Solar on England’s Best Farmland: Allowed by the Rules, Rarely on Grade 1

The worry runs like this: solar farms are going onto England’s best farmland. CPRE, the countryside charity, calls for “a ban on ground-mounted solar on scarce Grade 1 and 2 farmland”, the top two of the five grades. The focus on Grade 1 alone is ours: our own research brief for this piece started from it, and its example, Cleve Hill in Kent, was “Grade 1 Kent farmland”.

We went to the record: eight solar schemes, seven of them through their own planning documents, the only census of England’s large solar farms we could find, and Natural England’s land-grade data, queried directly.

Part of the worry holds. Solar on best and most versatile land, the planning category that covers Grades 1, 2 and 3a, is real and in places large: one consented scheme in Lincolnshire puts its arrays across 257 hectares of it. The Grade 1 part holds only in a few places. Grade 1 is the rarest case inside that category: two of the eight schemes record it in a source we read, 106 hectares between them, and Cleve Hill has none.

Nor were these schemes slipped past the rules. Every scheme here passed the planning test in force when it was decided, and none of the policy texts we read makes Grade 1 a bar. If solar on the best land is a problem, the place to look is the policy, not the decisions made under it.

What “best and most versatile” means

Agricultural Land Classification (ALC) grades land from 1 to 5. It is one system for England and Wales, applied under criteria the agriculture ministry, MAFF, set in 1988. Planning in England treats the top three as one category: “BMV agricultural land is graded 1 to 3a,” in the words of Natural England’s guide for planners.

The grade measures what land can do, not what it is doing. Natural England’s technical note TIN049 is explicit: “The current agricultural use, or intensity of use, does not affect the ALC grade.” A field of Grade 1 under grass is still Grade 1.

How scarce Grade 1 is, and on which map

Only one national dataset gives every hectare of England a grade: Natural England’s provisional ALC map. Queried on 27 September 2026, it puts 354,585 hectares in Grade 1. That is 2.7% of England’s land area, or 3.1% of the land the map grades as agricultural.

Against today’s farmed area the share is higher. Defra counts 8.8 million hectares of utilised agricultural area in England in 2026, and the same Grade 1 hectares are about 4% of it. That ratio divides one survey by another taken 52 to 59 years later on different definitions, so it is a reading, not a measurement.

The map itself is old. It rests on a one-inch-to-the-mile series published between 1967 and 1974, which shows only five grades because its preparation “preceded the subdivision of Grade 3”. Of those sheets, TIN049 says “They have not been updated and are out of print”, and that they are “not sufficiently accurate for use in assessment of individual fields or development sites”. The 1:250,000 map queried here is, in the same note’s words, “based on the same information”, and Natural England’s current guide says it too is “not at a scale suitable or accurate for assessment of individual fields or sites.”

The best evidence of a field’s grade is a detailed site survey, at “one boring per hectare for a detailed assessment”, which TIN049 calls “the most definitive source”. It also says “There is no comprehensive programme to survey all areas in detail.” Natural England’s archive of detailed surveys since 1988 covers 284,173 hectares, 2.18% of the area the provisional map covers.

Two lines need correcting here. “Less than 3% of total farmland”, a description of Grade 1 our own research brief started from, is wrong as worded: under 3% is its share of all of England on the old map, towns included; it is not a share of farmland on any denominator above. And the map cannot total BMV at all, because it has no Grade 3a. The national figure for the top grades is Natural England’s, from TIN049 in December 2012: “Grades 1 and 2 together form about 21% of all farmland in England; Subgrade 3a also covers about 21%.” Those are estimates, the note does not give their method, and they should carry their date.

How much Grade 1 is under solar

CPRE’s census

The only census of operational large solar farms we found is by CPRE, a campaigning charity that calls for “a ban on ground-mounted solar on scarce Grade 1 and 2 farmland”. Its July 2025 report, Getting solar off the ground, covers 38 operational farms over 30 MW in England, from the government’s Renewable Energy Planning Database as it stood in June 2025.

Grade 1 appears on two of the 38: 45.4 hectares, 1.7% of the 2,668.3 hectares those farms occupy. We use 45.4 because it is the figure in the report’s Table 2 and the sum of the rows in its appendix; the report’s text says 45.0 and its press release 45.

Part of that figure is pro-rated. 39.4 hectares of it is Sutton Bridge, where CPRE applied the site survey’s Grade 1 share, 33.3%, to the database’s 118.3 hectare site area. The survey itself records 48.0 hectares of Grade 1 on the 144 hectares it covered.

CPRE’s release words its 59% as a share of farms. In the report it is a share of area, 31% BMV plus 28% Grade 3b, which is not BMV. By site, on our count of the report’s appendix, 25 of the 38 have any BMV or 3b. Across all 38, BMV comes to 827 hectares, against Grade 1’s 45.4.

The release’s headline that three farms are “located entirely on highest-grade farmland” needs the same care. Its own text says “entirely on BMV farmland”, and in the report’s site-by-site table two of the three, Goosehall and Black Peak Farm, have no Grade 1 at all: their land is Grades 2 and 3a.

The census is a snapshot: on our count, the July 2026 edition of the same database lists 56 operational ground-mounted sites over 30 MW in England.

Eight schemes, each from its own record

The eight schemes came from the case our brief named, CPRE’s appendix and recent national decisions, not from a search for Grade 1.

Scheme	Decided by, and when	Capacity	Grade 1 recorded
Cleve Hill, Kent	Secretary of State (BEIS), 28 May 2020	about 350 MW consented, 373 MW built	none
Sutton Bridge	South Holland District Council, 12 March 2021	49.9 MW	48.0 ha
Larport, Herefordshire	Herefordshire Council; record not read	45 MW	6 ha, on CPRE’s figures only
Mallard Pass	Secretary of State (DESNZ), 12 July 2024	350 MW	none
Sunnica	Secretary of State (DESNZ), 12 July 2024	500 MW	not stated
Heckington Fen, Lincolnshire	Secretary of State (DESNZ), 24 January 2025	500 MW	58 ha
Springwell	Secretary of State (DESNZ), 8 April 2026	800 MW	not stated; Grade 1 and 2 fields kept in arable
One Earth	Secretary of State (DESNZ), 8 July 2026	740 MW	none

Capacities for the national schemes and Larport are the database’s installed figures, July 2026; Sutton Bridge’s is its application’s. Larport is in CPRE’s census of operating farms, so it was permitted; how its council weighed the land was not read.

The two with Grade 1 in a record we read:

Sutton Bridge, approved by the council as a “Change of use from agricultural land to solar farm”. The site survey, by Soil Environment Services for Arcus and dated 3 April 2020, found 48.0 hectares of Grade 1 (33.3%) and 96.0 hectares of Grade 2, nothing lower.
Heckington Fen. The decision letter records the applicant’s surveys of the 524 hectare solar park site: 58 hectares of Grade 1, 11.1%, and 39 hectares of Grade 2.

Together that is 106 hectares, for scale about 0.03% of the Grade 1 on the national map. Larport would add 6 hectares on CPRE’s word; its own survey and decision were not read.

A wider screen, our own, points the same way. We tested the grid reference recorded for each of England’s 1,060 operational ground-mounted sites in the July 2026 database against the Grade 1 areas of Natural England’s national map. 36 points fall inside one: 3.4% of sites and 3.1% of capacity (304.4 of 9,746.8 MW), close to Grade 1’s 2.7% share of England. On this screen, solar is neither avoiding nor seeking those areas. This is not a count of Grade 1 solar: a point on a 1:250,000 map says nothing about the grade of a field.

Cleve Hill, the case we started with, has no Grade 1

Cleve Hill Solar Park in Swale, Kent, was our research brief’s example of Grade 1. The examining authority’s report of 28 February 2020 records the survey: “The ES estimated that the Proposed Development would affect approximately 370ha of arable land, comprising approximately 2ha of ALC Grade 2, 9ha of ALC Grade 3a, and 360ha of ALC Grade 3b.” The soil was held back by wetness, which “limits agricultural quality to ALC Grade 3b.” The arable land is drained marsh, “progressively drained from the medieval period.”

An objector contested the survey throughout. His final figure, revised down from an earlier claim of “over 75%” of the land, was 164.7 hectares of Grades 2 and 3a, never Grade 1. The examining authority found the survey “robust”, and the Secretary of State’s decision of 28 May 2020 gave the loss of agricultural land “little weight”.

Cleve Hill’s size depends on which area is meant. The consented site is 491.2 hectares, the land under panels “around 176 hectares”, and the planning database records a solar site area of 364 hectares. Capacity was consented at “around 350 megawatts”; the operator gives 373 MW built, operational in summer 2025. Scaled to 70 GW on the database’s 364 hectares for 373 MW, that is about 68,000 hectares, the figure our comparison of solar and nuclear land use now uses.

The bigger story is BMV, and Lincolnshire

Grade 1 is tens of hectares per site. The BMV figures are of another order.

Heckington Fen: 257 hectares of the solar park site, 49%, is BMV, in the decision letter’s account, and the arrays are “proposed across the 257 ha of BMV land”. A further 62 hectares of Grade 1 and 2 land was taken out of the site before the application.
Mallard Pass: 360 hectares of BMV inside the scheme boundary, 100 of them Grade 2; 216 hectares of the BMV lies within the array areas and field margins, 35 of them Grade 2. The decision letter records that the applicant chose Lincolnshire partly for “the relative lack of Grade 1 agricultural land”.
One Earth: 660.9 hectares of BMV, all Grade 2 and 3a. The examining authority called it Grades 1 and 2; the Secretary of State’s letter corrected it, saying it had “incorrectly referenced” the grades. A Grade 1 claim in any document short of the survey is a claim to check.
Springwell: about 541.2 hectares of BMV, not split by grade, with fields “that comprise solely of Grade 1 or 2 BMV land” kept for arable production.
Sunnica: the applicant’s surveys found 37.3 hectares of BMV used; objectors, using Natural England’s predictive map instead, argued that, as “a reasonable worst case”, 82% of the scheme area was 60% or more likely to be BMV. The Secretary of State followed the surveys, as Natural England, the statutory adviser, had accepted them.

Concentration is a separate question from grade. The Heckington Fen applicant, as its decision letter records, put the national schemes in Lincolnshire at about 9,743 hectares of panels, about 2,597 of them BMV as a worst case. Mallard Pass’s applicant, in that scheme’s letter, put those in Lincolnshire and Rutland at about 2,114 hectares of BMV, “around 0.5%” of the BMV there. CPRE says solar already covers 7% of the land in the Sleaford and North Hykeham constituency, without stating its method.

Every scheme here passed the test in force

Who decides depends on size and date. Until 30 December 2025, a solar scheme in England over 50 MW was a nationally significant project decided by the Secretary of State; since 31 December 2025 the threshold is 100 MW, measured in AC under the revised EN-3. Below it, the council decides. Sutton Bridge, at 49.9 MW, went to South Holland.

The test has been restated several times since 2011. None of the four versions below, the ones we read, makes Grade 1 a bar. (The version of the Framework in force when Sutton Bridge was decided in 2021 was not read.)

2011: National Policy Statement EN-1, applied at Cleve Hill, said applicants “should seek to minimise impacts on the best and most versatile agricultural land”.
15 May 2024: a written ministerial statement said “the highest quality agricultural land is least appropriate for solar development and as the land grade increases, there is a greater onus on developers to show that the use of higher quality land is necessary.” Five of the schemes here were decided after it, and each letter cites it: Mallard Pass and Sunnica in July 2024, Heckington Fen in January 2025, Springwell and One Earth in 2026.
6 January 2026: the revised EN-3 says poorer land “should be preferred”, “avoiding the use of ‘Best and Most Versatile’ agricultural land where possible”, and that solar on BMV “is not prohibited”.
17 August 2026: the National Planning Policy Framework tells councils to take into account land quality “and its grade” and to “use areas of poorer quality land where possible”. Its Annex A lists the May 2024 statement as incorporated into the Framework “or no longer represents up-to-date government policy”, without saying which.

Each national decision read applied the test of its day. Cleve Hill’s Grade 3b loss was “neutral in the planning balance” to its examiners. At Mallard Pass, the examiners gave the harm “a little weight” and the Secretary of State raised it to “moderate negative weight”, while agreeing the effect on food production “in the national context is negligible”. Sunnica’s examiners gave BMV “moderately negative weight”; the Secretary of State disagreed and lowered it to “slightly negative weight”. At One Earth, the Secretary of State found the applicant had “reasonably evidenced the required use of BMV land” and gave effects on agriculture and soils “a moderate negative weighting”.

Heckington Fen, the one national scheme here with Grade 1, was also weighted lightly: “a little negative weight” for BMV, in its decision letter. The Secretary of State accepted the applicant’s case on alternatives, noting the site “is owned by a single landowner and accompanied by an available grid connection at Bicker Fen”, and gave the cumulative effect across Lincolnshire “moderate negative weight”. The letter counts 2.8 hectares of BMV as a permanent loss.

No national decision read here was refused, and none turned on Grade 1. Sutton Bridge’s officer report, on the council’s planning portal, exists only as scanned images and was not read; what is known is that the council approved it.

The case for a stricter test

CPRE’s position deserves stating at full strength, because it is the only criterion read here under which these schemes fail. Its proposed ban covers Grades 1 and 2, not Grade 1 alone. Applied as written:

Three fail on land a source puts under panels: Mallard Pass (35 hectares of Grade 2, with field margins), Heckington Fen (arrays across the BMV that holds its Grade 1 and 2), and Sutton Bridge, whose survey found nothing below Grade 2, so any panel stands on Grade 1 or 2.
Three more include Grade 1 or 2 in the scheme without a source putting it under panels: Cleve Hill (2 hectares of Grade 2 among 370 “affected”), One Earth, and Larport, on CPRE’s figures.
Two cannot be judged: Springwell and Sunnica, whose BMV is not split by grade.

On that test, failures are not rare. They are mostly Grade 2.

The second argument is about evidence. The grade that sets how hard a scheme is tested comes from a survey the applicant commissions. The 2024 statement itself proposed certification, citing “concerns about the perceived inaccuracy and unfairness of soil surveys”. The department that decides national solar schemes also sets the solar targets they meet. Against that, Sutton Bridge’s survey graded a third of its own client’s site Grade 1, a finding that did not favour the client.

What the trade-off is, in the government’s figures

The government’s land figures come from the department promoting solar, and they are UK-wide. DESNZ estimates “up to 0.4% of total UK land” for solar at its Clean Power range of 45 to 47 GW, and says “we do not believe that increased solar deployment poses a threat to food security”. The 2024 statement said solar would occupy “less than 1% of the UK’s agricultural land” even in the most ambitious scenarios.

England’s measured figures are incomplete. Defra’s June Survey found 9.4 thousand hectares of land on farm holdings used for solar in 2026, 4.4 thousand of them also grazed or farmed, but it excludes large solar parks not on a registered holding. The planning database records a site area for 794 of England’s 1,060 operational ground-mounted sites, and none for the rest; summed, on our count, those areas come to 15,731.1 hectares. Neither is the figure, and neither says which grade the land is: the database has no land-type field.

No national food-production figure for solar was found, only applicants’ own. The Heckington Fen applicant, in the Secretary of State’s account, said its site “could produce 4200 tonnes of wheat a year – 0.03% of the national production”, without defining “national”, and estimated that moving to poorer land would cost about 360 tonnes of winter wheat. Those are an interested party’s numbers.

What this does not show

Eight schemes are a sample, not a census. The other decided national schemes in the database, from Little Crow to Peartree Hill, were not read, nor any council decision except Sutton Bridge. The 106 hectares of Grade 1 is a floor.
Botley West, an 840 MW scheme in Oxfordshire, is undecided, with its deadline moved to 10 November 2026. Its land grades were not read.
Every grade here comes from a survey, a decision letter’s account of one, or CPRE’s appendix. Nothing here re-surveys a field.
Policy has moved since most of these decisions. A national scheme decided under EN-1 in 2020 was not tested against the 2025 EN-3, which governs national schemes now (the Framework does not set policy for them), and a council scheme of 2021 was not tested against the 2026 Framework. Nothing here says how either would fare today.

Sources

Natural England, Guide to assessing development proposals on agricultural land, updated 24 July 2026
Natural England, Agricultural Land Classification: protecting the best and most versatile agricultural land (TIN049), second edition, 19 December 2012
Natural England, Provisional Agricultural Land Classification (England), data service, queried 27 September 2026
Natural England, Agricultural Land Classification Post 1988, data service
Defra, Agricultural land use in England at 1 June 2026, 24 September 2026
CPRE, Getting solar off the ground, July 2025
CPRE, Two-thirds of mega solar farms built on productive farmland, 13 July 2025
DESNZ, Renewable Energy Planning Database: July 2026
Examining Authority, Cleve Hill Solar Park: Report to the Secretary of State, 28 February 2020
BEIS, Cleve Hill Solar Park decision letter, 28 May 2020
Cleve Hill Solar Park Ltd, project website
South Holland District Council, planning application H18-1126-20 (Sutton Bridge)
Soil Environment Services for Arcus, Agricultural Land Classification, Sutton Bridge, 3 April 2020
DESNZ, Heckington Fen Solar Park decision letter, 24 January 2025
DESNZ, Mallard Pass Solar Farm decision letter, 12 July 2024
DESNZ, Sunnica Energy Farm decision letter, 12 July 2024
DESNZ, Springwell Solar Farm decision letter, 8 April 2026
DESNZ, One Earth Solar Farm decision letter, 8 July 2026
Planning Act 2008, section 15, as amended from 31 December 2025
Written ministerial statement HCWS466, “Solar and protecting our Food Security and Best and Most Versatile (BMV) Land”, 15 May 2024
DESNZ, National Policy Statement for renewable energy infrastructure (EN-3), 2025, in force from 6 January 2026
MHCLG, National Planning Policy Framework, August 2026
DESNZ, Solar roadmap, 30 June 2025
Written ministerial statement HCWS304, “Energy Infrastructure Planning Projects”, 2 September 2026