Green roofs in the metric: substrate depth and scoring

A green roof earns units in the statutory biodiversity metric only if it is built as habitat, and most of the roofs drawn on planning applications are not. The difference is substrate: depth, composition and variation across the roof. A sedum mat rolled out on a thin growing layer supports a handful of species and dries out every summer, while a biodiverse build-up with varied depth, sand and gravel mounds and bare areas supports invertebrates and ground-nesting birds. Both look green on a visualisation. Only one of them will hold the condition the gain plan assumed at year five.

How roof habitats are classified

The statutory metric places roof habitats in its urban habitats, with a condition assessment sheet that sets out the criteria the roof has to meet. Type names, distinctiveness bands and criteria have changed between versions of the metric, so read the current published version rather than relying on a schedule copied from an earlier scheme.

What has stayed constant is the logic. A roof scores as the habitat it actually is, judged on what has been built and what is growing, not on what the specification says. An assessor visiting in year three is looking at a roof, and if the substrate is 30 mm deep and the vegetation is a uniform sedum monoculture, that is what gets recorded.

The baseline matters too. On a demolition and rebuild the existing roof is normally scored as developed land with no vegetation, so the new roof is a gain. On a retrofit over an existing biodiverse roof, the baseline is higher and the gain is smaller. A Preliminary Ecological Appraisal that records the existing roof properly is what the calculation is measured against.

Substrate depth and what it buys

Depth is the single variable that does most of the work. In the UK sector, a biodiverse roof is normally specified at a minimum of around 80 mm of substrate, varied across the roof up to 150 mm or more in mounds, while a sedum mat system typically sits on 20 mm to 40 mm. That difference decides water holding capacity, root volume, thermal buffering and whether anything survives a dry August.

Variation matters as much as the average. A roof at a uniform depth produces uniform vegetation. A roof with shallow areas, deeper mounds, exposed aggregate, log piles and small areas of bare substrate produces the structural variety the condition criteria reward and the invertebrate assemblage that follows.

Three specification points are worth fixing early:

• Substrate composition, which should be a crushed mineral aggregate with a low organic content rather than a topsoil-based mix that will compact and support rank grasses.
• Depth profile, drawn as a plan showing the ranges across the roof rather than quoted as a single figure.
• Seeding and plug planting of species suited to a thin dry substrate, with areas left to colonise.

What the depth does to the structure

This is where the argument usually happens, and it happens with the structural engineer rather than with the ecologist. Saturated substrate is heavy, and moving from a thin sedum system to a varied biodiverse build-up adds a real load across the whole roof plate, concentrated further where mounds are proposed.

On a steel or timber frame that load has to be designed in from the outset. Retrofitting depth onto a roof designed for a sedum mat is usually impossible without strengthening, which is why a scheme that promises a biodiverse roof at planning stage and details it at technical design frequently ends up with the thin version. Fix the loading allowance in the structural brief at concept stage, and state the assumed depth range in the application.

Falls, upstands and drainage layers all need coordinating at the same time. A roof with a 1:80 fall and a 150 mm mound at the high point behaves differently from the flat model in the ecologist's head.

Access, irrigation and the maintenance you can actually deliver

A habitat roof needs annual maintenance: removal of woody seedlings, control of dominant grasses, replacement of failed planting, and a check on drainage outlets. That needs safe access, which means a permanent means of reaching the roof, an edge protection system or mansafe line, and a route for bringing materials up.

Irrigation is the awkward one. A roof specified as drought-tolerant should not need permanent irrigation, and a roof that does need it is arguably the wrong specification. But the establishment period does need watering, so plan for a temporary supply in the first two growing seasons and say so in the management plan.

The 30-year management commitment under the Environment Act 2021 regime applies to roof habitat as it does to anything else, and a habitat management and monitoring plan that ignores access arrangements is not deliverable. Councils are increasingly checking for exactly that, in line with the duty explained in our note on the biodiversity duty and how councils must apply it.

Urban greening policies are a separate calculation

Several authorities, and London in particular, operate an urban greening factor that scores each surface type by area and weighting to produce a single figure for the site. It is a different calculation from the biodiversity metric, with different weightings and a different purpose.

The two overlap on roofs and the overlap causes confusion. A roof can score well on an urban greening factor and add few biodiversity units, because the greening factor rewards area and the metric rewards habitat condition and distinctiveness. Run both calculations and present them separately, labelled, so an officer is not left to work out which figure relates to which policy. A Green Infrastructure Plan is the usual place to set out the greening side of the argument.

Evidence to submit at planning stage

The submission that survives scrutiny contains four things. A roof plan showing substrate depth ranges and the position of mounds, aggregate areas and features. A specification covering substrate composition, seed mix and planting. A structural note confirming the loading has been allowed for. And a draft management and monitoring plan with the access arrangements described.

What tends to be submitted instead is a landscape drawing with a hatch labelled as a biodiverse roof and a sentence in the ecology report claiming a condition score. That is the version that fails at discharge of conditions.

Overstatements to avoid

Four claims recur and all four are checkable. Claiming good condition for a roof specified at sedum mat depth. Claiming a habitat type the substrate cannot support. Claiming units for a roof that the structure cannot carry, which surfaces when the engineer's calculations arrive. And counting the same roof twice, once in the metric and once as compensation for something else.

Where the numbers depend on assumptions about depth, access or management, write those assumptions into the gain plan explicitly, because that is what the condition will be drafted against. The interaction with off-site compensation, and the discount applied when habitat is delivered away from the impact, is covered in our note on spatial risk multipliers in the biodiversity metric.

Biodiversity net gain requires a competent person under the Environment Act 2021, and a Biodiversity Net Gain Assessment drafted with AI assistance is a starting point for that professional rather than a substitute for their work. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval.