko with access to a toilet bowl.

I’m glad someone on the public health side has weighed into the debate. I get a sense that many people are overestimating the capabilities of this technology (as good as it is) and underestimating the risks. Prof. Collignon is right to put public health at top priority. Even if (and this is by no means proven) advanced water treatment is cheaper than other options, safety - and the perception of safety - is I think worth paying extra for. Neither can public health be gambled or experimented with. We have yet to fully exploit alternatives on water either on the usage reduction or the resource side, but here we are reaching for probably the least desirable solution.

And perhaps if some problem did occur with the recycled water, you would hope that the authorities would at a minimum be open and honest with the public. Unfortunately, this is the kind of behaviour we are more likely to observe from authorities in this country.


rwindsor said...

Dear Stuart
I guess the sheer stupidity of all this is the fact that we unquestioningly accept fluoride in our water supply despite the fact that there has never (in my knowledge) been a double blind, placebo controlled trial which in any way suggests efficacy. We also accept levels of aluminium which are described as toxic in some studies. On the other hand , we fail to accept that low magnesium levels have been correlated with poor cardiovascular health and could easily be rectified by additives in the water supply (MgCl preferred)and also the correlation between high boron levels and low levels of arthritis.
It is a queer old world.
In reality, we are beholden to big business and the ill effects of contaminants in the whole food chain are ignored. All of Sydney's market garden areas have soil cadmium levels in excess of WHO safe standards. Does anyone give a toss? It is impossible for the home gardener to buy complete fertilizers which are not contaminated with heavy metals (they carry warning signs attesting to it)
Thus, arguing over the safety of RO is a bit like shutting the door after the horse has bolted. The system has broken completely

Stuart Khan said...

G’day Mark,

Thanks again for your comment.

I think pilot scale testing has been a very productive approach to improving our understanding about all types of water treatment technologies. There is much valuable data that can be gained and I would recommend it for all new advanced recycling schemes. However, this does not mean that there is not extensive experience with running full-scale plants. The Water Factory 21 Scheme in California had operated a full-scale reverse osmosis plant since 1976. It is currently being expanded as part of the Orange County Ground Water Replenishment Scheme.

I have heard people make the comment about terrorist attacks before. But frankly, I just can’t make sense of this suggestion. Only the most dunce terrorist would try to poison a city by placing some toxin in a water supply prior to advanced treatment including reverse osmosis and advanced oxidation. The water treatment scheme to be used by SEQ is extremely effective at removing all chemicals…there is not a single one that can slip through unaffected. Furthermore, an IPR scheme provides the opportunity to divert water as a wastestream (eg. as reverse osmosis reject), which most other water supplies do not have. Surely any terrorist worthy of the title would recognise these facts and have the sense to carry out an attack on any water supply at a point after the advanced treatment processes, -the most logical being the water reservoir. In other words, cities like Brisbane would be no more vulnerable with an IPR scheme than without. In fact, an advanced water treatment plant would be a rather handy thing to have in such a circumstance.

Prof. Collignon is indeed right to put public health as the top priority. I think you will find that he is not alone on that. I know of many scientists and regulators who have worked on risk aspects of water recycling for many years and they too also put public health as top priority. Your comment that safety is worth paying extra for is truly an understatement. IPR schemes would be significantly more simple and significantly less expensive if people did not consider safety worth paying extra for. Building a high degree of treatment redundancy into schemes is now standard practice.

Stuart Khan said...

Hello RWindsor,

I take your point about relative risks very well.

I have previously given some thought to the question of why people ascribe such high risks to drinking water compared to (for example) food. Practically any chemical that is likely to be found in any drinking water source is equally or more likely to be found as often and at higher concentrations in food. People seem to forget that food is a complex mixture of millions of chemicals including many man-made industrial chemicals at very low concentrations. If we could analyse most food anywhere near the detection limits at which we are able to analyse water (parts per billion), I’m sure it would be an eye opener for many. I think the difference is that water is practically the only substance that we consume in an almost chemically pure form. This makes contaminants, which may be present at say 0.0000001 per cent seem very significant compared to other more complex exposure sources. Well, that’s my take on it anyway!

Nonetheless, people clearly do associate a higher degree of risk on water quality compared to other exposure sources and they have a right to do so. It is then the responsibility of scientists and regulators to address concerns by demonstrating that safe water can be and will be provided.

rwindsor said...

Dear Stuart
It is obvious that you don't read the labels on your food :-) We are almost approaching the science fiction state of a "meal in a pill". Seriously, if the same standards of adverse event reporting were applied to foods as are applied to water and , more importantly, pharmaceuticals, then there would be a lot fewer additives in our foodstuffs .
Also, if there was the same morbidity and mortality associated with food or water as is associated with pharmaceuticals, the place would be in open revolt.

Mark said...

Stuart: "Only the most dunce terrorist..."

Frankly I can't make sense of anyone wanting to do something like that either, but I would not be blasé about the possibility. Your assessment that no toxin can pass through all of the barriers without being reduced to a poofteenth of its original concentration is probably valid. On the other hand, the assumption that no agent (not necessarily a toxin) could be used to degrade the barriers themselves is quite possibly invalid. If the dunce accomplished this much, he would not need to add anything toxic to cause a problem.


Stuart Khan said...


Without wanting to attract the attention of ASIO by discussing how (or at least how not) to carry out a terrorist attack on a water supply, I think your proposed approach would be among the most difficult, most expensive, most likely to be noticed and least likely to succeed.

But still, I do agree that it does make some sense to include ‘vulnerability to terrorist attack’ as a component of a comprehensive risk assessment for a major scheme. However, this is a study that you might not expect to be made public even if it were undertaken.

I worked in the USA for much of 2002/03 and at that time, President Bush was busy redirecting funds away from the EPA towards the newly created Department of Homeland Security. In order to obtain research funding, many water scientists switched their focus towards (for example) designing early warning systems for attacks on supplies. If any of this research was fruitful, there may well be numerous tools and strategies in place that we don’t even know about ;-)

Kevin said...

You said - The water treatment scheme to be used by SEQ is extremely effective at removing all chemicals…there is not a single one that can slip through unaffected.

Are you - or any government - willing to guarantee it?

Stuart Khan said...

Hi Kevin,

I don’t speak for any Government. But I consider their mandate to be to provide water that is fully protective of human health, -not to be held over a barrel on a scientific technicality.

But as for myself, yes –my reputation is already staked on every statement I make on this very public blog. Mark had expressed concerns for substances that a terrorist may discharge to the sewer and whether it might wind up in drinking water. For any substance, there are numerous barriers between the sewer and Brisbane’s kitchen taps that are effective for its removal, -either by biodegradation, adsorption, membrane exclusion, oxidation or volatilisation.

Mark said...

Hi Stuart,

My hope is that whoever has responsibility for this process uses all the available sensor and analysis technology even if it is top secret. Although as far as I can figure, the technology that should really be there does not yet exist [eg. M. Salgor et al. / Desalination 187 (2006) 29-40].

Here’s another free idea: I hope that someone will introduce depots here where people can dump liquid wastes to discourage pouring unwanted chemicals down the gurgler. Because even though sewage is what it is, and I’d love to believe that the French process can get rid of absolutely everything in virtually unlimited quantities, I’d still feel more comfortable if the cocktail already there can be as simple and consistent as possible.

Stuart, can you explain why the recycling process is described as having "multiple barriers", whereas as I understand it, most of the "barriers" target specific classes of contaminants. For example, RO might remove the bulk of the salts floating around in the water, but presumably UF and UV cannot reduce these significantly. Likewise, low MW organics may be zapped by UV, but RO or UF might do almost nothing to these. Again, some fine particulates need to be removed by UF because they may damage RO and UV can't knock them down. It would seem that for many kinds of contaminants, there is sometimes only one barrier doing the removing work. Therefore, don't you need every single one of the "barriers" because each performs a unique function? And if so, is it technically correct to describe any of these barriers as being "redundant"?

Another terminology issue: If you do "indirect" recycling, but you do it at such a scale that the residence time is much shorter, and the amount of recycled material is much greater than usually used in any other "indirect" scheme, is it really fair to call it "indirect"? Stretched to the ridiculous, if we have a near empty pond, and we pump water to it and extract it back again a few hours later, is this really achieving the intent of "indirect" recycling? What about a few days? Don’t you need a few months for natural biological activity to kill off some of the bugs?


Stuart Khan said...

G’day Mark,

Excellent comments and questions (as always!),

Liquid waste collections or depots are an excellent suggestion. Most large cities already have such depots or collection services, for example Sydney, Brisbane and Melbourne. However, I’m sure that it is fair to say that community awareness of such services is often not as high as it could -or should- be. This is particularly important for cities like Sydney where much of the sewage is discharged to the environment with practically no treatment. Nonetheless, I wouldn’t want to rely on such schemes for the protection of human health by a potable water recycling scheme, -simply because I am familiar enough with human nature to know that full compliance will never be achieved. It remains necessary to ensure that treatment processes are able to deal with anything that might be rightly or wrongly discharged to the sewerage system.

In the context of potable water recycling, the term “multiple barriers” is generally used to describe barriers to hazardous substances. Hazardous substances are those which may reasonably be expected to have a detrimental effect to public health or the environment. Thus it is true that some substances, such as light, very soluble inorganic salts (eg salts of sodium, potassium, chloride, iodide, etc) may have only a single significant barrier (reverse osmosis). However, this is a very effective barrier for these chemicals which are most unlikely to present a significant human health risk.

Low molecular weight organics do, indeed, have multiple barriers in planned IPR schemes. Chemicals which have an electrochemical charge are well removed by most reverse osmosis membranes (we are, after all, able to desalinate seawater, which contains much higher concentrations of very small, charged chemical species). Those organics which are not well removed by reverse osmosis are small uncharged chemicals. By their nature, they tend to be either lipophilic and/or biodegradable and/or volatile. These attributes make them well removed by primary and secondary sewage treatment processes (and conventional drinking water treatment processes at the drinking water treatment plant). As you have indicated, they are also susceptible to advanced oxidation processes.

I accept your point regarding the “intent” of indirect potable recycling. And yes, it is fair to say that certain barriers (such as the environmental ‘buffer’), will be more effective in some schemes than in others. But this actually applies to all barriers. For example, a UV disinfection process may be significantly more effective in one scheme than it is in another, -in fact, they can be highly variable depending on the design and selection of lamps. So when a scheme is being designed it is indeed important to carefully consider the actual effectiveness of each barrier, -rather than assume that it will be equivalent to those in existing schemes.

Nonetheless (and I know we have dealt with this point previously – but for the benefit of new readers), to cease being defined as an ‘indirect’ IPR scheme and become a ‘direct’ IPR scheme, it would be necessary to plum the advanced water treatment plant directly into the water supply mains (a la Windhoek in Namibia).

Mark said...

Hi Stuart,

Thanks especially for that tip about the waste collection. I certainly had never heard of it. I agree that you can't rely on such things but it could be a help at least (recycling both wastewater and storm water) especially if they advertised it and maybe even kept their web page (the Brisbane one) up to date.

I'm surprised that there doesn't seem to be a better definition for 'direct' and 'indirect'. It would seem that if Namibia wanted to play semantics, they could replace a section of closed pipe at the plant outlet with a short section of open launder, and that would convert it into an 'indirect' PRS. Should there not be a stronger scientific definition for 'indirect' as this definition seems open to abuse?

Good point too about those UV lamps. I read that the USEPA no less recently found they need to up the UV dose four-fold to kill some bugs to the level the old dose was supposed to [M.V. Yates et al., J AWWA, (2006) 98(6) 93-106] and this is going to make running these things quite expensive. I wonder if Brisbane is going to use the old dose or the higher one. Incidentally, I reckon that you would definately not go hungry if you had $10 for every time you read a paper in your field with words like "Many significant data gaps need to be filled ...".


Stuart Khan said...

Hi Mark,

I don’t quite understand where you’re coming from about the need for a “stronger scientific definition” for 'indirect’. Simply labelling a scheme as ‘indirect’ is not a risk assessment and does not win approval from regulators. All schemes need to be individually assessed in terms of safety and there is no rubber stamp that comes with the word ‘indirect’.

Perhaps you are concerned that the proponents of a scheme may unreasonably claim the endorsement of the US National Research Council by stretching the definition of ‘indirect’? That would be a valid point, but simply the fact that the NRC has taken the position that indirect potable recycling has been shown to be a viable option (and did not extend their position to include direct potable recycling) is not good enough to then automatically label any proposed scheme ‘viable’ or ‘safe’. In fact the NRC specifically stated that all proposed schemes should require thorough individual assessment in terms of the barriers that are proposed. If a scheme is not ‘up to scratch’, then it is simply not up to scratch, -regardless of what words may be used to describe it.

Regarding the UV lamps for the SEQ scheme: I don’t know the specific design specs, however I am aware that the proposal is for ‘advanced oxidation’ (not just UV disinfection). Advanced oxidation schemes require much more powerful UV radiation than plants designed simply for disinfection. You may be interested in an earlier post that I wrote, which includes some discussion of the various types of lamps that are suitable for disinfection vs advanced oxidation.

Yes, there are always data gaps to be filled and opportunities to do things better (I hope to be in a job for a little while yet!). However, I think you will find that the need for more research is routinely stated by researchers in all fields, -it justifies our on-going existence! I certainly wouldn’t knock back that $10...

Mark said...

Hi Stuart,

I'm concerned that there is a gray area when it comes to the labeling of a scheme as 'indirect'. Clearly, regulators and academics in the literature associate indirect PRS with a lower level of risk than the 'direct' variety. And rightly or wrongly this leads to a perception that indirect=good and direct=bad. But it is not clear to me how much un-recycled water it ought to be mixed with, or the residence time it should held for, water quality parameters or whatever else is used to determine that a given scheme is 'up to scratch' and in truth a lower risk than a direct PRS.

Plainly, pumping water to a stream does not in itself guarantee any reduction in risk unless the mechanism that achieves that reduction in risk is clearly identified and controlled. For example, in the Toowoomba case, they thought it necessary to hold treated water for 2-3 years. But in the Brisbane case, holding time will be orders of magnitude less and I haven't heard any scientific explanation why a shorter holding time is suddenly OK.

So what I'm trying to discover is, what this barrier there for, and how good will it be? Obviously, if hypothetically it were to be a lousy barrier in terms of reducing risk, they could drop the pretenses and save a lot of money by adopting a 'direct' PRS. Given that they are implementing this barrier, what scientific evidence justifies its efficacy at the lowest projected dam levels?

Also, are you saying that the US NRC is the body that has endorsed the SEQ scheme? If so do you have any documentation from their assessment?


Stuart Khan said...

Hi Mark,

I think its an unreasonable simplification to suggest that regulators and academics make decisions on the basis of some perception that indirect=good and direct=bad. Its really not as simple as that and all schemes need to be thoroughly assessed in terms of the safety offered by their specific attributes. Safe=good and unsafe=bad is a more appropriate simplification.

Environmental residence and dilution with other waters can often provide some degree of reduction of concentration for dissolved contaminants, but exactly how much reduction that can reasonably be assumed is highly scheme-specific. As you point out, its really not possible to generalise.

In many cases, the environmental barrier can be very effective (the 17 kilometres of Nepean River between Penrith and North Richmond seem to do a reasonable job). However, as you suggest, we have less control over it and less precise understanding of it than we have for some engineered barriers. So if we rely on the environmental buffer, we need to do so with a considerable margin of safety.

Even though Wivenhoe may be close to empty by the time it is supplemented with recycled water, its important to remember that it if the discharge point is sufficiently far upstream of the dam wall, it will still involve a flow of more than 50 kilometres before it reaches the water treatment plant at Mt Crosby. This is more than many ‘unplanned’ IPR schemes which do not include advanced treatment. It is a layer of redundancy consistent with the philosophy of multiple barriers.

No, I am certainly not saying that the US NRC has endorsed the SEQ scheme. I was merely saying that the NRC has given general endorsement of the concept of indirect potable reuse as a viable option. However, this should not be interpreted as an endorsement of any specific scheme since all schemes need to be individually assessed.

Mark said...

Hi Stuart,

I should have been clearer: the perception that indirect was safer/better than direct is one that a member of the public could be expected to draw. The fact that many people who advocate potable recycling refuse to endorse 'direct' PRS reinforces this perception for me. That tells me that there is a world of difference between the two.

Thanks for your patience and I'm sorry to keep asking about this, but I still have real trouble understanding this environmental barrier in terms of how it works. I apprecate that our understanding of the environmental barrier may not be as developed as for the engineered variety, but there is clearly a major distinction between PRS that either have (indirect) or exclude (direct) this environmental barrier. The environmental barrier therefore must be one of the most important of all the barriers. Unfortunately, this barrier more than any other seems to be a mystery, and you acknowledge that its effectiveness is "highly scheme-specific".

Given what you say, how can one guarantee whether the environmental barrier will be effective when one looks at a given scheme? Perhaps someone somewhere has assessed the safety/effectiveness of the environmental barrier in the SEQ case and there is a report floating around? And for an existing system like the Napean example, what do you measure to demonstrate that it is doing a 'reasonable job'? I'm still wondering about the Toowoomba example where the design criteria for the environmental barrier was explicitly stated to be holding time rather than how far the water travelled.


Stuart Khan said...

Hi Mark,

Apologies for slow response. I’ve been a bit distracted these last few days.

In my opinion (and it really is just my opinion and may not be shared by all), the real significance of the apparently important distinction between ‘direct’ potable recycling and ‘indirect’ potable recycling is somewhat historical. As we have discussed, many towns and cities throughout the world have relied upon ‘unplanned’ ‘indirect’ potable recycling for many decades. Water systems like the Mississippi River and Thames Water Valley have been able to safely recycle secondary or tertiary treated municipal sewage effluent by returning it to an effective environmental system before extracting the water again and treating by conventional drinking water treatment processes. All of the early planed indirect potable recycling schemes also relied (and continue to rely) on treatment processes which are perhaps less sophisticated than reverse osmosis and advanced oxidation. As a component of the overall system, the environmental buffer is presumed to be a very important component of these multi-barrier systems. Thus the apparent importance of the distinction.

Regardless of how good more modern advanced treatment systems may be, environmental buffers do still add a degree of additional surety (and they also make the system simpler to balance in terms of water coming in and water going out). If I saw a proposed scheme with a relatively limited environmental buffer, I would expect that the overall safety of the scheme should be balanced (‘swings and roundabouts’), by more effective engineered barriers. It is the safety of the overall multiple-barrier system that matters.

So why do we persist with environmental buffers even when highly effective engineered barriers are planned? I think the best reason is that most environmental buffers do still provide a nice piece of redundancy because they are effective at inactivating pathogens and degrading many chemicals. Yes, they can be highly variable, but all redundant safety is a bonus.

How can one guarantee whether the environmental barrier will be effective when one looks at a given scheme? Obviously for a scheme which currently only exists on paper we can’t test it by taking samples to analyse. Therefore we typically have to rely on environmental modelling methods. There are a wide range of approaches that different people will use. No environmental model is perfect, but some are more sophisticated than others. I have some experience with a modelling approach known as ‘fugacity modelling’ developed (predominantly) by Dr Don Mackay who recently retired from the Canadian Environmental Modelling Centre at Trent University. The trick, of course, is having a good understanding of the capabilities and (especially) limitations of the model. If you understand the limitations, you can make a conservative estimate of how chemicals can be expected to be removed from a well characterised environmental system.

I haven’t been involved with any real analysis of the SEQ proposed system, so am not aware of any assessment that has been undertaken. If something becomes available, I will be certain to take a look at it on this blog. I also don’t know anything about the Toowoomba assessment, however I imagine that in both cases, the environmental barriers would have been considered largely as a ‘bonus’ layer of surety rather than a barrier requiring precise quantitative understanding.

Yes, you are correct that in the case of pathogens (and to a large degree, chemicals), environmental residence time is generally the more significant factor rather than travelling distance. I mentioned my rough estimation of travelling distance above simply because I don’t know the details of how fast water travels through the Wivenhoe system, but I figure that realising that there are numerous kilometres involved would provide at least some perspective.

Mark said...

Thanks for those comments - this environmental barrier is making more sense to me now. Let me surmise that what we have in the SEQ case is an environmental barrier that will be quite effective when the dams have a decent level, but quite probably nowhere near good enough at lower dam levels. The SEQ project team already know this, therefore they are belatedly considering implementing an artificial wetland.

Artificial wetlands thrive on high phos and nitrogen water and can be very effective at removing pathogens and heavy metals (so I read). I wonder though, if anyone has operated an artificial wetland on advance-treated water that has most of the plant food (phos and nitrogen) missing. I suppose they can always add fertilizer.


Jared said...

You mentioned how big (or small) the pore sizes on the filters used in reverse osmosis are. How big is a water molecule? Would it ever be possible to filter by individual water molecules? Also, if the smallest commercially available pore size was scaled up to be the size of a grain of salt, how big in comparison would the smallest disease carrying substance be?

Stuart Khan said...

Hi Jared,

Thanks for asking this question, it was a useful exercise for me to sit down and calculate it through. I think my sums are right, but let me know if you think I’ve made an error anywhere.

The size of a water molecule is a little difficult to define since it is triangular and also the atomic radius of the oxygen atom gives it a very real third dimension. However, according to this website, a water molecule is approximately 2.75 Angstrom (which is 0.000275 micrometres).

Although many people tend to refer to ‘pore sizes’ for water treatment membranes, it is not technically correct in the case of reverse osmosis (RO). The reason is that if you look at an RO membrane under a microscope, you wont actually see any pores. Instead the molecules which pass through the membrane, do so between the gaps between the actual atoms that form the dense polymer matrix. In other words, there are no ‘gaps’ in the actual polymer matrix, but the ‘gaps’ are intrinsic to the polymer matrix itself.

RO membranes are variable depending on their application and what they are designed to do. However, typical water-treatment RO membranes have what we might call an ‘effective pore size’ of around 0.5 nanometres (0.0005 micrometres).

According to this website, a typical grain of salt is 0.5 mm (500 micrometres).

Therefore, the ‘scale-up’ factor from an RO pore to a grain of salt is 1 million (10^6). In other words, we can simply change the units from micrometers to metres.

According to this website, the smallest virus is around 20 nanometres (0.02 micrometres). Scaling this up gives 0.02 metres, -about the size of a soccer ball.

A study recently published in the journal Nature showed that for a prion to be infective, it needs to have a minimum radius of around 10 nm (thus also a diameter of 20 nm). Therefore the smallest infective prion has a particle size about the same as the smallest viruses.

Other infective agents such as bacteria and protozoa (Cryptosporidium and Giardia) are much larger than viruses and prions.

You asked “Would it ever be possible to filter by individual water molecules?” I expect that there is no reason why a membrane could not be designed with such a dense polymer matrix that the “effective pore size” is little more than 3 Angstroms. In fact, such polymers presumably already exist, we just don’t think of them as membranes since they are so impervious. My assumption would be that such a high pressure would need to be applied to pass a reasonable flow of water that it would be a challenge to design a material that was physically strong enough to prevent breakage. But I’m only guessing about this! I have seen some research on new membranes being designed with in-built ‘carbon nanotubes’ to allow the passage of very small molecules. Perhaps this would be the way to go. We need a materials scientist to get a good answer on this one.

AJ said...

Sorry to take this back a bit but I have just tapped in to this amazing blog. Firstly, Stuart I am so impressed by your responses to people's questions, the information provided is in plain English and very informative. The links you provide throughout are greatly appreciated. I have finally found an interesting site dedicated to merde!

rwindsor I completely agree that a debate about water recycling in the ACT needs to come from an informed base and it appears that ACTEW has been slow in providing info on their website. I am still wading through trying to find their rationale for IPU. In fact, in previous reports on ACT's water future they discard supplementing potable water supplies with recycled water, mainly due to anticipated public opposition.
I think the first thing the local govt and ACTEW need to do is detail why an alternative water source is necessary and provide costings, benefits etc of the options.
ACTEW must work harder on an education program to ensure an open debate, resistant to political hijacking and scare campaigns can happen. Stuart, have you thought about offering your services to ACTEW, perhaps set-up an open Q&A forum on the water2water site?

On another note, I am also shy of criticising a Professor's comments but I find it strange that there were clear and factual responses alleviating the majority of the Professor's concerns. Wouldn't Professor Collignon already know the answers to the questions he asked?
The questions are worth asking, but when the questions are published by an expert in this field, I think the majority of people would believe that these 'questions' are serious problems with the technology yet to be resolved. The Professor's piece scared me and I'm an advocate of recycled water. I had to go away and regain my senses.

Stuart Khan said...

Hi AJ,

Thanks for the very kind comments. Dedicated to merde indeed!!

No, I haven’t considered offering such services to ACTEW as you suggest. My area of “expertise” is really limited to issues related chemicals in water and water treatment processes. I’m happy to step somewhat outside of that on this blog as a means of discussing broader issues, but I don’t want to present myself as an expert on things that I am not.

In any case, I quite like the independence of being able to say what I like and such a relationship would presumably jeopardise that freedom. I would be very happy to provide a report or information to ACTEW, but I don’t want to become their ‘mouthpiece’.

Regrading Prof. Collignon, I would politely suggest that he is (irrefutably) an expert in the field of microbial diseases. However, this is a broad field and I expect that his experiences do not extend to advanced municipal water treatment processes. Nonetheless, I am of the opinion that his entry into the public discussion is of significant value and will help to ensure that such issues are closely scrutinised. Debate is healthy and I am confident that Prof. Collignon’s participation will prove to be highly constructive.

Stuart Khan said...

This post has a sequel of sorts, here.

Stuart Khan said...

I was just checking through my calculation in the response to Jared, above. While I still think the calculation is correct, I see that one of the interpretations is not. The sentence “Scaling this up gives 0.02 metres, -about the size of a soccer ball” is clearly wrong. 0.02 metres is more like a golf ball than a soccer ball...

Closer attention required on homework tasks!

Post a Comment