Integrating Rainwater Harvesting into SuDS Design

Rainwater harvesting in SuDS design sits in an odd position. Collection for non-potable use is Priority 1 in the national discharge hierarchy, ahead of infiltration, a watercourse and any surface water or combined sewer. Yet on most planning applications the harvested volume ends up counting for nothing against the storage the drainage strategy has to provide. Both are true. A tank sized to supply water is not a tank that is empty when the storm arrives, and that single fact governs almost everything else about how the technique is treated on a planning application.

That gap between policy preference and drainage credit is where schemes get caught, usually late, usually at the point where a Lead Local Flood Authority reads a strategy that has quietly assumed the harvesting tank will absorb a storm it was never sized to absorb. This article sets out what the 2025 National Standards for Sustainable Drainage Systems actually award rainwater harvesting, what Lead Local Flood Authorities will and will not accept in the calculations, and what a system has to demonstrate before any of it can be claimed.

Rainwater harvesting has a deemed-compliance route for interception under Standard 2. It earns nothing automatic against the attenuation storage required by Standard 3.

Does rainwater harvesting count as SuDS?

Yes. Rainwater harvesting is a SuDS component in its own right, with its own chapter in the CIRIA SuDS Manual (C753), and Standard 1 of the 2025 National Standards places collection for non-potable use at the top of the destination hierarchy, which is source control in its purest form. What being a SuDS component does not do is reduce the attenuation storage the rest of the scheme still has to find.

The policy weight behind that hierarchy changed in August 2026. Policy F8 of the revised NPPF now requires sustainable drainage to "be designed in accordance with the National Standards for Sustainable Drainage Systems", which turns a set of non-statutory standards into a decision-making test. The five-tier hierarchy and the wider content of the 2025 standards are covered in full elsewhere on this site; what matters here is that a designer who skips Priority 1 now has to justify the decision on evidence, and that higher cost, on its own, has never been an acceptable reason for dropping down a tier.

That is why reuse belongs in the surface water drainage strategy for planning from the first sketch layout rather than being retro-fitted once the plot positions are fixed. A tank that has to serve a real daily demand needs a catchment, a plant room or chamber, a distribution route and a maintenance access position, and none of those are easy to add late. It also has to sit alongside whatever else the surface water design needs, including the infiltration testing that governs Priority 2. Late is expensive.

What do the 2025 National Standards actually require?

Standard 1 does not make rainwater harvesting compulsory. It makes it a question that has to be asked, and answered, in three defined situations. Where none applies, a designer can move down the hierarchy on the evidence. Where one does apply and reuse is dismissed anyway, expect the drainage consultee to say so, because requirement 1.3 puts the burden of proof on the applicant and rules out cost as a reason on its own.

Requirement 1.12 is the one most schemes fall down on. It asks that any system "intended for the management of large rainfall events" be designed to BS EN 16941 and demonstrate performance "against an appropriate rainfall time series". That means a continuous simulation against real rainfall. Not a single design storm applied to a conveniently empty tank. It is also the evidential standard an authority is entitled to ask for before it credits anything.

One correction worth making, because it is still repeated across most of the trade literature: BS 8515 has been withdrawn. It was superseded in 2018 and the current document is BS EN 16941-1:2024, on-site non-potable water systems for the use of rainwater. Its scope expressly excludes attenuation and infiltration. Read that again, because it settles most of the argument: the standard the National Standards point to for sizing rainwater harvesting is not a stormwater standard at all.

How much water will the roof actually yield?

Good rainwater harvesting design starts here, with the supply, rather than with a tank size taken from a product range. Yield is the annual volume the catchment will deliver into the tank, and it is a straightforward calculation: contributing area multiplied by annual rainfall, multiplied by a yield coefficient for the surface, multiplied by the filter efficiency. Get it wrong and everything downstream is wrong with it: tank size, mains water savings, and any drainage claim resting on either.

The green roof line is the one that catches people out. Green roofs and rainwater harvesting are often specified together because both read well in a planning statement, and the yield is then taken straight from the plan area as though the roof were a conventional tile pitch, which quietly overstates the supply by something close to a factor of two before the design has even reached the tank. A substrate that holds rainfall back for evapotranspiration is doing its job precisely by keeping water out of the tank. Our free rainwater harvesting calculator and the rest of the drainage calculator suite will give you a first pass at the arithmetic before anyone commissions a design.

A 100 m² tiled pitched roof in a 700 mm-a-year rainfall area yields roughly 50 m³ annually. The same area of green roof may yield half that.

How big does the tank need to be?

Storage is sized on the lesser of annual yield and annual demand rather than on either figure alone, and the recognised rule of thumb then takes five per cent of whichever governs, which works out at roughly eighteen days of supply. That is deliberately conservative in the opposite direction from what a drainage engineer would want, because water sitting unused in an oversized tank goes stale.

1. Establish the contributing catchment: the roof and hard surfaces that can be routed to the tank without contaminating the supply.
2. Calculate annual yield from area, local annual rainfall, the yield coefficient and the filter efficiency.
3. Calculate annual non-potable demand from the actual end uses: WC flushing, irrigation, vehicle or plant washdown, process water.
4. Take the lower of the two figures and size the storage at around five per cent of it.
5. Check the result against a rainfall time series if any drainage benefit is to be claimed, as requirement 1.12 asks.

Notice what step four does. A tank sized this way is meant to fill and empty repeatedly through the year, and the Environment Agency's guidance is explicit that it should overflow at least twice a year to clear floating debris. That is not a flaw in the method. It is the method working as intended, and it is exactly why the volume cannot simply be handed over to the drainage calculations as though it were free storage.

A correctly sized harvesting tank is designed to be full for much of the winter. That is precisely when the design storm turns up.

Where the drainage credit is real, and where it is not

There are three distinct claims a scheme might make for a harvesting system, and they have three different answers. Separate them. An authority that sees all three run together as one undifferentiated benefit will usually refuse the lot, whereas the same authority will often accept two of the three when each is evidenced on its own terms.

• Interception: yes, and it is deemed. Standard 2.3 states that where rainwater harvesting means no runoff leaves the development for lower-order design storms, the system is deemed compliant with Standard 2, on calculations to BS EN 16941. The system must be based on regular daily demand, not seasonal garden use. Interception and the first 5 mm are covered separately.
• Volume reduction: yes, with evidence. C753 treats the difference between developed and greenfield runoff volume as long-term storage, and says that volume should be prevented from leaving the site by rainwater harvesting or infiltration. Water genuinely reused is water that never reaches the outfall.
• Peak attenuation: no, not automatically. Nothing in Standard 3 credits harvested storage against the 1 in 100-year-plus-climate-change requirement. That credit has to be earned separately, and it is earned by design, not by assertion.
• Water quality, amenity and biodiversity: no. susdrain rates rainwater harvesting poor on all three. A buried tank contributes nothing to the other three SuDS pillars, which is why harvesting complements nature-based drainage rather than replacing it.

Why do LLFAs model a rainwater harvesting tank as full?

Because they cannot verify that it will be empty. A reviewing officer has no way of knowing whether the occupier will be at home in February, whether the irrigation system will be running, or whether the pump will still work in year fifteen. So the standard position across England is the cautious one: assume the worst case, and require the attenuation separately.

Devon is not an outlier. The same reasoning about rainwater harvesting appears in guidance from authorities with very different drainage contexts, and the national standards take the same line on water butts. They do not guarantee that storage will be available unless they are designed to, and where a scheme says otherwise, detailed calculations are required.

Read together, these are not four different rules. They are one rule stated four ways: the volume counts to the extent that you can prove it will be there when the design storm arrives, which is a modelling exercise rather than a declared intention, and it is the exercise most submissions skip. It is also one of the more common reasons behind an LLFA objection to a planning application.

What a dual-purpose tank has to do to earn credit

It has to hold two volumes. Not one volume doing two jobs, but two, separated by level and by function. The lower volume is retention: harvested water drawn down by demand. Above it sits an attenuation volume that is kept empty in dry weather, fills during a storm, and empties through a flow control at the permitted discharge rate. The overflow sits above both. Only the upper volume is attenuation, and only the upper volume can be counted as such.

Sizing that upper volume honestly is expensive. HR Wallingford's work on sizing rainwater harvesting tanks for stormwater control found that designing to a yield-to-demand ratio below 0.95 needs "between two and six times more storage than is usually provided when sizing tanks for water supply purposes only". The same report offers the practical way round the biggest source of uncertainty: a communal rainwater harvesting system serving around ten properties "removes the uncertainty associated with occupancy rates", because one empty house no longer breaks the assumption.

That is the honest position to take into a pre-application discussion. A single-dwelling tank will almost never earn a peak-flow credit, because one household on holiday is enough to break the assumption the whole calculation rests on. A communal system on a housing scheme can, provided it is sized against a rainfall time series with a controlled release below the overflow. The modelling that proves it is a known quantity, not a research project.

The compliance problems that decide whether it survives

A harvesting system that passes planning can still fail in service, and the evidence on how often that happens is uncomfortable enough that it belongs in the design conversation rather than in a footnote at handover. Water fittings inspections have found most installed systems non-compliant, with a third of the failures involving cross-connections between the rainwater and drinking water systems, usually improvised repairs made after a pump or a filter failed.

Around 70% of existing rainwater harvesting systems inspected failed to comply with the water fittings regulations, according to evidence submitted to Ofwat in 2023.

• Fluid category 5. Harvested rainwater is the highest contamination category. Any mains top-up must be through a physical air gap, not a valve or a non-return device.
• Notification to the water undertaker. Regulation 5 of the Water Supply (Water Fittings) Regulations 1999 requires prior notice and consent for the qualifying work, with ten working days for a response. It is routinely missed, and the water company may inspect before energising the mains supply.
• Marked pipework and labelled outlets. Non-potable distribution has to be identifiable to whoever works on the building in twenty years' time, and every outlet labelled as not drinking water.
• Pump energy. The Environment Agency's own research found that for most systems monitored, pumping harvested water to end uses produced higher carbon emissions than supplying mains water. Worth stating plainly in a sustainability statement rather than leaving it to be found.
• Abstraction. The Environment Agency's regulatory position statement means no licence is needed for roof-collected water kept isolated from inland waters or groundwater, but harvesting must not affect normal watercourse flow.

What should the drainage strategy actually say?

The submission decides the outcome, not the concept. A strategy that proposes harvesting without the supporting arithmetic invites the consultee to discount it entirely. One that separates the claims and evidences each of them usually gets what it asks for. What drainage information a planning application needs varies by authority, and several now want the detail set out on a SuDS proforma or in the national strategy templates, but the harvesting content itself does not change.

1. State the position in the hierarchy explicitly: what non-potable demand exists, what catchment serves it, and where reuse cannot be taken further.
2. Give the yield calculation with every input shown: area, rainfall, coefficient and filter efficiency, alongside the answer.
3. Give the demand calculation from real end uses and occupancy, and show which of yield or demand governs.
4. Claim interception under Standard 2.3 where daily demand supports it, with the BS EN 16941 calculations behind it.
5. Keep any attenuation volume separate, sized by the hydraulic model, with the flow control and overflow levels shown on the drawing.
6. Set out the maintenance regime and who carries it, so the drainage condition can be discharged without a second round.
7. Keep the planning and building control positions consistent, because the two regimes test different things and a system approved on one can still be queried on the other.

Where a scheme is in a seriously water-stressed area the balance shifts, because the driver stops being drainage. In Greater Cambridge, more than 9,000 homes have been held up over water availability, and the emerging local plan proposes a consumption standard of 80 litres per person per day on schemes of 100 dwellings or more, a figure it accepts will need water reuse and dual-pipe distribution. That is a water resources argument rather than an attenuation one. It still lands in the same drainage drawing.

Frequently asked questions

If you are weighing up whether harvesting can carry part of your drainage case, we can tell you early and in writing. Unda prepares surface water drainage strategies and wider drainage strategies for planning across England and Wales, often alongside the flood risk assessment, and every one is signed off by a senior consultant. Call +44 (0) 1293 214444 or email enquiries@unda.co.uk.