Dale Paul Gould v The Environment Agency

There has been a water mill at Langley Mill, in the Colne Valley in Essex, since at least the seventeenth century and possibly since Roman times. It is many years now since grain was ground there, and today there is no mill wheel; but the river Colne, running from Great Yeldham to the sea at Brightlingsea, flows beneath the building as it has done for centuries.

It also flows around a channel to the north of the mill, and there lies the quarrel about this beautiful property. The parties are in dispute about liability for maintaining the channel, and also about work done by the Environment Agency over twenty years ago which the mill owner, Mr Dale Gould, says is causing and will cause damage to his property.

Mr Gould has brought an action in the County Court for declarations about the construction of a deed made between his predecessor in title, Major Courtauld, and the Essex River Board being the statutory predecessor of the Agency. He has also made a claim in the Upper Tribunal against the Agency for compensation under paragraph 5(1) of Schedule 21 to the Water Resources Act 1991.

The two actions were heard together. This judgment is in three parts. Part 1 explains the role of the Environment Agency and sets out the factual background to the two actions. Part II comprises the reasons of Judge Cooke, sitting as a County Court Judge, for the County Court order made today. Part III is the decision of the Tribunal (Judge Cooke and Mr Peter McCrea FRICS FCIArb) in the compensation claim.

Mr Gould was represented in both actions by Mr John Bates, and the Environment Agency by Mr Ned Westaway and Ms Caroline Daly; we are grateful to them all. We visited Langley Mill on Monday 12 June 2023, on a gloriously sunny day when the yellow irises were in flower, accompanied by Mr Gould and by Mr Willis of the Environment Agency; we are most grateful to Mr Gould for showing us round.

The Environment Agency, established in 1996, is a non-departmental public body. It has a range of duties: to develop, publish, maintain, apply and monitor a national strategy for flood and coastal risk management (s.7 of the Flood and Water Management Act 2010); to co-operate with other Risk Management Authorities in the exercise of its flood risk management functions (s.13); and to maintain and publish the main rivers map (s.193, Water Resources Act 1991).

It also has a range of powers, some of them conferred by the 1991 Act in relation to main rivers (of which the River Colne is one); relevant to this claim is the Agency’s power under s.165, which we refer to below.

The facts set out here are relevant to both claims, and are not in dispute except where the contrary is stated. Findings of fact in this Part are made both by the County Court and by the Tribunal.

While Langley Mill is not specifically named as one of the three mills on the Colne in the Doomsday Book, there is no doubt that a mill has been present on the site for some centuries. It was used for the grinding of grain and the fulling (finishing) of textiles and is one of a series of mills on the Colne. It lies about two miles downstream (east) of the Essex town of Halstead, and three miles west of Earls Colne.

A water mill is a means of harnessing the energy of a river. The amount of energy required from a water mill depends on the use to which it is put – grinding corn using a heavy stone needs more energy than flax and threshing mills – but in any event it is generally necessary to generate a head of water to drive the wheel. (Footnote: ¹) A common way to so do is to divert the natural flow of the river into a mill race at a higher level than the natural river floor. The race has a shallower gradient towards the mill (but still having a fall to ensure the flow of water) than the river, generating a head of water at the wheel before the flow returns to the river at a lower level. The natural river channel remains as a means to prevent flooding and damage to the mill machinery. Simple and effective, but the design means that if the natural energy which would be used to drive the wheel is not so used and is instead released along the channel, an accelerated flow velocity leads to more stream power which can erode and scour the riverbed or banks.

Langley Mill was set up in this way. The river bed is 1 – 3 metres above the valley floor as it approaches the mill to form a mill race upstream, to the west of the building, which is about 10-11 metres wide but owing to vegetation is restricted to around 8-9 metres, and a mill pool to the east. Some 40 metres upstream, on the north side of the mill race is the entrance to what we shall refer to as the channel. The channel flows north then north-east for about 40 metres (which the experts termed “Reach 7a”) before turning through almost 90 degrees to the south-east (“Reach 7b”) to join the natural valley floor and then rejoin the mill pool about 20 metres downstream of the mill. The total length is some 155 metres. The general arrangement is shown on the plan below.

There are two culverts beneath the mill. One is 1.5m wide and has a sluice gate at its inlet which can be raised or lowered using a rack and pinion gear. The sluice is undershot – when raised the water flows underneath, although in flood conditions water can flow over the gate. Operation of the gate affects the levels of water in the mill race, the channel, and the mill pool. The second culvert is about 2.5m wide and is simply an open concrete weir. As we shall explain later, the relative heights of this weir and the point where the channel joins the mill race are important.

On the west and east sides of the mill building there is decking across the river; thus a walk round the house from the drive to the south-east takes one across the river on decking, along the north bank, and back across the river on decking. The sluice gate and weir can only be accessed via the decking. A small area of decking on the west (upstream) side of the building is positioned beneath the building and over the sluice gate culvert. In the 1964 deed referred to below it is pictured in the diagram and labelled “open decking over”. It is possible to stand on the “open decking over” to operate the gate, although (as we saw on our visit) it is equally possible to operate the gate without doing so.

Major George Courtauld was the owner in the mid-1960’s, the mill having been in his family’s ownership for more than a century. On 22 May 1964 the Major entered into a deed (“the Deed”) with the Essex Water Board, under which the Board acquired rights in relation to the control of water around the mill; those rights are now exercised by the Agency as the statutory successor to the Essex Water Board. The Deed is described in more detail later.

Until around 2001, water ran into the channel from the mill race over an earth bank (approximately where the “new side weir” is shown in the plan above). By 2001 the bank was suffering from erosion, such that there was a depression in the middle, causing the bank to fail. The Agency carried out a programme of works to replace the earth bank, comprising a concrete inlet weir and apron, curved brick side walls and below the apron a stone-filled gabion mattress. It is common ground that the mattress has now failed, with the cages deteriorating and stones being carried down into the channel. We saw as much on our view of the site.

Mr Gould thought that the works had been carried out in 2002 but the only witness that could speak with any direct knowledge of the works was Mr Martin Willis, who was part of the team that carried them out, and we are satisfied from his evidence that they were carried out in the summer of 2001. At that time, the owner of Langley Mill was a Mr Eddy Shah. Mr Willis said, and we accept, that Mr Shah welcomed the works. We heard evidence from Ms Natasha Manning from the Agency, who produced an internal Agency memo dated 6 February 2001 which recorded that an inspection of the mill had taken place on that date, that Mr Shah’s concerns had “been resolved” and that “a plan of action decided upon”. Ms Manning had spoken to Mr Barry Scott, another of the 2001 works team, who told her that Mr Shah said that too much water was going into the channel, and not enough into the mill race, causing a smell. While Mr Scott did not give evidence, it seems more likely than not that this was right – it is common ground that the earth bank had failed, with a central dip caused by erosion, which can only mean more and not less water entering the channel. (Footnote: ²)

There is a dispute, which we deal with below, about the height at which the Agency set the level of the new weir, relative to the culverts under the mill, and as to whether it was wider than the old earth weir. Mr Gould believes that the purpose of the works was to divert more water into the channel to alleviate the Agency’s requirement to attend the mill regularly to operate the sluice gate.

Mr Gould made two contentions about the works. First, that they were part of a planned flood relief scheme for Halstead, following the town’s flooding on 21 October 2001. Secondly, that they were designed to make more water flow around the channel so that the Agency would not need to attend to open the sluice gate in times of potential flooding. As for his first contention, we are satisfied that the works had already been completed before the Halstead flooding, based on Mr Willis’s evidence and Ms Manning’s evidence that the initial engineering appraisal for flood alleviation works was not completed until August 2002.

The Agency disclosed a table prepared by Ms Manning contained in an email dated 29 July 2022 in which she listed any Agency employee who might have been involved with Langley Mill. The involvement of Mr Gary Cockett, now retired, is described as “initiated withdrawal of (sic) mills programme in Essex, requested modelling and asset inspection carried out in ~2016. Has met Claimant on site previously, until complaint was filed.” An Agency spreadsheet dated 24 October 2018 had also been disclosed, listing mills on various rivers, and against which there was a forecast date by which the current level of operation and maintenance would cease. Mr Gould contended that the Agency’s intention to withdraw from the mills had started by the early 2000’s. He referred to the example on the spreadsheet of Tinkers Lane on the River Stour against which was noted “adjacent weir constructed in 1990’s to alleviate need to visit and open gates”.

Ms Manning was not employed by the Agency until much later, but we accept her evidence that the programme did not commence until around 2011/12 – at any rate much later than the 2001 works. In the same document the date of cessation of the current level of operation and maintenance was forecast to be 31 December 2019. For these reasons we do not accept Mr Gould’s second contention.

When Mr Gould purchased the mill in 2011, the vendor told him that she had a good relationship with the Agency, which responded promptly when she raised any concerns. Its Field Team would attend the site at least monthly and discuss any issues she had; they would also carry out regular bank maintenance, clearing the river and channel, maintain the sluice gate and operate it when required, and repair the gangway and decking as necessary.

This changed shortly after Mr Gould bought the mill. While the formal withdrawal from mills programme was not enacted until much later, the Agency began to scale back its operations at the mill. Ms Manning explained that the pressure of constrained budgets and limited staff resources meant that the Agency began to prioritise areas where it could prevent flooding or deliver the greatest flood risk management benefit. Langley Mill was a site where the control measures did not provide any wider public benefit, and there was no compelling justification for the use of public money to continue to maintain the banks of the channel or mill pond at the mill. We accept Mr Gould’s evidence that the level of the Agency’s involvement as described to him by the vendor did not continue in his ownership.

Since 2001 the Agency has carried out further works in three tranches. In 2006, gabions were installed to the north and south banks of the mill pond, largely to prevent further erosion which was threatening the integrity of the vehicular drive; in 2012 further gabions were installed on the southern bank of the pond; in 2016 willow spilings were installed to the northern bank of the pond and to the eastern end of the channel banks. By 2019 some of these had died, so the Agency installed aqualogs and a geotech membrane. Mr Gould disputed that work was done in 2012, but Mr Simon Hipkin gave evidence for the Agency that he was the site foreman for the works in 2012 and we accept his evidence.

We do accept Mr Gould’s evidence that from Autumn 2011 he was in contact with the Agency endeavouring to persuade it to continue the previous maintenance programme. From 2014 to 2017 this involved informal discussions with local management, escalating to the Eastern Region area management, and then escalated to the highest level. The issues were not resolved, and in due course Mr Gould initiated proceedings in the Court and in the Tribunal.

It is common ground, and was evident during our site view, that the gabion mattress installed in 2001 has failed. The Agency accepts responsibility for repairing it (pursuant to the Deed), and for any consequent erosion caused by its failure. It is also common ground that there has been erosion in the channel. The extent to which the failure of the gabions was the cause of that erosion is, as we shall see, in issue.

The Deed, dated 22 May 1964, recorded that Major Courtauld, as owner of Langley Mill, was entitled to control the flow of water in the River Colne where it passed through the mill subject to the statutory powers of the Essex River Board but agreed to grant to the Board rights in relation to the control of water in the River.

By clause 1, and “in consideration of the covenants by the Board hereinafter contained” the Major granted to the Board:

“(a)

The right to control water in the River Colne at Langley Mill … in order to achieve such level as the Board may from time to time consider to be necessary or desirable.

(b)

The right to carry out such repairs or alterations to or reconstruction of the existing flood gates adjacent walls and inverts grass spillway and millhead bank at Langley Mill… shown on the plan annexed hereto and to construct any new works which the Board may from time to time consider to be necessary to supplement or replace and of the above mentioned works due regard being paid to the interests of the owner.

(c)

The right for the Board its employees agents and contractors with or without plant or vehicles at all times to have full access to the said flood gates grass spillway and millhead bank and other works hereinafter mentioned and to any new works constructed by the Board hereafter for the purposes of operation maintenance repair or reconstruction.”

The Board covenanted at clause 2 that it would:

“(a)

Operate the said flood gates and any control works which it may hereafter construct as it may from time to time consider to be necessary or desirable and maintain the same in good working order

(b)

Carry out at its own expense such works of repair or reconstruction of the said flood gates grass spillway millhead bank the brick walls up to ground level adjacent to the flood gates the inverts of the mill-race and any new works which it may construct as it may from time to time consider to be necessary or desirable.”

The Deed contains a plan showing the layout of the mill, the mill race and the channel, and diagrams showing a cross-section of the sluice gate and the weir beneath the mill, and an aerial view of the mill building, with decking to east and west and the “open decking over” beneath the building and on the downstream side of the sluice gate.

It is agreed between the parties that the fixed crest weir (beneath the mill), the sluice gate and the concrete weir at the entrance to the channel are all “control works” under the 1964 Deed, so subject to the maintenance obligation in clause 2(a). The Agency accepts that the gabion mattress immediately downstream of the side weir is part of that control work, and also accepts that it is responsible for maintaining the “open decking over” as marked on the plan, which is the small section of open decking adjacent to the sluice gate.

In April 2021 Mr Gould brought proceedings in the High Court – which were later transferred to the County Court – seeking a series of declarations. The particulars of claim were amended by agreement in the course of the proceedings, and the issues the court now has to decide are:

Is the decking within paragraph 2b?

The Particulars of Claim state that the Agency is obliged to maintain the “timber decking, over decking and gangway areas” on the plan attached to the Deed. As noted above, the diagram depicts the decking on either side of the building as well as the “open decking over”. The Agency accepts that it is obliged by clause 2b to maintain the “open decking over”; it is worth noting that that small section of decking, while it is not essential as a stance for the person operating the gate, certainly has no other function.

But the Agency denies any responsibility to maintain the rest of the decking. It is not referred to anywhere in the wording of the Deed. Clause 2b refers specifically to the existing features in 1964 for which the Agency took on responsibility (albeit only insofar as it considered it necessary or desirable to repair or maintain them). And the rest of the decking is not within “any new works which [the Agency] may construct”.

Mr Bates argued that there is an implied term that the Agency will repair the decking.

The principles on which terms may be implied into a contract are well-known and are not in dispute. The authority is Marks & Spencer plc v BNP Paribas Securities Trust Co (Jersey) Ltd [2015] UKSC 72, where the Supreme Court re-stated the principle that the question is not whether it would be reasonable to imply a term, nor whether it would be fair, but whether it is necessary to do so “in order to make the contact work” (paragraph 77) – or, in the older phrasing, in order to give business efficacy to the contract.

There is no basis for the implication of a term that the Agency will maintain the decking, aside from the “open decking over”. Major Courtauld granted to the Agency’s predecessor in clause 1(c) “full access” to the works mentioned in the Deed. On the assumption (of which more will be said later) that Mr Gould is bound by that grant, he has to maintain the decking insofar as that is necessary to give the Agency the access to which it is entitled. All the Agency needed was a grant of access; if the decking falls into disrepair and Mr Gould does not maintain it the Agency can enforce its right and thereby require him to repair it or to provide some other means of access. Mr Bates’ argument that there is a “plain and obvious gap” in the Deed, such as that identified in JN Hipwell v Szurek [2018] EWCA Civ 674 is unsustainable.

It is argued that the fact that the Agency has repaired the decking in the past indicates the intentions of the parties to the Deed. Whilst the conduct of the parties may be an aid to construction where a document is ambiguous, there is no ambiguity here; in any event it seems that the Agency before 2011 or so went beyond its obligations in dealing with Mr Gould’s predecessor in title in a number of respects, and there is nothing to indicate that anything it may have done to the decking in the past was anything other than voluntary.

It is also suggested that the obligation to maintain the decking is required as consideration for the grant of access. That is without foundation. This is a deed, so no consideration is required in any event; insofar as the deed does expressly contain consideration then there is no need, in this or any contract, for a specific item of consideration to match a specific grant such as the access in clause 1(c).

The decking, other than the “open decking over”, is not within the Agency’s obligations in clause 2(b) and no term can be implied in the Deed that the Agency will repair it.

Is the channel a “control work” within clause 2(a)?

The second practical point that Mr Gould asks the court to determine is that the channel is among the “control works” that the Agency is obliged to maintain under clause 2(a).

The channel was not made by the Agency and it is difficult to see how it can fall within “any control works which it may hereafter construct”. The term “control works” is not defined in the Deed; as Ms Daly (who presented the argument for the Agency in the County Court claim) points out, clause 1(a) is helpful here in that it grants the right to control the water in the River Colne, and control works may be supposed to be structures designed to help with that. The Agency accepts that the side weir and the gabion mattress are control works; the construction work done in 2001 was done in accordance with the Deed.

Mr Bates’ argument is that the side channel has “become an integral part of the system”, and he relies upon the expert evidence of Dr Brookes to the effect that the side weir, the gabion mattress and the channel “are all integral parts of the same water control measure”.

Dr Brookes is not a witness in the county court action and there is no permission to adduce expert evidence here. Nor should there be; the action is about the construction of the Deed. Clause 2(a) is very specific; it refers to the flood gates and to later works; there is no basis on which natural or other physical features can be included within the clause because of their connection to such works. In any event the channel itself does not control water levels; levels are controlled by the side weir, the mill weir and the sluice gate, and the side channel simply receives water.

The Agency is not obliged by the Deed to maintain the side channel.

Does the Deed have perpetual effect?

The third declaration sought by Mr Gould is that the Deed cannot be brought to an end.

No-one is seeking to do so. The request to the court for a declaration that the Deed continues in perpetuity arises from an anxiety that the Agency may seek to give notice to terminate the Deed – an anxiety brought on by a remark by one of the Agency’s staff in correspondence.

Mr Bates in his skeleton argument addresses this issue by arguing that clause 2(b) remains part of the Deed. The Agency has not argued that it does not and it is difficult to see why Mr Bates puts his argument this way. The Agency in its pleaded response observed that the obligation to carry out maintenance in clause 2(b) is only to do so as the Agency may from time to time consider it necessary or desirable. It filed a witness statement by Mr Aaron Scott, its Operations Team Leader in the Essex Flood and Coastal Risk Management Department; he said that the Agency does not consider it necessary or desirable to maintain the structures listed in clause 2(b), because it takes the view that the arrangements at Langley Mill make little contribution to protecting downstream and upstream communities from flooding.

Subject to the contractual requirement to act rationally (see Braganza v BP Shipping Limited [2015] UKSC 17) the Agency is entitled to take that view, and that has no bearing on the construction of the Deed; this is not an action for breach of contract. In argument at the hearing the point in issue was clearly not what the Agency is choosing to do in pursuance of its obligations under the Deed but whether or not the Deed could be terminated by either party giving reasonable notice to the other.

Mr Bates says that the Deed was clearly intended to be perpetual; certainly its wording gives no indication of any limitation in time. He argues that the Deed has not been frustrated; but no-one is suggesting that it has. He argues that the Agency is wrong to put upon Mr Gould the responsibility for maintaining and operating the flood gates for the amenity of others and to prevent flooding, and that it is failing to act in accordance with the common purpose of the parties to the Deed which was for the Board (and now the Agency) to use the control works to fulfil its responsibility; but this is akin to an argument about breach of contract or perhaps about public law, and either way is not relevant to the construction of the Deed.

Essentially the application for a declaration that the Deed cannot be brought to an end is a pre-emptive move; Mr Bates urges the court to make a declaration, rather than leaving the parties to litigate afresh if and when the Agency decides to try to bring the Deed to an end.

The Agency argues that there is simply no need for a declaration; Ms Daly points out that the Agency has not sought to bring the Deed to an end and is observing its obligations (hence the promised repair of the gabion mattress). But if the court is minded to make a declaration she argues that this is the sort of arrangement in which the court will readily imply a term that the contract can be brought to an end on reasonable notice.

The court declines to make a declaration for two reasons.

First, there has been little or no analysis by the parties of what the Deed is. Ms Daly thought it was a contractual licence, but that is hard to accept. It is a Deed; it expressly “grants” certain rights. Mr Bates said it is a statutory easement. Neither party had really delved into full argument about what the Deed is and whether the obligations within it do bind the successors in title to the original parties, and without that it would be premature to make a decision about whether it can be brought to an end.

Second, the only way to terminate the Deed – absent a frustrating event – would be by implying a term, and at present there does not seem to be any particular reason to do so. The Agency does not point to something in the present circumstances and argue that because of it there is a need to imply a term about termination by notice in order to give business efficacy to the contact. But the court is not prepared to say that that would not be possible in the future.

It is not appropriate for the court to make a declaration now that no term could be implied that would enable either party to terminate, when that would bind the parties for ever, regardless of what might happen in the future. Either party might come to regret it.

Accordingly the court declines to make a declaration as to whether there is an implied term in the Deed that it can be terminated on reasonable notice.

We now turn to Mr Gould’s claim for compensation under paragraph 5 of Schedule 12 to the Water Resources Act 1991, both for costs he has already incurred in repairing the bed and banks of the overflow channel and for costs to be incurred in the future. Costs already incurred are said to amount to £1,775; costs to be incurred are estimated at £12,230 for urgent work and further work estimated to cost £34,000. The Tribunal was not told how much of what is claimed will be rendered unnecessary by the Agency’s commitment to repair the gabion mattress and the erosion its failure has caused.

Mr Bates opened the compensation claim with a discussion of Rooke’s Case (1597) 5 Coke Reports 99b, which he said establishes that it is wrong for one person to bear responsibility for protecting a number of other riparian owners from flooding. The burden should be shared. Therefore, it was argued, the Agency’s withdrawal from its mills programme (see paragraph 19 above) was wrong in law, and it was put to Ms Manning in cross-examination that the programme was initiated without consultation and was driven by lack of resources. The claimant’s view is that the Agency should continue to maintain the river banks and control the operation of the sluice gate at Langley Mill as it has done in the past.

This was a puzzling argument. A challenge to the withdrawal from the mills programme on procedural or legal grounds could only have been made in the Administrative Court by way of judicial review. The relevance to the compensation claim of the argument about the withdrawal of the programme is not immediately obvious, since the basis of that claim is that the works done in 2001 were defective and have caused damage.

It may be that the argument was made in order to bolster the claim in nuisance, which requires a wrongful interference with the claimant’s use or enjoyment of land. Or it may have been intended to head off one aspect of the Agency’s response to the compensation claim, which is that the erosion could have been avoided had Mr Gould managed his sluice gate properly.

Whatever the intention of the argument, we are not persuaded by it. We see no substance in the argument that the withdrawal from the mills programme was unlawful. The Agency has no statutory duty to control the flow of water through the mill or to keep the river banks in repair; Mr Bates did not suggest that there is any such statutory duty, and as we saw in paragraphs 6-7 above the Agency’s statutory duties are far more general and certainly do not extend to the management of mills or the maintenance of riparian properties. It appears that the Agency has gone beyond its statutory duties in the past and no doubt when Mr Gould bought the mill he hoped it would continue to do so, but the Agency’s past actions do not change the legal position and the Agency is perfectly entitled to cease to do things it is not obliged to do.

Rooke’s Case is of no assistance to Mr Gould. It concerned a liability to taxation imposed by the Commissioners of sewers who had authority pursuant to the statute 6 H. 6 cap 5 (meaning the fifth statute of the sixth year of the reign of Henry VI, 1428) to impose a levy for the repair of riverbanks. The plaintiff brought an action claiming (among other things) that he should not bear a tax liability as a riparian owner when his neighbours, whose land was also “subject to drowning, if the said bank be not repaired”, did not. He was successful; the Court of Common Pleas stated that “the commissioners ought not to charge him only with the whole, but ought to tax all who had land in danger.” The decision turned on the words of the statute (“for the statute … has precise words in the said commission, that no person of any estate or condition shall be spared”). No principle of common law was involved and none was established.

A case about the construction of a fifteenth century taxation statute, long repealed, has no relevance to the present case and has no bearing on the liability of the Agency under statutes now in force or pursuant to the common law.

For all these reasons we regard the argument about the withdrawal from the mills programme as irrelevant to Mr Gould’s claim. For the same reasons, his statement at paragraph 10 of his witness statement dated 15 November 2022 that the Agency “is responsible for controlling the flow of water through the mill” is incorrect. So is Dr Brookes’ assertion that the Agency has “specific responsibilities” for the operation and maintenance of the main river channel and the side channel (paragraph 20 of his report). The covenant by the Agency in the Deed requires it to operate and maintain the sluice-gate and maintain any works it might install (such as the side weir) if it considers it “necessary and desirable” to do so, but that does not detract from Mr Gould’s own ability and responsibility to control the flow. Nor does it detract from the common law responsibility of riparian owners such as Mr Gould to their neighbours; they will be liable in nuisance if they cause flooding or other disturbance to other landowners for example by opening or closing sluice gates, or by the way they manage vegetation.

As the owner of a mill Mr Gould does have one special right at common law, namely to the “accustomed flow of water”, described in Roberts v Gwynfrai DC [1899] 2 Ch 608 by Linley MR at 612 as follows:

“The right of the plaintiff as the owner and occupier of his mill is to have the water flow down the stream, which has its origin in the lake, in the accustomed way. That right is subject to the rights of the other riparian proprietors higher up the stream; but, subject to those rights, there is no right whatever to alter the flow of the water in its old accustomed way. If it is said that the alteration of the old flow is an improvement, that is a matter of opinion. There is no right to interfere with the accustomed flow of the water.”

We shall have more to say about that right shortly.

Section 165 of the Water Resources Act 1991 provides the Agency with a general power to carry out works relating to flood defence and drainage. The version of s.165 in force in 2001, gives the Agency power in relation to a main river:

“1 (a) to maintain existing works, that is to say, to cleanse, repair or otherwise maintain in a due state of efficiency any existing watercourse or any drainage work;

(b)

to improve any existing works, that is to say, to deepen, widen, straighten or otherwise improve any existing watercourse or remove or alter mill dams, weirs or other obstructions to watercourses, or raise, widen or otherwise improve any existing drainage work;

(c)

to construct new works, that is to say, to make any new watercourse or drainage work or erect any machinery or do any other act (other than an act referred to in paragraph (a) or (b) above) required for the drainage of any land.”

The Agency accepts that in carrying out the 2001 works, it was relying on its permissive powers under s.165 in addition to its contractual obligations under the Deed.

Section 177 gives effect to Schedule 21 to the Act, making provision for imposing obligations as to the payment of compensation in respect of the exercise of powers under s.165, among others. Paragraph 5 of Schedule 21 provides:

“(1)

Where injury is sustained by any person by reason of the exercise by the appropriate agency of any powers under section 165(1) to (3) of this Act, the appropriate agency shall be liable to make full compensation to the injured party.

(2)

In case of dispute, the amount of any compensation under sub-paragraph (1) above shall be determined by the Upper Tribunal.”

It is common ground that in order to obtain compensation under paragraph 5, a claimant must show that what they complain of would otherwise be actionable as a tort. The tort alleged here is nuisance, which is a wrongful interference with another person’s use or enjoyment of their land. The claim here is founded on the allegation that the works done in 2001 were defective; the effect of the Limitation Act 1980 is, it is agreed, that if a nuisance is proved the claimant can claim compensation only for damage that has occurred since 2016 (the claim having been issued in 2022) and for future damage.

That, then, is the common law and statutory background to the compensation claim, to which we now turn.

Mr Gould’s case is that the work done in 2001 was poorly or incorrectly executed (in a number of ways that we explore below) and that as a result too much water flows through the side channel which has been and continues to be eroded and that he has suffered and will suffer the expense of repair to the channel and its banks and of clearing debris.

The Agency’s case, in summary, is that the work in 2001 was carried out competently, using the right materials and making the right choices. The erosion of the side channel has been caused by the natural flow of water which could have been alleviated had Mr Gould operated the sluice gate and met his responsibilities as a riparian owner.

The Agency also says that in any event the landowner in 2001 consented to the work. It is convenient to deal first with that final point because it can be swiftly dismissed. There is no evidence of what Mr Shah agreed to in 2001. There is evidence, as we have seen, that he was content with the plan in February 2001 (see paragraph 16 above). But as Mr Bates points out, in circumstances where the Environment Agency has statutory powers to carry out the work whether or not the landowner consents it is difficult to say that he did so. If the landowner actually requested the work that might well indicate consent sufficient to make it impossible for him or his successor to claim compensation for it, but there is no direct evidence that he did so. Accordingly Mr Gould’s claim cannot be dismissed on the basis that the work was done with the then landowner’s consent, and we turn to look at the substance of the claim.

Mr Gould has to prove the elements of the tort of nuisance: a wrongful act interfering with the claimant’s use and enjoyment of land. Mr Gould said that the weir was constructed incorrectly, and that that caused erosion in the side channel; he also complained that since 2001 there has been a “preferential flow” of water in the side channel and a reduction in the accustomed flow of water under the mill which, if proved, would in itself be tortious being both a wrongful act and an interference. So we look first at the 2001 works, in terms of the design, materials used, the drop from the weir to the channel, and finally the gabion mattress. We then consider whether there has been any change in the flow of water under the mill and over the side weir.

Central to Mr Gould’s case is the argument that the work done in 2001 was badly designed and inappropriate, leading to increased flow over the side weir and increased erosion, because:

In the paragraphs that follow we assess those propositions in turn. We were assisted by the evidence of two experts. For the claimant, Dr Andrew Brookes, who is a geomorphologist, and for the Agency Mr Jeremy Benn who is a hydrologist. Between them, the experts have decades of experience in water matters, as well as a plethora of additional qualifications and Chartered membership of various august bodies, including the Charles Close Society for the Study of Ordnance Survey maps. Both have visited the site, and they carried out a joint site inspection in January 2023.

We start with the issue of whether the Agency was right to replace the previous earth weir with a concrete version. Mr Gould thought the Agency should have replaced the earth weir like for like. Dr Brookes agreed.

Whilst Dr Brookes accepted that by 2001 there was a need to do something to repair the earth weir, he thought that an earth and clay bank would have been better. The new concrete structure had a smoother face and an entirely different structure that would, and did, result in erosion to the bypass channel. Even if the concrete weir was a like for like replacement of the earth weir in terms of size and location, in his view the different materials resulted in erosion because of the higher velocity of water passing over the smoother surface, resulting in turbulence and therefore erosion. The new weir would also allow more debris to pass over.

Mr Benn said (in paragraph 6.3(c) to (d) of his report) that the earth weir was “an unusual and not a particularly resilient arrangement”, and that he was not surprised that by the late 1990s the spillway had become eroded. His view was that the concrete weir was typical of the type that would have been constructed at the time and was normal for an overflow structure on a river. A design today would be very similar, save for incorporating a fish and eel pass. His evidence was that had the Agency not installed the concrete weir, there would have been progressive collapse of the left-hand bank of the mill race and likely flooding to the garden of Langley Mill. He noted that the transition between a hard structure and natural riverbanks does often result in erosion, which can be mitigated either by provision of a stilling basin or by stone-filled gabions as in this case. He was sceptical about the idea that the smoother weir would allow a faster flow of water; only if there was a close and level grass cover on the earth weir would that have slowed the flow to any appreciable extent. As to debris, he thought that the side weir was wide enough to prevent anything getting jammed, and expressed the view that large logs would in any event be caried straight towards the mill rather than making a left turn into the channel.

We found Mr Benn’s evidence about the use of concrete convincing. An earth weir is obviously more liable to wear away than concrete; it is agreed that the use of concrete can result in erosion, and we accept Mr Benn’s evidence that the use of a gabion mattress was appropriate to mitigate that effect.

We turn next to the height of the replacement weir.

Dr Brookes’ initial, and admittedly approximate, survey suggested that the replacement weir was set a few centimetres lower than the culvert under the Mill, the Agency said half an inch. Dr Brookes and Mr Benn carried out a measured survey which showed the crest of the inlet weir to the channel at 28.883m AOD. The culvert beneath the mill was not level across its width; at its lowest point in the middle its height was 28.88m and at its sides 28.968 and 28.97m. The experts agreed that the inlet weir to the channel was therefore 3mm higher than the lowest point in the culvert. They also agreed that it was 5.6mm higher than the average height of the culvert (which they had as 28.393m). We wonder whether that is quite right. On our calculations the average height of the culvert is 28.939m (not .393) and so the height of the bypass weir is 5.6mm lower than the average height of the culvert. However, we are sceptical that in reality measurements can be taken to that degree of accuracy. Mr Benn in his report at paragraph 6.1(h) stated that he and Dr Brookes agreed that “To all intents and purposes the two weirs have the same level.” This is consistent with Mr Willis’s evidence that the height of the weir entrance into the channel was set so that in normal conditions, with the sluice gate shut, water would just “spill over” the new weir, as well as going over the concrete weir below the mill.

But there is no record of the height of the previous (failing) earth weir. The only direct evidence we have is that of Mr Willis, who said that the level of the weir was reinstated to its previous height, before the earth bank started to erode. We accept that evidence.

As for the width of the channel (agreed to be approximately 4 metres) while in correspondence with Mr Gould, the Agency had previously said that a section of the bypass channel was widened as part of the works, Mr Willis said that was incorrect. He explained that the brick “wing walls” were rounded off to make them more resistant to erosion, and that any increase in the size of the mouth of the channel was very slight. Again, we accept Mr Willis’s evidence. Furthermore, Mr Benn at his paragraph 6.1(i) stated that the two experts agreed that the “current width of the overflow channel is similar to that shown on historic OS maps.”

The evidence given by Mr Willis is decisive on the issue of the height and width of the side channel because he did the work. We accept the Agency’s case that the side weir was built to approximately the height of the old earth weir so that in normal conditions water would flow under the mill but also spill over into the side channel, and was not materially wider than the old weir.

This aspect of the claimant’s case was developed at the hearing rather than in the pleadings or the evidence, and it arose we think from an appreciation in the course of the hearing as a result of Mr Benn’s evidence of the importance of the drop from the concrete weir into the channel. Any water that does not pass beneath the mill is diverted over the concrete weir, and it has to drop to the natural level of the river bed, in this case about a 1.4 metre drop.

The slope, while less sudden than the drop over the wheel, dissipates the energy of the water and is a cause of erosion. Mr Bates therefore suggested to Mr Benn in cross-examination that when carrying out the work in 2001 the Agency should have eliminated, or at least, smoothed out, the slope in order to minimise erosion in the channel.

We regard that as an unrealistic suggestion. It was not supported by expert evidence from Dr Brookes and was not an option that he considered. Mr Benn dismissed it as a practical possibility in 2001; if the slope had been somehow evened out (perhaps over the length of reach 7a) Mr Gould’s garden would have repeatedly flooded because the bed of the channel would have been higher for some of its length. What actually happened was, as Mr Willis said, that the Agency’s team matched the slope they found. The purpose of the installation of the gabion was to absorb the energy as the water dropped down the slope; and it is easy to see how that happens as some of the water is going to fall through the gaps between the stones within the gabion so that the force of the flow is dissipated.

We accept the evidence of Mr Benn and Mr Willis that there the Agency matched the slope it found, and we do not accept the argument that it was incorrect to do so or that the drop to the natural river level could somehow have been eliminated.

A gabion mattress was installed to provide a more solid bed to the river, to prevent erosion of the base of the channel; we accept Mr Willis’ evidence that the channel was dug out sufficiently to ensure that the gradient of the gabions was the same as that previously.

So the slope from weir to channel bed was inevitable. That slope was a potential source of erosion, and the Agency’s response to that potential problem was to place a gabion mattress over the slope from the weir to the bed. As Mr Benn put it at paragraph 4iv of his report:

“The erosion to the banks of the bypass channel for a length of approximately 20-40 metres downstream of the overflow inlet weir IS influenced by the 2001 works… The transition between a hard structure and natural riverbanks does often result in erosion which can be mitigated by use of large stone-filled gabions or provision of a stilling basin.”

The gabion mattress has now failed and, as we said above (paragraph 15 ) is visibly causing erosion as water rushed around the gap between the gabion and the bank of the channel. The Agency has now offered to repair (or to pay for) the gabion and the erosion caused by its collapse, in exercise of its powers and pursuant to its responsibilities under the Deed.

The Agency cannot be liable for failure to repair the gabion mattress until now; it is well-established that the Agency cannot be liable for failing to exercise any of its powers, only for what it does in exercise of its powers. The claimant’s case is that the gabion was defective because they have failed too early. In support of that proposition Dr Brookes provided a chart of the expected life of a gabion in the UK, from the website of a supplier, which indicated that while gabion in a polluted industrial area with a high salt spray would last about 25 years, and in an urban light industrial area with a low salt spray would last for about 50, gabion in a rural area could last 100 years or more. Accordingly he said that gabions properly designed and implemented in a rural to suburban area such as Langley ill would last 80 to 100 years. Dr Brookes agreed in cross-examination that the chart referred to gabion that is in dry conditions, save for spray, and not under water. Dr Brookes had no information about the expected lifespan of gabion under flowing water; Mr Benn said he would be pleased to get 20 years’ use out of gabion in a river, and we accept that evidence.

Accordingly we find that there was nothing wrong in the Agency’s choice of a gabion mattress to absorb the energy of the water as it flows down from the concrete weir to the channel bed, and there is no substance in the argument that they have failed earlier than they should have done.

In conclusion to the material under this and the two previous sub-headings, we find no fault with the way the side weir was constructed in 2001.

We turn now to the other major component of Mr Gould’s case: his argument that the result of the construction of the side weir was an alteration to the flow of water under the mill, so that there was a preferential flow over the side weir and his “accustomed flow” was reduced.

This is an important element of Mr Gould’s case. A change in the “accustomed flow” of water beneath a mill is in itself a tort (private nuisance) so that, even if no fault can be found with the 2001 works in themselves, if Mr Gould can show that since 2001 the flow has been changed then he can succeed in his claim.

In the absence of measurements of the flow taken at a number of times of the year before the installation of the side weir and in a range of conditions, it is difficult to assess whether the flow has changed. Dr Brookes’ opinion was that it has, for a number of reasons. One reason was the evidence of maps, which he said show that there was no water passing over the earth weir into the side channel before the concrete weir was built. Second, he said that much more water flows over the side weir than under the mill, on the basis of measurements he has taken and on the basis of observations at the site. Third he said that no erosion occurred in the channel prior to 2001 and that the erosion that has occurred since is more than normal for this stretch of river, which must itself show that the flow has been altered.

Dr Brookes sought to argue that water did not flow from the mill race into the channel before 2001, and that the channel’s water arose from springs in the ground beside it. He relied upon what he said was an Ordnance Survey map that pre-dated the works in 2001, and which showed at the point of the side weir a gap in the water of the river and the water of the side channel. At the near corner of the side channel there was a single blue dot. In Dr Brookes’ experience, this dot represented a spring or seepage of water out of the ground. In his view, the plan indicated that before the works there was no surface water where the new side weir had been installed. This analysis ran aground in two ways. First, the date of the plan he relied upon could not be confirmed. Secondly, Mr Benn produced a version of the plan confirmed to be from 2023, so well after the works, which while not having the blue dot, also showed a gap between the two blue areas of water. Accordingly we derive no assistance from a comparison of the plans.

Mr Willis recalled that there was water in the channel when the team arrived to do the work in 2001. Ms Manning referred to Mr Scott’s recollection that the channel was a running channel. Mr Benn’s opinion was that the channel had always been an active conduit for water. He pointed to three pieces of historic evidence: first, that the channel has been designated as a main river (see paragraph 7 above) since the 1960s and it would be odd for a dry watercourse to be so designated; second, that it has been shown as a channel with a left and right bank on OS maps from at least 1855-1882 and would have been shown as a single line if it only occasionally caried water; and third because the 1884-1914 and 1938 OS maps show a “waterfall/overflow” at the inlet to the bypass channel and that label would only have been used for a feature that was actively carrying water.

We find that there was usually water in the channel in 2001 and prior to that date.

Dr Brookes carried out an analysis of the “flow split” between the channel and the mill race. He did this by visiting the site on 29 September 2022, a time of very low flow, and holding a bucket under the fixed weir beneath the mill, timing how long it took to fill up, converting the result to “cumecs”, or m³/s. Unfortunately, the same method could not be applied to the weir at the channel, because of its length and lack of a V shape to enable 100% of the water to be captured and measured. He then took the flow gauging station at Earls Colne to extract data from the same time as his bucket test. The Earls Colne data was adjusted to estimate the total discharge across the two weirs, to allow for the ungauged catchment and the River Bourne which enters the Colne upstream of the Earls Colne gauge. The adjustment adopted was 25%, based on area, with a sensitivity analysis ranging from 15% to 35%. In the result, the flow over the channel was between 9-12 times greater than over the fixed weir, with the sluice gate closed.

Mr Benn pointed out the inevitable weaknesses with this approach. A proper comparison can only be made by using the same method for both the mill weir and the side weir, for instance by blocking all but a small element of the channel weir by sandbags. As it was, whilst Dr Brookes’ method was appropriate where there was only one unknown, in this case there were two (the side weir, and the River Colne). He expressed the view that the experiment could have been repeated several times with different results.

The other relevant evidence here is the observations made by Dr Brookes and Mr Benn at their site visit. They asked Mr Gould to open the sluice gate, which he did. When it was fully open, water continued to flow approximately 5 centimetres deep over the side weir, while the weir under the mill was dry and the water 8 – 9 centimetres below the sill. Dr Brookes was not able to explain how this could happen in light of the agreed evidence that the two weirs are pretty much the same height.

Mr Benn’s explanation was that the vegetation growing thickly on the north side of the river between the side weir and the mill narrows the channel (as we noted at paragraph 11 above) and obstructs the flow towards the mill. In light of the fact that the two weirs are at about the same height we accept that explanation.

Mr Benn’s evidence, which we accept, was that the water that the flow split at lower flows is determined primarily by whether the sluice gate is open, and secondly by the amount of vegetation in the mill race. At higher (flood) flows, the influence of opening the sluice gate reduces, and the flow will depend upon the relative capacity of the mill race (which itself will be determined by vegetation in the channel) and the overflow.

Finally Dr Brookes argued that the erosion itself in the side channel demonstrates that there is a preferential flow because in his opinion there was no erosion in the channel before 2001.

Dr Brookes used erosion pins driven into the bank of the side channel to demonstrate that erosion is happening. But it is not in dispute that it is happening. The experts disagreed about just how much material had been washed away in the last twenty years, but again we do not think this assists us. Dr Brookes also pointed to the deposit of sediment at the end of the channel, where the water rejoins the river; certainly there is sediment there, which we saw on our site visit, and we accept that sediment is what happens when material is washed away from the banks in the course of erosion. But none of that tells us whether the 2001 works made any difference.

In Dr Brookes’ opinion the River Colne is “a low energy river type that does not have the competence to significantly erode its channel bed and banks. He carried out a “Fluvial Audit”, which we understand to mean an analysis of the geomorphology of the river and a visual inspection, generating a description of the river in the Langley Mill area and the upstream catchment. Dr Brookes saw no significant erosion upstream. He asserted that the shape of the side channel did not change in the 150 years before the effect of the works began to be felt, and that what should have been an inactive channel was now suffering serious erosion.

Dr Brookes based that opinion on an aerial photograph of 2005 – not long after 2001 – and the 1876 OS map; when the line from the map was overlaid on the photograph it could be seen that the route and shape of the channel had not changed. However, in cross-examination it became clear that what Dr Brookes had drawn on the photograph was the 2005 line of the channel, which shows reach 7a extending north-east at an angle; the 1876 map shows reach 7a going north-east along a line much closer to north. The channel has moved and the angle of the bend to reach 7b has become shallower.

Mr Benn had walked downstream from Langley Mill and found a number of instances of erosion. He did not accept that the River Colne has reached an equilibrium and no longer erodes its banks.

We find Dr Brookes’ assessment of the historic position unconvincing. The comparison of the map and the photograph did not support his view; on the contrary it showed how the shape of the channel has changed and the sharpness of the bend has been worn away. And Dr Brookes’ opinion ignores the inevitable 1.4 metre drop from the mill race to the natural valley floor. We remind ourselves of the way in which energy is stored for use at a water mill (see paragraph 10 above), by raising the river bed to create a head of water at the wheel. Where the energy is not being used to grind corn it has to go somewhere; in practical terms it is going to be dissipated where it flows back down to the natural level of the river, whether under the mill or over the side weir. As it flows downhill the force of the water is bound to cause some erosion. As Mr Benn put it, “While the erosion can be mitigated by protection works, it cannot be completely prevented – at least while the mill race and sluice remain and that the channel lies at an artificially high level through the mill.” Although Dr Brookes was certainly aware of the topography of the immediate area we think that he failed to appreciate its importance. He therefore came too readily to the conclusion that there was no erosion in the side channel prior to 2001, and concluded that the only cause of erosion in the channel since then must be the 2001 works.

Accordingly Mr Gould and his witness have not proved that there has been more erosion in the side channel since 2001 than there was before.

Even if there has, there are a number of available causes. Mr Benn said, and Dr Brookes to some extent agreed, that there have been more frequent floods in the last 20 years than there used to be. Vegetation can divert or disrupt the flow and can cause erosion. And the elephant in the room is that the flow of water in the side channel is under Mr Gould’s own control by means of the sluice gate.

Dr Brookes outlined his view of the effect of opening the sluice gate under the mill on the water entering the channel in different conditions. First, in normal circumstances, when the gate was closed more water would flow over the side weir, causing erosion. Secondly, at high water flows, Dr Brookes said that with the sluice gate partially open, erosion below the side weir would occur in the channel for a period of minutes to hours until the gate is opened sufficiently. Those minutes to hours, on each occurrence, would in aggregate have significant erosive effect. But once the sluice gate was open sufficiently the rate of erosion in the overflow channel will decline until negligible. Thirdly, in flood conditions, even with the sluice gate open significant flow would, he said, be observed down the channel. Water would flow back up from the gate opening and under house weir, and discharge across the side weir, causing erosion in the channel.

Dr Brookes said that Mr Gould is “allowed” to operate the gate himself, but rarely did in the first ten years of residence, because the Agency usually attended during periods of high flow.

Mr Gould’s evidence was that he has rarely opened the gate; essentially he said he has done so only when the Agency told him to. The Agency advises riparian owners when there is a flood alert; its staff no longer do the “gate runs” on which they used to visit and check all the weirs on the River Colne, but it still gives telephone advice and will visit where necessary. Mr Gould gave evidence that the Agency staff would no longer visit since the commencement of litigation, by reference to Agency memoranda to its staff; but it became clear in cross-examination that that was not correct and that Mr Gould had misinterpreted the memos.

We find that Mr Gould could have taken a far more active role in controlling the flow of water in the side channel. The sluice gate is essentially under his control; neither common law nor statute, nor the 1964 Deed nor the Agency’s own policies and advice have taken that control from him. His indignation at the Agency’s withdrawal of support from the mills on the River Colne has led him to ignore his own powers and responsibilities, and his failure to operate the mill is likely to have caused much of the erosion of which he complains.

The claim under the 1991 Act fails. Mr Gould cannot show a tortious act by the Agency, whether in the way the 2001 works were carried out or in the diversion and reduction of the accustomed flow of water under the mill.