We spent for ever trying to pin this one down as a foundation of the chalk strategy and finally settled along the lines of: abstraction should not reduce chalk stream flows by more than 10% from natural at Q95 (generally late summer), albeit there’ll be places where a less stringent target is acceptable (lower reaches of larger streams that are very heavily modified) and also there will be places where – as targets are neared – habitat improvements could arguably make more difference to ecological health than the last few increments of abstraction reduction.*

(*worth mentioning here that there is an inherent issue with flow targets in chalk streams: where do you measure the impact, given it will vary as you go upstream? Chalk Streams First has proposed abstraction as a % of recharge figure (10%) as a simpler way to limit impact, partly to overcome this issue: this will become relevant further in to this post)

Now, the eminent hydrogeologist Rob Soley has opened Pandora’s sustainability box again and is questioning whether, in the environmental ambition of the draft regional plans, the concept of sustainable abstraction has lost its holistic inclusion of cost and human need and is too narrowly focused on flows alone. This is an important debate and it needs to include all of us. So, this will be a long blog post but a worthwhile read, I hope.

For a while after I joined the WRSE environmental panel advising on the draft regional plan I was completely lost in a sea of acronyms and diagrams of water resource scenarios that looked like the wiring plan of a 1970s Fiat. I didn’t really know what anyone was talking about.

But then one day Tom Perry from the EA produced a spreadsheet of proposed reductions in abstraction for the chalk streams of the Colne and Lea catchments and for me the discussion suddenly morphed from the abstract to the concrete. This was where wiring diagrams met the real world and I realised what a game-changer the regional planning process could be. 

For the first time we were looking at evolving our siloed, localised, groundwater- and gravity-dependent water-supply systems – which inevitably exert a large environmental impact where the supply/demand balance is most out of step – towards better interregional planning, transfer schemes, water recycling and future desalination. 

And not before time. As long ago as the mid 1800s a miller on the River Gade, whilst arguing with the Grand Union Canal and London Spring Water companies about the impacts their abstractions had on his stream* had suggested that the long-term sustainable answer to London’s water supply deficit was to transfer water from north Wales, where it rained a lot and where there weren’t many people. 

(*Impacts they denied, of course. Worth noting too how the GUC and LSWC recruited the best scientific minds of the day to advance their case and argue that, for example, the water came from an infinite subterranean source incapable of exhaustion and completely isolated from the surface!!)

Here we were, 172 years later, finally thinking about the miller’s sensible, but ignored, idea. Though it sounded dryly bureaucratic, I realised the national framework planning process had the potential to bring about truly significant improvements to our really heavily abstracted chalk streams. The deficits to acceptable flows identified by Tom and the EA fairly closely matched those indicated by our Chalk Streams First modelling. We were on the same page and we were talking about truly meaningful reductions. 

But Tom’s first schedule really only covered the chalk streams. Later, when I saw the more global figures for all rivers and waterbodies, I started to worry in the other direction: they seemed so massive. Not from a purely ecological point of view, of course, so much as from a practicable point of view. How on earth would we deliver reductions that amounted to almost 2,500 Ml/d – the more ambitious of the potential scenarios – in the WRSE region alone? And could leak reduction and demand management really take up all the slack until 2050?

To put that 2,500 Ml/d into perspective, Abingdon reservoir could potentially add 400 Ml/d to our regional output. 

To put that into another kind of perspective, the total deficits to acceptable flows across ALL of the 55 chalk-stream catchments which we identified in our CaBA chalk stream A%R report amounted – according to the A%R assessment methodology which is not the same as the EA flow-based methodology – to only 399 Ml/d (excluding the lowermost Lea and Colne) and even that included about 100 Ml/d from chalk streams in the south west and eastern regions. Overall, we had deliberately surveyed most of the chalk streams which we suspected to be under the greatest stress.

The danger seemed obvious: Ofwat would look at the cost of these ambitious scenarios, deem them unaffordable and throw the chalk stream baby out with the “enhanced scenario” bathwater. Although I sat on the WRSE panel, I nevertheless responded to the first iterations of the draft plans, stating the importance of prioritising these proposed reductions so that we didn’t mix the “essential” up with the “to-no-great-gain” and the importance of planning / delivering the reductions in such a way that the investment added up to conspicuous improvements. 

The last thing we want is to expend a great deal of effort and money to no great or obvious effect, undermining the case for environmental restoration in the first instance. It would be perfectly possible to. Arguably this is what has happened thus far, where abstraction reductions have been made in a piecemeal way, often offset by increases within the same catchments, so not actually amounting to net reductions at all, consequently allowing a narrative to develop that abstraction reductions don’t really yield much of an increase in river flows, or not one that you’d notice, or not one that is justifiable given the cost. 

Recently, Rob Soley has been leading a bit of a kickback campaign questioning the sense of the proposed reductions in the national framework. Of course Rob’s concern may include a worry that overblown / poorly planned abstraction reductions could run the danger of ruling ALL abstraction reductions out of court: but mostly Rob is worried about turning our backs on a valuable water resource (the chalk aquifer) at exactly the time when water resources are coming under greater stress from a changing climate and when the cost of living is putting a squeeze on the nation’s finances: all to what Rob argues will be arguably of little ecological benefit. 

So, I instinctively want to disagree with him, but I find myself partly agreeing with him, even if we are coming from different points of view.

Rob gave a webinar presentation on this subject last week to the Geological Society’s groundwater forum.  The audience was – as you may have guessed – comprised almost entirely of hydrogeologists. There was much agreement with Rob’s argument and so the kickback will be well supported by experts, I suspect. You can watch the presentation HERE, but I’ve written a layman’s précis of the argument in blue below. 

Even though Rob apparently argues for something like the exact opposite of what we (ie. Chalk Streams First) have been arguing for (the restoration of flows to headwater chalk streams and the chalk-stream tributaries of systems like the Colne) I am very keen to find common ground, for the reasons expressed above: unless we get these abstraction reductions right – properly planned and judiciously delivered in the right places, at meaningful scales and allied with habitat and water quality improvements – we will undermine the case for them altogether and we’ll be back at square one. 

This is a very finely balanced argument: for the first time ever the regulators are planning a national re-plumbing that allies public supply resilience with environmental restoration. Fantastic! But it’s expensive work and we have to get best value from the investment. And we also have to make a powerful case for the continuing inclusion of environmental restoration in the definition of best value: it will fall out the door if we spend money unwisely.

In an effort to find that common ground I have made notes (in red) against Rob’s argument (in blue), pushing back here and there but agreeing where I think it is right to agree. Rob says “hydrogeologists unite”, but in the end we want hydrogeologists, environmental regulators and environmentalists, Ofwat accountants and our political leaders all to unite, for without common agreement on the problem and the best solutions we’ll be on an adversarial merry-go-round for ever.

I will post comments from anyone who wishes to join the debate over Christmas.

Précis of Rob Soley’s webinar presentation, organised by the groundwater modellers’ forum of the geological society:

Rob’s words in blue.

(discussion/comments by CRW in red)

Rob emphasised that he is “standing on his own soapbox” to air concerns about proposed potentially enormous groundwater abstraction reductions in the national framework draft plans. The elephant in the room, he says, is vaguely greenwashed and represents water resources folly in the pursuit of questionable ecological benefits.

‘Sustainable’, Rob argues, once encompassed ‘environment + people + carbon + money’. Now, river-flow standards are being pursued regardless of benefit or cost and we are on the verge of giving up a huge amount of groundwater storage. The altar of naturalness has made smart management a ‘dirty’ concept.

(Worth remembering that this pursuit of river-flow standards is only a potential pursuit at the moment: right now many chalk streams around London and Cambridge and into Kent are very heavily abstracted – abstraction totals of between 30% to 60% of recharge from rainfall – and will continue to be for years to come, no matter what. 

I agree that ‘smart management’ should not be sacrificed on the ‘altar of naturalness’)

In the early noughties the EA developed an abstraction-pressure screening process, consisting of a natural-flow duration curve, below which hung an impacted but idealised flow threshold (the Environmental Flow Indicator or EFI): in the places where the actual flows dipped below the idealised impacted flows a sustainability-assessment screening process was conducted which considered the environment + people + carbon + money. Over the past 20 years that definition of sustainable has changed, until now it considers only the environment, with the target an imperative. 

(This isn’t quite true. Rather, over time the weighting and even definition of these values has changed and therefore the balance within the definition of sustainable has changed. For example, what people are prepared to pay for has changed as their feelings about the importance of the natural world have changed: the definition still includes people, what they care about as well as the cost of water from their taps. The carbon cost may change too and certainly will when humankind cracks nuclear fusion, at which point the carbon costs of groundwater abstraction will become the absence of wetlands, and the definition of sustainable will once again have to shift to reflect this.)

And the target has been tightened.

The risk is that many of the reductions will not deliver the anticipated benefits. The process of national re-plumbing is going to be hugely expensive. With climate change our lows flows are likely to fall and so we need to consider how to make our environments more resilient.

Previously the map of groundwater (GW) abstraction impacts included the good (green), bad (amber) and ugly (red). The desire now is to turn the whole map green, by reducing the current national total groundwater abstraction total from 6,300 Ml/d to 3,500 Ml/d. We will attempt this by moving from a system of local water for local people and building more pump transfers, desalination plants, surface reservoirs: and all just to stand still (ie. reduce g’water abstraction) and before we address sustainability of water supplies.

(The environmental impact of “local water for local people” varies greatly depending on population density versus rainfall. For the chalk streams around London that ratio and the ‘local for local’ principle condemns the rivers to chronic and significant flow depletion. Local systems are gravity-dependent and thus subsidised by the environment while that environment gets bypassed in the supply-discharge loop. There is always a cost: it’s just a matter of the degree to which nature picks up the bill versus what we are prepared to pay in order to retain/restore nature.)

Switching off GW abstractions on this scale will sterilise our access to a vital resource while the environment will continue to be stressed by droughts and climate change. 

(Worth pointing out here that the Chalk Streams First concept is not proposing to “sterilise access to groundwater”: as a concept CSF still sees the aquifer as a water-resource reservoir. But CSF argues for using the stream as the means of delivery, rather than bypassing it: hence the name Chalk Streams First. Note that later in his talk Rob tries to show the difference between groundwater abstractions that subtract directly (both temporally and spatially) from the stream flows and those that subtract from distant storage with a delayed impact.)

We should carefully prioritise GW-abstraction reductions, alongside better management and habitat creation, yielding quicker, cheaper, and better ecological outcomes. .

(Conditionally agree: quicker, cheaper, better than ? … abstraction reductions per se, or poorly planned abstraction reductions? It shouldn’t be ‘either, or’.)

How did we get to these levels of proposed reductions?

The EA’s first national screening process was the CAMS ledger, incorporating 1200 assessment points of flows often at flow gauging station to improve confidence, and a screening process of natural versus recent actual and fully licensed flows.

The water framework directive (WFD) added 8,300 assessment points, with concomitant upstream migration of assessment. The assessment tool became the Water Resources GIS (WRGIS) linked to CAMS ledger spreadsheets. The screening process fed into WINEP.

WRGIS + the “FIXIT” tool fed into the national framework, looking at places where flows are below the indicator and calculating the reduction in abstraction needed to meet the target. It worked iteratively down the catchment, waterbody by waterbody. This developed the numbers the EA sent to water companies. Initial estimates of the groundwater abstraction  reductions required (to meet the ‘Business as Usual’ environmental flow standards without any water body exclusions and before accounting for climate change) indicated a total reduction of 2713 Ml/d from 6228 Ml/d (recent actual) (and a reduction of 4682 Ml/d from 11300 Ml/d fully licensed).

The chalk stream strategy has catalysed the further upstream migration of assessment: now quite seriously on the table is the need to protect chalk springs “where they are”. Protecting perennial springheads where the Q99 is close to zero: 5% of no flow is no abstraction. The conclusion of this is that “this is not an aquifer for people anymore”.

(The chalk strategy made the point that flows at the waterbody boundary aren’t necessarily representative of flows higher up the catchment, especially when the assessment point is below a significant discharge. There are plenty of examples, but the Ivel, for one, is assessed downstream of the Pix tributary and a sewer discharge: it is assessed as “supports good” for flow when the upper, chalk stream reach is very frequently dry. Concluding that arguments for the river’s restoration amount to “this is not an aquifer for people” arguably does not reflect how much people have come to care about their local chalk streams. In fact, the chalk strategy has proposed that we need strategic abstraction realignment in the places where abstraction is currently 20, 30, 40 or even 50+%* of recharge. This does not have to mean a 100% net loss to supply at all, but rather an evolution away from old-school infrastructure towards intelligent, conjunctive systems that allow the streams to flow whilst also utilising the water for public supply. This is perfectly possible, but it will cost some money.

*Interestingly, Rob agrees that A%R is perhaps a better way to assess abstraction impact and reduction prioritisation, because driving flow targets higher and higher up the catchment leads to increasingly stringent and unrealistic abstraction reductions: this is because chalk streams are naturally ephemeral in their upper reaches and trying to achieve no more than a 10% reduction from a tiny amount of flow or even no flow effectively amounts to zero abstraction: the meaning of his penultimate line in the preceding blue paragraph.)

Also the approach to the environmental flow regulation has tightened. We’ve gone from an assessment where sustainable included considering things more broadly to one where you just have to meet the flow standard. There is a mechanism of local override but it is usually impossible to make that stick because the data isn’t there and it is impossible to move away from the precautionary standard. 

(This is rather as if the precautionary principle has triumphed and we are in world of fully naturalised flows. It hasn’t and we aren’t.)

The national framework has tightened this yet further, so that now there is a notion that all chalk streams should be ASB3, all salmon streams should be ASB3. There hasn’t been a great deal of evidence presented to argue why that should be. There’s just the notion that these things are crown jewels, “without considering whether those standards are what the ecology really needs”.   (without clear impact evidence).

(Abstraction sensitivity bands (ASBs), as the chalk stream strategy shows, evolved out of flow targets which were originally determined by river type and expected macrophyte, invertebrate and fish communities (and their dependence on flow for health). ASBs , therefore, should be determined by a river’s type / potential, not current condition. But ASB designations have become muddled in places by a somewhat automated process that managed, for example, to designate streams like the upper River Nar as ASB1 (least sensitive). Every single chalk stream in England once contained salmon and nearly all still contain trout and often sea trout (the presence of salmonids, especially migratory salmonids should mean automatic ASB3 designation). Chalk streams’ natural macrophyte and invertebrate communities are also rheophilic (flow loving) and if the actual communities are now different (ie dominance of chironomid lava and cyprinids) this will ALWAYS be the result of anthropogenic modifications, usually a combination of abstraction, impoundments, canalisation and pollution. Therefore, there is a very good case for arguing that all chalk streams should be ASB3 because, the ecology of a chalk stream depends on flow. However, in the chalk strategy we have also conceded that ASB3 might not be appropriate (ie it would signal unjustifiably large abstraction reductions) in the lower reaches of big systems, which would benefit from upstream flow recovery anyway, or systems that are so heavily modified as to make restoration of ASB3 flows rather pointless, because the bigger constraints are immovable navigation locks or such-like.) 

All of that is true at the moment however bonkers the cost.

(The total reductions mooted at the moment do appear to have been generated by a computerised exercise (probably unavoidable as a first pass) without the important next step of prioritising and strategising. This next step is really important.)

And at this stage of the WRMP all the water companies are saying “okay, we’ll put it into the plans if that is what you want” and that’s going to go round the hoop until it gets to Ofwat, who may well tell us that we can’t afford it.

We need to get back to a more holistic, less siloed approach (flow + m’ment + habitat + water quality) to deliver real local biodiversity benefits more quickly and affordably “without trashing our resilient GW supplies on a massive scale” “with some understanding nationally of how much money we want to spend on this problem”. 

In parallel with this story there has been a development of regional groundwater and river-flow models: you can use these models to look locally at each stream cell and the river waterbody outflow points to check the compliance with these standards. 

There are lots of examples where both modelling and experience show ‘disappointing’ returns at low flows relative to the amount of abstraction reduction.

(For various complex reasons a unit of abstraction (or abstraction reduction) tends to have a greater impact on high flows than low: ie. if you reduce abstraction by a given unit (say 1 Ml/d) the impact on flows will be well over 100% at high flows and as little as 30% to 20% at low flows*.

*The impact on flows is expressed as an impact at the time, but this is not the full picture, because the impact is not a direct one but takes place via the impact on groundwater levels. A unit of water abstracted at low flow periods may have a smaller impact on the force which drives flow (because groundwater levels are low), but it nevertheless removes that unit from the total flow discharged from the valley: in fact it creates a debt to future flow, by delaying the rise of groundwater levels during recharge. This is why the oft repeated idea that abstraction has no impact on flows once groundwater levels have fallen below the river bed – and the river has dried – is misleading. It has an impact on future flows just like accumulated debt has an impact on future spending power.)

There is a good example on the south coast near Chichester of a chalk source on the River Wallington, between the River Meon and Bedhampton springs. Outflow points from the aquifer are at different elevations. The Wallington is relatively higher than either the Meon which gets baseflow most of the year round where it flows off the chalk and the Bedhampton springs at the base of the chalk. The impacts of the abstraction show that most of the impacts on the Wallington occur at high flows when GWLs are high and the chalk is spilling. On the Meon the impacts are constant through the year and at Bedhampton the impacts at low flows are actually higher than at high flows (when the aquifer overflow is pouring out of the Wallington). All because of the different elevations of the drainage points from the aquifer. The variable pattern of impact is true in other settings, for a combination of additional reasons to do with the non-linear behaviour of the aquifer, the variations of permeability with depth, winterbourne path-lengths changing as the aquifer fills or drains, the layering of the aquifer, evapotranspiration rates as groundwater comes close to the surface etc etc.

(The impact of abstraction is shared between three sources of discharge from the aquifer: the Wallington, the Meon and the Bedhampton springs: very simply you could see them as three vertical columns of holes with different uppermost elevations in the side of a bucket: the Wallington holes extend much higher up and as the water runs out of the bucket so the flow deserts these holes first, whereupon the impact of the abstraction becomes progressively more exerted upon the other two columns of holes. In this case, therefore, the impact of abstraction on the Meon and Bedhampton, but especially the latter is actually higher at low flows than high, because the impact is shared across three streams, one of which switches off almost completely.)

Therefore it is important to use these latest models to “wise-up” the EA tools such as the CAMS ledgers and the water resources GIS, because if not WINEP decisions are based on calculations that may not reflect reality.

Recently, and in response to the chalk strategy, the EA has proposed assessing flows at the perennial head of chalk streams, called Point X. The Fixit tool can be used to determine how much abstraction reduction would be needed to meet the flow targets at all the points of perennial outflow. 

Looking at the Cam Bedford Ouse model: to meet the WFD flow targets at all of the waterbody boundary assessment points abstraction would have to be reduced from the recent actual volume of 144 Ml/d down to 61 Ml/d. While to meet the same target at the springs (rather than the waterbody outfall) abstraction would have to come down to 18 Ml/d.

In the Quy sub-catchment currently 14 Ml/d of g’water abstraction leaves a non-compliance deficit of 7 Ml/d at the outfall. Plug in the reduction suggested for meeting the outflow compliance and there is still non-compliance in the headwaters. Plug in ASB3 numbers and this reduces, but not completely. Effectively you have to turn off abstraction to remove non-compliance at the spring-heads. This then creates a 5 Ml/d surplus (ref the target flows) at the outfall. 

(Worth noting that the generic CSF lumped parameter figures more or less match the sophisticated modelling figures in this case: reduce the abstraction to the point where flows re-naturalise and 12 out of 14 Ml/d manifest as surface flows at the outfall of the catchment. The Quy is on the Bedford Ouse system and is linked, therefore, to Grafham reservoir and in time to the proposed Fenland reservoir. In theory, therefore, we could restore flows to the Quy as in the CSF model, with a net average loss to supply of only 14%, albeit there would be additional costs involved in treatment and pumping and infrastructure. The proposal is not, therefore, to cease groundwater abstraction and source 100% of the reduction from completely different sources, but rather greatly reduce groundwater abstraction, find the 20% net loss through methods like efficiency improvements and then re-plumb in order to capture and use the recovered flow lower down the system. It is not accurate to represent the chalk strategy or CSF proposals as completely hands-off aquifers.)

Work by APEM for Severn Trent assessed the health of water bodies across Midlands against degrees of variance from natural flow. Pooled evidence suggests that at Q70 most sites with impacts of less than 50% are at good or better, (against EFI allowable reductions of 15% to 24%). This suggests we could adopt a less precautionary approach.

(I’m not convinced this case provides a useful application to chalk streams: were the assessed rivers in the Midlands salmonid streams with rheophilic ecologies? Rob’s main point though was that pooled hydro-ecological modelling should be carried out for chalk streams too, in order to see what the evidence suggests across the range.) 

The CSF model is a very simple lumped model. But “it fits really well”. Even with a simple non-linear representation of the head driving flow into a chalk stream, you get an indication of lower flow recovery at low flows compared to high flows. CSF is a good example of a modelling tool that is a lot cheaper and quicker to run but it doesn’t incorporate an understanding of the location of individual abstractions and how that influences their different impacts on river flows, which is why we can’t rely on it for making some of the really big decisions.

(Agreed: CSF is not designed as a decision-making tool at a micro-scale. It is partly designed to bridge the knowledge gap between hydrogeologists and the layman, in order to democratise the discussion. It is also designed to provide quick and simple assessments of abstraction impact. The fact that it fits really well again and again must also say something!)

An example of where the CSF model would not yield the requisite level of detail and where the impacts of abstraction reduction may not be justified:

Upper Itchen, “Source A”: on the syncline between the Wey and Candover, a long way from the perennial spring heads. Where and when does this abstraction impact flows in the Rivers Wey, Candover and Dever? A steady-state abstraction impacts these three rivers with a similar pattern of variance as the Wallington Source: the biggest impact is on the Wey, almost 7Ml/d at high flow, down to 1Ml/d at low flows. The Dever impact on the other hand is much more steady 1 to 0.5 Ml/d. Added all together the impacts are 18 Ml/d at high flows and 3 Ml/d at low. In other words, at low flows “you might not notice it that much”.

(This is where Rob is trying to draw a distinction between chalk groundwater sources which subtract directly /almost immediately from stream flows and those that subtract from distant storage with a delayed impact. This upper Itchen source, like the World’s End source impacting the Meon, Wallington and Havant springs uses groundwater storage with, as Rob argues, minimised environmental impact. 

The reason: “this is a source that develops drawdown in the summer: that’s a good thing. It’s not mining. It’s taking water from storage rather than directly from river flows.” The process here is the same as why good groundwater augmentation schemes work … they provide immediate summer low flow gain in discharging to the streams (providing it doesn’t leak back in) because the impact of the abstraction is not instant – it is offset to the next recharge season.

Candover augmentation scheme also scheduled for decommission: in the past it has supplied up to 22 Ml/d for abstraction in droughts. It can also supply on a more frequent basis enough water to keep flows close to natural in the Candover. So, the aim of having flows close to natural can be achieved without turning off all these abstractions. But there is a mantra now that we have to restore everything “to natural” and that we can only achieve that via switching abstractions off.

(I agree with Rob that the concept of a natural aquifer is potentially a distraction from finding workable and good conjunctive solutions. Combined with downstream surface-water abstraction off-takes, the upper-catchment groundwater storage here (ie. aquifer) surely has the potential to deliver almost naturalised flow volumes via the stream: I don’t think we should rule out ideas like this. Augmentation has a bad reputation but we should question why, if it gives us a way to retain water supply resilience and ensure close-to-natural flows. We should try to arrange some open discussions between Rob, the CaBA chalk stream group, Chalk Streams First, and Natural England / Environment Agency on this topic.)

The upstream abstraction reductions of 11 Ml/d Source A 3 Ml/d Source B 26 Ml/d augmentation add up to 40 Ml/d. Plus another 40 Ml/d from downstream = 80 Ml/d of resilient drought supplies lost only to be replaced by effluent recycling / desalination / distant reservoirs and pumping: this is untried, costly (hundreds of £millions) and carbon heavy. The “environmental benefits uncertain”. Protected areas need active management to protect them … let’s not leave them entirely at the mercy of climate change without groundwater augmentation.

(As above: there are ways to reduce abstraction impacts without creating a 100% net loss to supply.)

Another example of where the cost of abstraction reduction may not be justified:

Avon: most of the Avon is within CSMG flow targets, except for the final 1km, where the flow deficit is -70 Ml/d, almost the entire supply for Bournemouth. To replace this will increase low-flow abstraction from the Stour, a river with less resilient base flows, via stream support, relocation of 30 Ml/d wastewater recycling, plus aquifer storage and recovery schemes, all at a cost of £300 million.

There is little evidence of ecological impact on this reach. None in terms of invertebrates. In terms of fish migration there is no clear correlation between fish migrating up the Avon versus flow or temperature, although there is clearly a spate flow response (Rob showed a chart of 2020 salmon runs into the Avon). In addition 2020 saw far greater returns than any of the previous years since 2006 through which time abstraction pressures have been constant.

(Salmon return numbers and timing depend on so many factors and the overarching control is undoubtedly conditions at sea, especially sea temperature impacted by global warming (unnatural) and the North Atlantic multi-decadal oscillation (natural). Our job is to ensure the natal streams produce as many salmon as possible: so the real question is whether a 70 Ml/d reduction in flows is likely to inhibit salmon from entering the system or reduce the salmon-producing potential of that system? At average flows in a base-flow supported system like the Avon, maybe not. In drought conditions, maybe yes).

Q: is the legal requirement to reach the SAC target worth 300 million? 

(I feel Rob has a point here, especially because the Stour is also a salmon river. There may be cheaper or better ways to overcome the real risk to returning salmon trapped in the estuary in drought conditions: for example, salmon are attracted to velocity and turbulence (as well as the scent of fresh rainfall) when running upstream and there must be ways to create both in the lower 1km?).

Ref the future we know flows are on a downward trend with climate-change impacts. Where will supplies come from if we abandon g’water?

What do g’water modellers need to do?

• need to make their models fit. CSF fits even if g’water modellers don’t believe the concepts*. But regional models which include the concepts don’t always fit very well. Need local refinements for credibility.

*(maybe the concepts fit too!?)

• have to engage in the model wising-up processes to improve the national WRGIS data

• need to use the models to give stakeholders engagement: make model easy to share and improve

• need to talk about ecological benefits versus cost and challenge the green myths which say that:

• vast GWABS reductions are good (unwise, costly, uncertain benefits). 

(Agree, we need to be strategic and judicious to avoid the risk of undermining the whole project.)

• GW augmentation a bad thing (actually supports river flow and abstraction). 

(Agree, we need to consider the role of using the stream to deliver water: this is not augmentation as such, but a different way to plumb water supplies. Even pure augmentation has its place.)

• we’ll easily use less water (a very uncertain outcome).

• we’ll easily reduce leakage (without digging up cities?).

(Agree, both demand and leak reduction offer uncertain, untested savings at this stage. We need schemes that provide certain relief to the highly stressed chalk streams and we need them soon!)

• wetlands will improve drought flows (may just increase evapotranspiration).

(Yes, but they are still ecologically desirable and they lock up carbon.)

• protected areas need to be left as “natural” (no such thing. active management is a good thing).

(Agree, we need to actively manage.)

• make the most of our aquifers for people and the environment. Embrace well located groundwater abstraction reductions. Encourage effective GW river support schemes.

(Agree.)

• work across organisations to deliver biodiversity benefits now.

(Wholly agree)

• reclaim a more holistic meaning of the word sustainable.

(Yes, but values change. Holistic must include those changes too.)

In summary, Rob is saying that we should be very wary about abandoning our climatically resilient (mostly chalk) groundwater supplies and investing many millions in an untried, untested, carbon-costly re-plumbing exercise, all in the name of uncertain ecological benefits. Instead, we should keep most of what we have in place, make reductions in a few locations where they will make a big difference to flows and ecology, and instead we should work on habitat improvements which will give us better outcomes for less money.

And I agree with some of that, certainly that we should carefully prioritise abstraction reductions so that they deliver tangible gains, something that will require transparency, knowledge-sharing and democratic discussion. But it is not a question of abandoning chalk aquifers so much as allowing chalk streams to become the means of delivery of water from the aquifer to the tap, meaning that restoring flows should amount to much smaller net reductions in public water-supply than the bare modelling headlines suggest. Habitat improvements can make a vast difference, but only once flows are back towards an acceptable level: in many chalk streams around London, Cambridgeshire and Kent where abstraction is over 20% and up to 50%+ of average recharge, flows are not acceptable. But all these deficits added together come to a manageable volume -400Ml/d) especially when it is possible to reclaim much of that water lower down the catchments.