Whitehall Has a Number for Net Zero’s Water. It Has None for the Harvest It Restricted.
£50 million to re-wet farmland. 1,506 orders restricting farmers’ irrigation. No official cost for the food already lost.
THE RATIONALS
This is the year water-company boards handed out millions in bonuses and retention payments while the industry still lost nearly three billion litres a day through unrepaired pipes — and the year the government found £50 million to re-wet farmland for the climate, published an exact figure for the water a hydrogen plant will need in 2050, and issued farmers 1,506 orders to take less water in the middle of the earliest harvest on record, without ever producing a single public number for what that harvest cost.
Whitehall has a number for hydrogen’s water three decades out. It has none for this year’s harvest. What follows is the paper trail of that refusal.
“Wetter farming systems” is the sort of phrase a civil servant produces when the plain version — paying people not to grow food on it — would raise questions in a select committee. It appears in Defra’s spring announcement of nearly £50 million for lowland peat, a net-zero land-use commitment inherited from the Climate Change Committee’s own advice.
Alongside it sits “paludiculture,” a word that sounds like a minor Roman deity and means, roughly, farming a bog. The money exists to raise water tables on farmland — not to irrigate crops, but to stop there being crops, more gently than a compulsory purchase order and with considerably better branding.
Three months later, on 29 July, the Environment Agency declared half of England to be in drought. The prose was brisk, as government prose about the weather generally is, a flash drought, unusually low July rainfall, repeated heat. Somewhere inside the brisk prose sat the number that mattered — 1,506 abstraction licence restrictions, up sharply from the start of the year, each one a legal instruction to a farmer to stop taking water the moment the river fell below a line drawn for him by someone in an office he has never visited.
Cereal harvesting had already begun early. Yields, the Agency noted in the same flat tone it uses for rainfall totals, were “already reduced.” Seven water companies had put temporary use bans on 23 million customers, a considerable achievement for a country that had spent the spring finding money to put water back on land nobody was allowed to farm properly.
Two documents. Same government, same year, same substance. Neither refers to the other. The coincidence has been left, with what looks like real diligence, entirely to the reader to notice — which is, in fairness, the traditional British compromise between transparency and shame.
Weather has done most of the damage, and no fair account pretends otherwise. Timing, however, is another matter, and the timing is the story.
By the last week of July the Agriculture and Horticulture Development Board’s first harvest report of the season confirmed 2026 as the earliest since its records began in 2006. Farms across the south and east were finishing two or three weeks ahead of the long-term average, some were working through the night simply to stay ahead of heat that would otherwise have finished the crop for them.
Early wheat surveys averaged 6.8 tonnes a hectare — thirteen per cent below the five-year mark and sitting uncomfortably outside the recent ten-year range. AHDB’s own drought briefing added the quieter details, potatoes likely to be short, tubers more prone to bruising in store, grass growth slowed dramatically and in places stopped altogether, livestock farmers already dipping into winter forage several months early.
Yields fall, farm incomes tighten, and some of the cost is felt later at the till.
None of this required the peat scheme’s involvement. It required only that the scheme happen at the same time — funded, staffed and announced — while the harvest unfolding beside it went uncounted in any comparable way.
It is worth being honest that 2026 is not a freak. English arable farming has been taking comparable hits with grim regularity for more than a decade. After the sodden, sunless summer of 2012 the NFU’s own survey put wheat at 6.7 tonnes a hectare, down fourteen per cent on the five-year average — a figure last seen in the late 1980s, a decade better remembered for other things falling.
Eleven years later the country managed the opposite trick. After the catastrophically wet winter of 2023–24, independent analysis for the Energy and Climate Intelligence Unit (working from AHDB area forecasts and Defra yield data, so nobody can accuse it of invention) put the combined output of wheat, barley, oats and oilseed rape down 17.5 per cent on the previous year and 21.2 per cent against the longer average. Winter wheat alone was off by roughly a quarter.
English weather, it turns out, can ruin a harvest by giving too little water or far too much, and does both with equal commitment to the bit.
The pattern is therefore not an aberration that caught anyone by surprise. A double-digit swing in cereal yields from water stress — too much or too little — has now arrived at least three times in fourteen years.
A state capable of building a national water-deficit model that stretches to 2055 has had every opportunity to notice that its own farmers keep losing something like a fifth of the harvest to water roughly once a decade. It has not yet produced an equivalent model for what that costs, nor a fund with anything like the peat scheme’s administrative speed to soften it.
To its very limited credit, the Environment Agency has at least priced one new claimant properly. Its National Framework for Water Resources 2025 states that carbon capture, storage and hydrogen production — both central to the statutory net-zero target — will together need 767 megalitres a day by 2050. The figure was produced for the regulator and sits inside its own national deficit model, it was not conjured by a pressure group with an axe to grind.
Durham University, working for the water retailer Wave, puts the wider decarbonisation total higher still and flags possible regional shortfalls as early as 2030. Data centres, expanding in many of the same catchments, receive no equivalent figure at all. The Agency has told Parliament, in terms admirably free of self-flattery, that it simply does not have reliable data on what they use.
One industrial appetite has been counted. The other has been waved through on the strength of nobody asking. Both, however, now sit inside the same planning landscape as the farmers whose water was formally restricted this summer.
It would be a mistake to conclude that being counted means being ranked with any seriousness. In the Agency’s own hierarchy of the coming deficit, 767 megalitres a day for hydrogen and carbon capture still sits below the reductions required to protect rivers and chalk streams, and below population growth. It simply arrives wearing a net-zero badge — which is enough to place it inside the official arithmetic while the harvest it competes with remains outside it entirely.
A number, once obtained, appears to function less as the start of an argument about priorities than as a convenient reason to stop having one. The water that grows food this summer was never given a matching figure, and therefore never entered the ranking at all.
There is, in fairness to the state, a long-term answer to all of this, and it even has a name, more storage. The trouble is the calendar. The last major public water-supply reservoir completed anywhere in England was Carsington Water in Derbyshire, opened in 1992, a year in which the internet was a rumour and John Major was still finding his feet.
Havant Thicket — the first genuinely new scheme in more than three decades — received planning permission in late 2021, began construction in spring 2025, and is not expected to deliver a drop until the early 2030s.
The Fens Reservoir, a £2.2 billion scheme intended to serve exactly the East Anglian catchments this article keeps returning to, is aiming for a development consent application in 2026, construction from 2029, and completion somewhere around 2036.
A further scheme for Lincolnshire is pencilled in for 2040. Between them, these projects span roughly fifteen years from the drought that made the case for them to the tap that finally answers it.
Set that timetable against the peat scheme’s own pace — announced and funded within a single financial year — and the comparison becomes almost too neat to be believed, except that both figures are published by the same government and can be checked by anyone with an afternoon to spare.
Water that will stop this year’s crop from drying out is fifteen years away. Water that will help stop the crop existing at all in the same form was found by spring.
A farmer needing irrigation this July cannot defer that need to 2036 the way a reservoir, or for that matter a hydrogen plant, can defer its opening date to match new supply. That is not a subtle asymmetry. It is the entire asymmetry, restated as a construction schedule.
Overlay these pressures on a map of eastern England and the argument stops being rhetorical. The Durham–Wave analysis names the Anglian Water region — most of East Anglia — as facing a decarbonisation-driven deficit as early as 2030. The Agency’s own regional figures already show agriculture, chiefly irrigation, as the largest non-public water user across much of that same territory.
Water UK’s evidence to a parliamentary committee puts the great majority of proposed new data-centre capacity in the same water-stressed east and south-east. Three appetites, one aquifer. The one already under formal restriction belongs, this summer, to the person growing the food.
It is worth pausing on why this particular geography keeps recurring across every version of this story, because it is not an accident of reporting.
East Anglia is flat, well-connected, with relatively cheap land by southern English standards, and sits atop some of the country’s most productive arable soil — precisely the combination that makes it attractive to a data centre developer, a hydrogen investor, and a vegetable grower simultaneously, and precisely the combination that means none of the three can simply relocate to somewhere wetter without losing the advantage that brought them there in the first place.
The region is not unlucky. It is popular, in the specific and unfortunate sense that several incompatible ambitions have chosen to be popular in exactly the same postcode at exactly the same time.
None of this is being decided by one department, which is precisely the point. The Land Use Framework and the peat funding sit with Defra. The 767-megalitre hydrogen and carbon capture figure sits with the Department for Energy Security and Net Zero, feeding into targets set under the Climate Change Act. The abstraction restrictions and the drought declaration are issued by the Environment Agency, which is Defra’s own regulator, and therefore in the peculiar position of restricting the water for a sector its parent department is simultaneously re-wetting for entirely separate reasons.
Reservoir delivery is a matter for the water companies, overseen by Ofwat. Four institutions, four sets of published figures, and — on the evidence of the documents themselves — no visible mechanism by which any of them is required to read what the others have written before publishing their own.
This is not a conspiracy. It would almost be more reassuring if it were, since a conspiracy at least implies someone in a position of authority has noticed the whole picture and decided what to do about it.
What the paper trail actually shows is closer to four separate, perfectly reasonable, perfectly well-evidenced departmental positions, each one correct within its own remit, arriving at a combined result that nobody appears to have signed off on as a combined result at all.
Coherence, in this system, is not something that gets decided. It is something that is assumed to have happened somewhere upstream, by someone else, in a meeting nobody minuted.
Wetter peat also fights climate change, runs the first, and this is true as far as it goes — carbon storage, flood mitigation, a genuine ecological case that few dispute in principle, even if they dispute the trade-offs.
What the documents never do is put that case on the same page as this year’s 1,506 abstraction restrictions and let the two arguments meet. They have been kept, with what can only be deliberate tidiness, in separate folders, with separate budgets, and with separate ministers standing in front of separate podiums. The climate benefit is announced. The food cost is not even calculated. That is not balance. It is avoidance.
At least the 767 megalitres is a real number, runs the second, which is true — and better than the data-centre alternative, which is no number at all. A number is not an argument. Nobody has yet explained, in public, why water earmarked for 2050 hydrogen should outrank a potato that needs it this summer. The figure is treated as settled. The harvest loss is treated as weather. Only one of those treatments required a decision.
It is only fifty million over four years, runs the third, and the reply almost writes itself. The sum was never the point. The point is that Whitehall found the administrative energy to design, fund and announce a scheme to re-wet farmland for the climate inside the same season it was formally rationing water for food — and found none of that energy, at any comparable speed, for the production side of the ledger. The state can move quite fast when the demand wears the right badge. It moves markedly slower when the demand is a crop.
Reservoirs are coming, this is just a timing problem, runs the fourth, and it is the defence that concedes the most while appearing to concede the least. A “timing problem” that runs from a 1992 reservoir to a 2036 one, spanning three governments and at least four droughts severe enough to be named in official reports, is not a timing problem in any sense recognised outside Whitehall. It is a planning failure that has had the good manners to describe itself with a word that implies patience will fix it. Patience does not irrigate potatoes.
Two arms of government can now tell you, to the megalitre, how much water Britain’s hydrogen ambitions will need by 2050. The same government can tell a farmer, to the litre, the exact flow at which his licence cuts him off this afternoon.
What it still cannot tell anyone — because it has never tried to produce the figure — is how many tonnes of British wheat, potatoes, onions and forage the summer of 2026 has already cost.
A state fluent in forecasting appetites it has not built yet remains, on the matter of harvests it is actively restricting, entirely innumerate by choice. That is not an oversight fifty years of civil-service reform has failed to reach. It is a choice about which numbers are worth the paperwork, made afresh every time.
Until the missing number exists, the claim that net zero, nature recovery and food security add up to one coherent programme is not an argument. It is a hope, filed in triplicate, under three different departments, none of which has read the others’ paperwork, and none of which — on the evidence available — has yet been asked to.
(UKR)