Drought, Leakage and “Forever Chemicals”: Is This Really How We Should Protect the Thames?

When a river is at its lowest, that may be precisely the time when we should be most careful about what we put back into it.

An interesting—and rather worrying—story appeared in The Guardian this morning.

It concerns Thames Water's proposed Teddington Direct River Abstraction scheme, designed to help maintain London's water supplies during periods of drought.

At first sight, the idea sounds quite clever.

Take up to 75 million litres of water a day from the Thames. Transfer that water towards the Lee Valley reservoirs, where it can eventually become part of London's drinking-water supply. At the same time, replace the water removed from the river with highly treated wastewater from the Mogden sewage treatment works.

The river level is maintained.

London receives additional water.

Wastewater is recycled rather than simply discarded.

It sounds, superficially at least, rather green.

Unfortunately, there is a problem.

According to technical pilot data reported by The Guardian, concentrations of the “forever chemical” PFOS reached 6.3 nanograms per litre in one test and 8.7 nanograms per litre in another.

The existing statutory environmental quality standard for PFOS in fresh surface water is 0.65 nanograms per litre. That makes the two pilot concentrations roughly 10 and 13 times the standard respectively.

And that raises a rather important question:

Are we solving a shortage of water by creating a water-quality problem?

First, the drought problem is genuine

It would be too easy to turn this into a simple attack on Thames Water and conclude that the whole scheme is ridiculous.

That would miss an important point.

South-east England genuinely needs to think about where its water will come from in future.

Thames Water says its region faces increasing pressure from population growth and climate change and forecasts a potential supply shortfall of around one billion litres per day by 2050. Its Teddington scheme is intended to provide up to 75 million litres a day during drought conditions.

And 2026 has given us a fairly good demonstration of why drought planning matters.

At the end of July, London's reservoirs were below their normal level for the time of year. River flows were reduced. The Environment Agency has subsequently restricted lock operation around Teddington partly to conserve water.

We cannot simply say:

“It will rain eventually.”

Climate adaptation means preparing infrastructure for the weather we are increasingly likely to experience.

The difficult question is how.

What exactly is Thames Water proposing?

There is an important distinction worth making because headlines about “putting sewage into the Thames” can give a slightly misleading impression.

This is not a proposal to pump untreated sewage into the river and then immediately pipe that same water into people's taps.

Thames Water's plan involves taking river water near Teddington and transferring it towards the Lee Valley reservoirs.

Meanwhile, final effluent from the Mogden sewage treatment works would undergo an additional tertiary treatment stage before being discharged into the Thames to compensate for the water being removed.

Water reuse itself is therefore not the scandal.

Indeed, recycling water will almost certainly become an increasingly important part of living in a hotter and more water-stressed Britain.

The question is whether the treatment is sufficiently good.

And that is where the PFAS results become important.

What are these “forever chemicals”?

PFAS stands for per- and polyfluoroalkyl substances.

It describes a very large family of synthetic chemicals that have been used in numerous industrial and consumer applications because of their useful resistance to heat, oil, stains and water.

Unfortunately, that chemical stability is also their environmental problem.

Many PFAS compounds degrade extremely slowly.

Some can accumulate in living organisms.

Some are associated with health concerns.

And conventional sewage-treatment plants were never designed specifically to remove them.

The Environment Agency says PFAS are now frequently detected in rivers, estuaries and groundwater.

One particular compound—PFOS, or perfluorooctane sulfonate—already has a statutory Environmental Quality Standard.

And this is where the pilot becomes uncomfortable.

What does “13 times the legal limit” actually mean?

We need to be careful with the terminology here.

The 0.65 ng/L figure is an Environmental Quality Standard for surface water. It is not the same thing as saying that somebody taking one mouthful of river water containing 0.66 ng/L PFOS is suddenly being poisoned.

Environmental standards are designed to protect ecosystems and human health from longer-term environmental exposure.

Nevertheless, the comparison matters enormously.

The pilot results reported by The Guardian were:

Measurement	PFOS
Existing statutory fresh-water EQS	0.65 ng/L
Pilot test 1	6.3 ng/L
Pilot test 2	8.7 ng/L
Approximate exceedance	9.7x and 13.4x

The newspaper reports that PFOS “breakthrough” through the treatment system occurred in less than a month and that maintaining compliance would require frequent regeneration, reactivation or replacement of carbon filtration.

That doesn't necessarily prove that the eventual full-scale plant will discharge water at those concentrations.

In fact, Thames Water makes precisely that point.

It says the pilot was intended to test different technologies and refine the final plant, that the results should not be regarded as representative of the eventual discharge, and that the project cannot operate without an Environment Agency environmental permit.

That is an important defence.

A pilot plant is supposed to identify problems.

If every experimental result had to be perfect, there would be little point running experiments.

But identifying a problem is only useful if the eventual engineering convincingly solves it.

And the science may actually be moving towards tighter PFAS standards

There is another particularly interesting aspect to this story.

In January 2026, the Environment Agency published new research into thresholds for four PFAS compounds.

It emphasised that the new values do not replace the existing statutory PFOS standard, so we shouldn't describe them as new legal limits.

However, the research derived an equivalent PFOS water concentration of just 0.015 ng/L as part of an approach intended to assess the combined effects of four PFAS chemicals.

That is striking.

The current statutory PFOS standard is 0.65 ng/L.

The emerging science is looking at concentrations much lower than that in considering combined PFAS exposure.

In other words, engineers designing water-treatment plants probably shouldn't assume that today's regulatory standard will necessarily represent tomorrow's expectations.

The river is most vulnerable when the scheme would operate

There is something else that concerns me.

The Teddington scheme is specifically designed to operate during drought.

Thames Water says it expects the scheme to be needed roughly every two years, normally between late summer and late autumn.

But think about what a river looks like during a severe drought.

There is less water.

Flow is slower.

Water temperatures may be higher.

Dissolved oxygen can become more problematic.

Pollutants receive less dilution.

Aquatic organisms may already be stressed.

Indeed, this August the Environment Agency confirmed a blue-green algal bloom affecting part of the lower non-tidal Thames between Sunbury and Teddington.

That doesn't mean the proposed recycling scheme caused the bloom—it isn't operational.

But it illustrates the important principle.

The ecological conditions under which drought infrastructure operates may also be the conditions under which a river is least able to tolerate additional environmental stress.

“Maintaining the water level” therefore isn't necessarily the same thing as maintaining the health of the river.

Chemistry matters too.

Then there is the question of iron

PFOS wasn't the only issue reported from the pilot.

According to The Guardian, technical documents also identified high iron concentrations that were considered capable of having a detrimental effect on the river.

Again, there may be engineering solutions.

But it reminds us that wastewater treatment isn't simply about producing water that looks clean.

Clear water may contain substances that we cannot see:

• PFAS;
• pesticides;
• pharmaceuticals;
• nutrients;
• dissolved metals;
• micro-contaminants;
• and potentially many other substances arriving from homes and industry.

Modern water recycling increasingly requires us to think at concentrations measured in nanograms per litre.

One nanogram is one billionth of a gram.

Our analytical chemistry has become extraordinarily good at finding these substances.

The next challenge is becoming equally good at removing—or preferably preventing—them.

The elephant in the room: leakage

And then we come to the statistic in the Guardian report that jumped out at me.

The proposed scheme would provide:

up to 75 million litres per day.

The Guardian reports Thames Water losing approximately:

571 million litres per day through leakage.

Put those numbers beside each other.

The maximum output from the proposed drought scheme is only about 13% of that leakage figure.

Or look at it the other way around:

The leakage figure is around 7.6 times the capacity of the Teddington scheme.

That doesn't mean fixing leaks eliminates the need for new water resources.

It doesn't.

No distribution network can realistically achieve zero leakage, and as leakage becomes smaller, finding each additional leak can become progressively harder and more expensive.

Population is also increasing.

Climate patterns are changing.

We need storage, new supplies, conservation and leakage reduction.

To its credit, Thames Water says it has reduced leakage by more than 15% since 2019/20, is fixing over 1,000 leaks a week and is deploying acoustic sensors, satellite technology and AI to detect more losses.

But the comparison is still impossible to ignore.

Before asking customers to accept complicated new infrastructure to obtain another 75 million litres a day, it is entirely reasonable for customers to ask:

How much more of the water we already have can be prevented from disappearing into the ground?

Recycling isn't automatically green

This connects with a theme I keep returning to in Going Green.

We have a tendency to assume that the word “recycling” automatically means environmentally good.

It doesn't.

Recycling is a process.

Its environmental value depends on what the process involves.

Activated-carbon treatment needs carbon media that eventually has to be regenerated or replaced.

Advanced filtration requires energy.

Reverse osmosis, if used, consumes electricity and creates a concentrated waste stream.

Pumps require power.

Treatment plants contain concrete, steel and chemicals.

Tunnels have embodied carbon.

None of that automatically makes water recycling a bad idea.

But it does mean we should assess the whole system, not merely attach the word “recycled” to the water and stop asking questions.

The greenest solution is the one that produces the greatest environmental benefit for the resources consumed.

Perhaps we need a water hierarchy

We are familiar with the waste hierarchy:

Reduce → Reuse → Recycle.

Perhaps our water system needs something similar.

I would put it approximately like this:

1. Prevent pollution entering water in the first place.
2. Reduce unnecessary water consumption.
3. Find and repair avoidable leakage.
4. Capture and store water when it is plentiful.
5. Reuse water locally where appropriate.
6. Recycle wastewater using treatment capable of removing modern contaminants.
7. Develop additional abstraction and supply infrastructure where genuinely necessary.

The exact order will differ according to circumstances.

But the principle matters.

We shouldn't automatically jump to the most complicated engineering solution before getting as much value as possible from the water already available.

PFAS is not really a sewage-treatment problem

There is an even bigger question lurking behind this story.

Why is PFAS in the wastewater in the first place?

Water companies did not manufacture most of these chemicals.

Sewage works receive what society sends to them.

Trying to remove persistent chemicals at the end of the pipe can become extraordinarily difficult and expensive.

Thames Water itself argues that PFAS should be banned and that manufacturers rather than water customers should ultimately pay for their removal from the environment.

On this point, I think there is an important principle worth considering.

Pollution is far easier to prevent at source than to extract from billions of litres of water later.

Imagine deliberately adding a tiny quantity of a persistent chemical to millions of litres of water and then attempting to extract every molecule again downstream.

That is essentially the technological problem we have created for ourselves.

This is why environmental regulation upstream matters.

What can the rest of us actually do?

It would be rather unfair to finish another environmental article by telling householders that they can solve all of this by turning the tap off while brushing their teeth.

They can't.

PFAS pollution requires industrial action, government regulation, monitoring and investment in treatment technology.

But household water conservation still matters.

I've written before about collecting water that would otherwise disappear down the plughole, using rainwater in the garden and thinking more carefully about whether we really need drinking-quality water for every household task.

There are plenty of small reductions available:

collect the cold water that runs while waiting for the shower to warm;

repair dripping taps and leaking toilets;

use full washing-machine and dishwasher loads;

use rainwater for garden watering where possible;

let lawns go dormant rather than trying to keep them emerald green through drought;

and ask whether every litre coming from the mains really needs to be used only once.

One household doesn't transform the Thames.

Millions of households changing demand can.

But the responsibility must be shared.

Consumers should conserve water.

Water companies should reduce leakage.

Government should control persistent pollutants.

Industry should take responsibility for chemicals that remain in the environment for decades.

And new infrastructure should have to demonstrate that it solves more problems than it creates.

Five questions I would like answered before this scheme proceeds

For me, the Guardian investigation doesn't prove that Teddington water recycling can never work.

But it does justify some very demanding questions.

What PFAS concentrations can the final plant reliably achieve—not for a few days, but throughout prolonged operation?

What happens when activated-carbon filters become saturated?

How quickly can operators detect breakthrough and stop a non-compliant discharge?

What other persistent chemicals and pharmaceuticals are being monitored?

And how does the environmental and financial cost of producing each additional litre compare with saving another litre through leakage reduction, efficiency or storage?

Those aren't anti-technology questions.

They are exactly the sort of questions we should ask of environmental technology.

Going Green must mean more than keeping the taps running

There is a temptation during every environmental crisis to search for one great technological solution.

More solar panels.

More batteries.

More recycling.

More reservoirs.

More treatment plants.

Technology matters enormously.

But environmental problems are systems problems.

Water scarcity, leakage, chemical pollution, sewage treatment, population growth, energy use and climate change are all connected.

The Teddington proposal may eventually become part of London's water solution.

Perhaps the pilot has done exactly what a pilot should do: exposed weaknesses early enough for engineers to solve them.

If so, good.

But “we will fix it in the final design” needs to be demonstrated with evidence rather than treated as an assumption.

Because during the drought for which this system is intended, the Thames itself will already be under stress.

A successful drought scheme shouldn't merely keep water in the river.

It should leave us confident that what remains there is worth protecting.

And perhaps that is the broader lesson for Going Green:

The aim isn't simply to find another 75 million litres of water. It is to build a system in which we waste less water, pollute less water and need less heroic engineering to recover what we have already damaged.