In December 2023 we argued for a new Bristol Tree Replacement Standard (BTRS) calibrated to the statutory biodiversity metric. That argument was about making the local standard consistent with national policy. It’s now about something much more critical: on half of all sites, a local standard is the only thing left to protect trees being lost to development.
What changed on 6 August 2026
The Biodiversity Gain (Town and Country Planning) (Amendments and Transitional Provisions) (England) Regulations 2026 (SI 2026/790) took effect for planning applications made on or after 6 August 2026. The three changes made matter here:
- A site of 0.2 hectares or less, measured by the red-line boundary of the application, is exempt from mandatory BNG unless on-site priority habitat is negatively affected.
- Development that’s wholly temporary, with permission granted for five years or less, is also exempt on the same proviso.
- For minor development above the area threshold, the mitigation hierarchy has been relaxed: off-site gains are now seen as just as desirable as created or enhanced habitat on the site itself, so a developer may go straight off site without first showing that on-site provision is unachievable.
The self-build and custom-build exemption has also been removed, on the reasoning that most such schemes now fall inside the new small-sites exemption anyway. While applications already submitted before 6 August continue under the old rules, Defra in its 15 April 2026 response to the consultation, estimated that ‘the new area based exemption for sites of less than 0.2 hectares is expected to remove more than 50% of the smallest planning applications which are currently delivering mandatory BNG, from the BNG requirement’.
Trees aren’t priority habitat
The priority habitat proviso sounds protective, and for some sites it is, but it does nothing for trees. Depending on where it is, an individual tree is recorded in the statutory metric as urban tree or rural tree, both of medium distinctiveness; neither is a habitat of principal importance under section 41 of the Natural Environment and Rural Communities Act 2006. Only ancient and veteran trees sit outside these two categories, as irreplaceable habitat, and they are a vanishingly small proportion of the trees felled for development.
So a mature street-side lime, a garden oak, or a row of maturing London planes on the edge of a school playing field — none of these will be protected if they’re in the way of a 0.2 hectare development. Below the threshold, they can be removed without any obligation to count them, replace them, or compensate for their loss anywhere. Unless the tree is protected by a Tree Preservation Order, or unless the local plan says otherwise, nothing now requires their loss to be made good.
This is the gap; and it’s much larger than the word ‘small’ implies.
Half of all development sites now fall outside the system
We measured site area for every gain plan in our dataset that came with a usable metric calculation — 305 sites across 59 authorities. Site area here is the baseline area-habitat total, net of any individual trees, taken from the statutory metric or, for small-sites metrics.
| Site size | Sites | Share | Baseline area HUs | Post-development area HUs | On-site change |
|---|---|---|---|---|---|
| 0.2 ha or smaller | 151 | 49.5% | 48.79 | 48.67 | −0.25% |
| Over 0.2 ha | 154 | 50.5% | 506.88 | 512.82 | +1.17% |
| All sites | 305 | 100% | 555.67 | 561.49 | +1.05% |
Habitat units are on-site figures from sections 6.3 and 6.4 of the biodiversity gain plan. Hedgerows and watercourses are excluded as linear features.
Just under half of the sites that have been going through the BNG process fall at or below the new exemption threshold — consistent with Defra’s own estimate. Two further things stand out in that table. First, small sites were barely holding their own on site even under the old regime that’s just been withdrawn: a baseline of 48.79 area habitat units became a mere 48.67 after development, a small net loss, with any gain bought off site from elsewhere. Second, the small-site total is dwarfed by the large-site total, which is exactly why the exemption looked politically cheap. In habitat unit terms these sites are a rounding error. In tree terms, though, they aren’t.
Small sites are where the trees are
Of the 320 metric calculations we could read, 163 relate to sites of 0.2 hectares or less. Sixty of these — 36.8% — record individual trees in the on-site baseline. Between them they carry 3.12 hectares of tree canopy ‘oversailing’ 15.22 hectares of land.
| Site size | Metrics | With trees | Site area (ha) | Canopy (ha) | Canopy as % of site area | Tree baseline HUs |
|---|---|---|---|---|---|---|
| 0.2 ha or less | 163 | 60 (36.8%) | 15.22 | 3.12 | 20.5% | 26.71 |
| Over 0.2 ha | 157 | 81 (51.6%) | 219.64 | 13.79 | 6.3% | 133.22 |
| All sites | 320 | 141 (44.1%) | 234.85 | 16.91 | 7.2% | 159.94 |
Site area is net of individual trees, because tree canopy oversails the habitat beneath it and is not additional land. Canopy can therefore exceed 100% of the site area it overlies.
That last comparison is the heart of it. On sites above the threshold, tree canopy is equivalent to about 6% of the site area. On sites at or below 0.2 hectares it’s 20.5% — more than three times the share. Small sites are disproportionately tree rich, probably because they’re disproportionately urban: they include infill plots, back gardens, corners of school and hospital grounds, and the fringes of existing built-up land. These are the places where we usually find the mature tree canopy of a town.
The concentration is sharper still for twenty of the 163 small-site metrics, which record canopy covering more than half the site area. Seven record canopy exceeding the whole of it. One example is a 0.18 hectare site at 22 York Road in Sutton, which carries 0.33 hectares of canopy across 18 recorded tree entries — canopy equal to 182% of the site, and worth 3.42 habitat units. At an average of 0.019 hectares of canopy each, these are mature trees, not saplings — on the replacement standard set out below, losing 18 medium trees would require 90 replacements. On an application made today, that site would be exempt and the trees could be removed without any obligation to replace them under the new BNG regime.
Expressed as trees rather than hectares, the 3.12 hectares of canopy on small sites is equivalent to 762 small trees, using the 0.0041 hectares the June 2026 metric assigns to a small tree. Because much of that canopy belongs to medium, large and very large trees, the actual number of trees is likely to be lower — and their replacement cost, in the metric’s own currency, correspondingly higher.
And they’re where the new trees were coming from
The exemption removes an obligation to plant as well as an obligation to compensate. Of the 163 small-site metrics, 72 record individual trees created as part of the post-development scheme.
| Site size | Metrics | With tree creation | On-site trees | Off-site trees | Total trees |
|---|---|---|---|---|---|
| 0.2 ha or smaller | 163 | 72 (44.2%) | 597 | 8 | 605 |
| Over 0.2 ha | 157 | 97 (61.8%) | 3,135 | 205 | 3,340 |
| All sites | 320 | 169 (52.8%) | 3,731 | 213 | 3,944 |
Tree numbers are the recorded habitat creation area divided by 0.0041 ha per small tree, rounded up. Off-site units bought from a registered gain site do not appear here, because they sit in the seller’s metric.
Over 600 trees across a dataset of a few hundred plans covering barely two years and 59 authorities out of 306. If you scale that across every planning authority in England and across the years ahead, the tree planting foregone is substantial.
BNG was never going to save urban trees anyway
It’s worth being clear that the case for a local tree standard doesn’t rest solely on the exemption. Even where BNG still applies, it often doesn’t keep compensation anywhere near the trees that were lost.
The Biodiversity Gain Sites register as of 19 August 2026 — 2,328 developments across 306 authorities — shows that 87% of developments in urban authorities offset outside their own authority area. Those authorities needed 1,835 area habitat units and hosted 889. In London the pattern is near total: 98% of the 113 registered London borough developments offset elsewhere, requiring 46.8 units and hosting 3.3.
The units are real, and the receiving sites may well be excellent. But a habitat unit delivered in a rural county doesn’t provide shade in Bristol, intercept its rainfall, cool it in a heatwave, or stand outside anybody’s window. Urban tree canopy is a local public good, yet BNG systematically moves its replacement to some distant field that few, if any, will ever visit.
Why we need a tree replacement standard
A tree replacement standard is a local plan policy, not a national biodiversity calculation. It attaches to the tree, not to the site area; it applies whether or not a metric has to be completed; and it keeps the replacement in or close to the place where the loss occurred. It’s the mechanism best suited to fill the gap the exemption has opened.
This is not a novel invention. Bristol has operated the BTRS since 2014 under Core Strategy policy BCS9, and a replacement table is carried forward into proposed local plan policy BG4. What we argued in 2023 was that the ratios should be recalibrated so that the standard delivers what the statutory metric would have delivered, plus the 10% uplift. Our proposal was not adopted then, though our argument stands. What’s changed is who needs it: not just Bristol, but every planning authority that wants any control at all over tree loss on the majority of its planning applications.
We would encourage authorities without such a policy to adopt one, and those with one to check that it’s calibrated to the metric rather than to a historic rule of thumb whose rationale was and remains hard to understand. As the standard is a local plan policy, it can be applied to exempt sites without cutting across the national BNG regime: BNG sets a floor for biodiversity, and a replacement standard sets a separate, local requirement for canopy.
Recalculating the standard against the June 2026 metric
Our 2023 proposal derived replacement numbers from the tree area equivalents then in force. Those figures have since changed. Table 15 of the Statutory Biodiversity Metric User Guide (June 2026), at page 62, now assigns the areas set out below. Applying the same method — the area of the lost tree divided by the area of one small tree, since a newly planted tree is recorded as small, plus the 10% net gain, rounded up because a fraction of a tree is not a tree — gives revised numbers.
| Tree size (DBH) | Area equivalent (ha) | Multiple of a small tree | Plus 10% net gain | Replacement trees | 2023 proposal |
|---|---|---|---|---|---|
| Small (>7.5–30 cm) | 0.0041 | 1.00 | 1.10 | 2 | 1 |
| Medium (>30–60 cm) | 0.0163 | 3.98 | 4.37 | 5 | 3 |
| Large (>60–90 cm) | 0.0366 | 8.93 | 9.82 | 10 | 10 |
| Very large (>90 cm) | 0.0765 | 18.66 | 20.52 | 21 | 20 |
Areas are from Table 15 of the Statutory Biodiversity Metric User Guide (June 2026), p62. Replacement numbers are the tree’s area divided by the area of one small tree, plus the 10% net gain, rounded up.
Two categories of tree size move. A medium tree is now worth almost four small trees rather than three, so its replacement requirement rises from three to five. A very large tree rises from 20 to 21. Large trees are unchanged at ten.
The small-tree row needs a decision rather than an arithmetic answer. One small tree replaced by 1.1 small trees rounds up to two, which is a real doubling for the commonest category in the surveyed population. The alternative is to hold the small category at one-for-one and take the 10% uplift from the scheme as a whole. We think the honest position is to state the arithmetic and let the policy make the choice explicitly, rather than round the uplift away in silence.
What we’re asking for
- That local planning authorities adopt, or retain and strengthen, a tree replacement standard in their local plan, applying to all development including sites exempt from mandatory BNG.
- That the replacement ratios be calibrated to the current tree area equivalents in the statutory metric, so that a local standard delivers no less than the national regime would have done.
- That replacement be required within the same authority area and as close to the loss site as is practicable, so that canopy is replaced where people live.
- That authorities continue to require a BS 5837:2012 tree survey with applications on exempt sites, so that the loss is properly recorded even where no metric is required.
- That Defra keep the 0.2 hectare exemption under review, and specifically consider whether the priority habitat proviso should be extended to sites carrying significant tree canopy.
The exemption was introduced to reduce costs and delays for small applications. Whatever one makes of that aim, it shouldn’t be paid for with the mature canopy of our towns and cities. A tree replacement standard is a cheap, well-understood, locally enforceable way of making sure it’s not.
About the data
Figures are drawn from the Bristol Tree Forum biodiversity gain plan dataset (v11): 505 gain plans logged from 59 planning authorities, of which 320 carry a readable statutory or small-sites metric calculation and 305 a usable site area. Plans are dated 2024 (58), 2025 (254) and 2026 (165). Register figures are from the Biodiversity Gain Sites register extract of 19 August 2026, covering 2,328 developments and 2,686 allocations across 306 authorities.
Two caveats:
- Site area in this analysis is the metric’s baseline habitat area net of individual trees, which is a close proxy for the red-line boundary the exemption uses but not identical to it; a handful of sites near the threshold may fall on the other side of it in practice.
- The sample is of applications that produced a gain plan, so it describes the population the exemption removes rather than all development across the country.
