Sizing a Soakaway From Three Timed Fills

One Hole, Three Fills, and the Slowest Answer Wins

The trial pit is dug where the soakaway is going and down to the level its base will sit at, because the horizon that matters is the one the water will actually meet. Topsoil takes water freely and says nothing useful about the clay 1.4 m (4 ft 7 in) beneath it, and a rate measured in the first spit is the commonest way a soakaway ends up half the size it needed to be. A pit long and narrow rather than square gives more wetted side for the same volume of water and behaves more like the finished thing.

BRE Digest 365 fills the pit to a marked effective depth, lets it fall, and times the passage between the levels at 75% and 25% of that depth. The middle half of the drop is used because the ends are the untrustworthy parts: the first minutes are dominated by whatever the excavator smeared onto the sides, the last by the base, which on a real soakaway is assumed to silt up and stop working. What comes out is a soil infiltration rate f in metres per second — the volume that left between the two marks, divided by the pit's wetted area up to half depth, divided by the seconds it took.

The area term is where this method is misapplied more often than anywhere else. For the trial pit the area is taken to half the effective depth including the base; for the finished soakaway it is the side area to half depth with the base written off, on the grounds that the base is where fines end up. Two different definitions, one page apart, and using the pit's figure for the soakaway inflates the answer by whatever fraction of the area the base happened to be. Check the definition against the edition of the Digest you are working to before the arithmetic is worth checking.

Three fills, because one is an anecdote. The first is heavily influenced by how dry the ground was when you started; the third is closer to the wetted condition the soakaway lives in through a wet February. Take the slowest of the three — the longest time, the lowest f — and design on it. On this plot a pit 1.5 m long by 0.5 m wide, filled to 1.0 m, held 0.375 m³ between the two marks against a wetted area of 2.75 m² to half depth, and drained in 9 h 40 min, 11 h and 11 h 50 min. The last of those gives f = 3.2 × 10⁻⁶ m/s, which is 11.5 mm/hr or 0.45 in/hr. That is a sandy clay: not a refusal, but nowhere near free-draining, and it is the number every remaining sum on this page runs on.

1. Dig the pit at the soakaway's own position and to its finished base level, and log what the sides show as you go — BS 5930 practice, even on a domestic plot, because the horizon the water meets is the finding.
2. Treat anything over 1.2 m as an excavation to be entered under HSE HSG185 rather than leaned into: measure from the surface with a tape and a float rather than climbing down to read a mark.
3. Mark the effective depth on a batten fixed to the side, then mark 75% and 25% of it — the marks, not the pit, define the test — and time the fall between them in seconds, with a watch rather than a memory.
4. Repeat twice more, refilling as soon as the previous fill has emptied, so the later tests run on progressively wetted ground.
5. Divide the volume between the two marks by the wetted area to half depth and by the slowest time to get f, and record all three results rather than only the one you used.
6. Leave the pit open overnight before backfilling and look again in the morning — water standing in it that you did not put there is groundwater, and that is a different conversation entirely.

The Storm That Sizes It Is Not the One You Would Guess

Storage is a volume problem, not a flow problem. Water arrives over the duration of a storm and leaves through the sides at f the whole time, and what has to be held is the difference between the two totals. Inflow over a storm of duration D is the impermeable area times the rainfall depth for that duration at the design return period; outflow over the same D is the side area to half depth times f times D. The storage you must provide is the largest value of inflow minus outflow across every duration tested — a search rather than a single sum, because inflow rises fast then flattens while outflow rises in a straight line forever.

Which duration wins depends on the ground. On free-draining sand the outflow term catches up quickly and the critical storm is short and violent. On this plot, at 3.2 × 10⁻⁶ m/s, it does not catch up for the better part of a day, and the critical duration lands somewhere between six and twenty-four hours. Anyone who sized this from a one-hour storm — the instinct, because that is the storm you can picture — would have built roughly two-thirds of what the ground needs. Run the durations the Digest lists and take the worst; do not decide in advance which one it will be.

The rainfall depths belong to the location and to no calculator on this site. BRE Digest 365's own tables are built on the Flood Studies Report rainfall statistics — the M5-60 depth and the ratio r — while current UK practice reads depth-duration-frequency data from the Flood Estimation Handbook, and North American work reads NOAA Atlas 14 or whatever the state manual has adopted in its place. The return period is not yours to choose either: it comes from Approved Document H and from whatever the local authority has adopted alongside it, and a climate-change uplift on top — in England, the Environment Agency's published peak rainfall intensity allowances — is now routine rather than optional. The five depths used below — 12 mm at fifteen minutes, 22 mm at an hour, 40 mm at six hours, 65 mm in a day and 80 mm over two — are placeholders chosen to show which term wins, and they are not data for your postcode.

A Box That Is Mostly Gap

The store holds nothing itself; it holds what fits between the bits of it. Clean single-size aggregate to BS EN 13242 gives about 30% voids, and geocellular cells give about 95% — the datasheets across the common systems, Polypipe's Polystorm, Wavin's AquaCell and ACO's StormBrixx among them, all sit around that figure, and the exact number belongs to the product actually bought rather than to a rule of thumb. That single ratio is the entire difference between the two ways of building a soakaway: 1.5 m³ of storage takes 1.6 m³ of crates or 5.0 m³ of stone, and the hole follows the material.

Storage volume itself is the least interesting multiplication on the page — plan area by depth by void ratio — and it is worth doing on a calculator only because it is the one place a unit slip goes unnoticed. What the page below cannot do is model a crate stack: it is a permeable-pavement page, its void-ratio field is bounded to the 0.25–0.45 range open-graded aggregate actually occupies, and 0.95 is outside it. So is the drawdown page in the section after next, which stops at 0.45 for the same reason. Use them for the stone version. The storage page in the section above is the BRE sum itself and accepts either fill, crates or stone, with the manufacturer's void figure in its void field.

Half Empty in a Day, or the Shape Is Wrong

The second check is the one that fails good-looking designs. A soakaway that holds the design storm but is still half full a week later is not available for the next storm, so BRE Digest 365 asks that it discharges half its stored volume within 24 hours. That is a different sum from the storage one and it can fail while the storage sum passes comfortably — which is exactly what happened here on the first attempt.

The first store drawn was a compact box: 2.0 m by 1.2 m of crates, 1.2 m deep, 2.74 m³ of storage against 1.93 m³ required. Ample. Its half-drain time came out at 30.9 hours, and it was rejected. Re-laid as a 4.0 m by 0.5 m trench — one crate module wide, twelve modules in total, three courses of four — it takes slightly fewer cells, 2.4 m³ of them against 2.88, holds 2.28 m³ of water, still comfortably over the 1.50 m³ that geometry needs, and half-drains in 18.3 hours. It got smaller and it drained faster, because what changed was the perimeter.

There is a clean reason for that, and it is worth deriving once because it kills a lot of wasted redrawing. Half the stored volume is plan area times depth times void ratio, divided by two. The side area to half depth is perimeter times depth, divided by two. Divide one by the other and the depth cancels completely: half-drain time is plan area over perimeter, times void ratio, divided by f. Depth does not appear. Making a failing soakaway deeper does not help it drain, and making it shallower does not either — only making it narrower does. Rearranged, the shape passes when plan area divided by perimeter stays below f × 86,400 ÷ void ratio. On this plot that limit is 0.29 m for crates and 0.92 m for stone: the compact box sat at 0.375 and failed, the trench at 0.222 and passed, and the same box built in stone would have passed on the void ratio alone.

The BRE side-area sum itself is the storage page two sections up, which gives the half-emptying time and the largest plan-area-to-perimeter ratio that meets the day beside the storage. The calculator below is a one-dimensional vertical model — stored water depth over a design rate, with a safety factor on the measured figure and a permitted window to check against — which is how North American stormwater manuals frame the same question and how a permeable pavement reservoir is genuinely assessed. It will not reproduce the numbers above, and like the permeable-pavement page it stops at a void ratio of 0.45, so the stone version is the one you can put through it. Use it for what it is good at: seeing how hard the safety factor pushes the answer around, before committing to a shape with no margin in it.

Stone Instead: A Longer Trench and a Much Bigger Hole

Stone wins on drawdown and loses on excavation, and on a plot with room at the bottom of the garden that is often the right trade. The same 46 m² of roof, built as a stone soakaway rather than a crate one, needs 6.48 m³ of clean aggregate to provide 1.94 m³ of storage — a trench 6.0 m long by 1.2 m wide at 900 mm (3 ft) of stone, or 9.0 m by 800 mm at the same depth if the garden is a strip rather than a square. It half-drains in 13 hours and 9.6 hours respectively, against 18.3 for the crates, because 30% voids move less water per unit of side area than 95% do.

Long, narrow stone soakaways usually carry a perforated distribution pipe along their length so the inflow is spread rather than dumped at one end, and that pipe displaces stone you would otherwise have ordered. Over 9 m of 150 mm perforated pipe — 160 mm across the outside, which is the dimension that displaces stone — it is about 0.18 m³: small, but it is a real deduction and it is on the same order as the delivery shortfall people usually blame on the supplier.

Ordering Aggregate Whose Whole Value Is What Is Missing

Two things go wrong when a soakaway's stone is ordered off a general aggregate calculation, and both come from the same place: a soakaway wants an open-graded, single-size, washed stone with the fines taken out, and most gravel sums are written for a base course that wants exactly the opposite.

The first is the compaction allowance. Every driveway and base-course calculation adds 15–25% because loose stone finishes compacted, and every one of those percentage points is a void you are paying to destroy. A soakaway's stone is placed, not rolled. Set the allowance to zero and order the volume you actually want to end up with.

The second is bulk density. The usual default of 1,600 kg/m³ describes a well-graded crushed stone whose small particles have filled the spaces between the large ones. A clean 20/40 single-size — the material a soakaway is built from precisely because it does not do that — comes in lighter, commonly nearer 1,350–1,500 kg/m³, and the honest figure is the loose bulk density on the supplier's own test data for that grade, determined under BS EN 1097-3 or ASTM C29/C29M. On the 6.48 m³ trench above, assuming 1,600 rather than being told 1,400 is 10.4 tonnes against 9.1 — more than a tonne you either paid for or did not get, depending which way round the mistake ran. The page below has 1,600 built into its tonnage and no field to change it, so it will hand you the 10.4: take the loose-volume line out of its breakdown instead and multiply by the density on the supplier's ticket.

The Hole, the Muck Away and the Run From the Downpipe

The excavation is bigger than the store in all three directions. The crate trench above is 4.0 m by 0.5 m of cells, but the dig is nearer 4.6 m by 1.1 m to give working space for wrapping the geotextile and for someone to stand while the cells are placed, and it goes 1.5 m deep to take 1.2 m of cells plus blinding and cover — around 7.6 m³ out of the ground. In a sandy clay that arises at roughly 1.25 times its in-situ volume once it is loose, so 9.5 m³ of spoil is looking for somewhere to go on a garden the extension has already occupied.

Almost none of it goes back. Cohesive spoil packed around a soakaway is the material the geotextile exists to keep out, and cover to a cell stack is normally a specified granular fill placed in shallow lifts, which means importing one load and carting another away — two vehicle movements down a domestic street, and the line most often missing from the quote.

The carrier trench from the downpipes is the other excavation and the more predictable one: 14 m of 110 mm pipe at 600 mm wide by 800 mm deep is 6.7 m³ out, about 7.6 m³ of loose granular back in once the pipe's own volume is deducted and a compaction factor applied. The calculator below reports the backfill as its headline; for the soakaway pit itself, take the total excavation figure from its breakdown and ignore the backfill line, since the pit is filled with cells and stone rather than with fill.

What Happens on the Day It Fills

Every number so far has been a volume. Two things on the job are flows, and neither comes out of a storage sum: the pipe carrying the roof to the soakaway, and the route the water takes on the day a storm larger than the design one arrives — because one will, and the only question is whether it runs across a lawn or into an air brick. For both, the Rational Method is the right tool and storage is the wrong one. Peak discharge from a small catchment is runoff coefficient times intensity times area, and it sizes a pipe, a gully or a channel. It will not size the soakaway, and the calculator below says so itself: it returns a peak with no hydrograph and therefore no volume behind it. Keep the two apart in the submission as well as in your head, because a peak flow quoted as though it justified a storage volume gets an otherwise sound calculation sent back.

The exceedance route is a drawing, not a sum, and it is the cheapest insurance on the job: an overflow at the top of the store, a level threshold it discharges to, and a fall away from the building rather than towards it. Approved Document H's minimum separation between a soakaway and a building exists in part for this — the widely applied figure is 5 m, and it is a floor rather than a target. On a shrinkable clay, or where the extension's foundations are shallow, or where the ground falls towards the house, further is the answer and the geotechnical advice governs rather than the round number.