Permeable Paving and Storage Below It

Reading the Section as a Tank, Not a Pavement

Every permeable pavement holds its water in one place: the space left when angular stone locks against angular stone. The concrete units store nothing. The geotextile stores nothing. Subgrade soil, on most sites, gives up a small fraction of what the stone above it holds. Strip away the vocabulary and a permeable section is a stack of open-graded layers, each rated by the share of its own volume that is air, and each capable of losing that share permanently to fines.

Gradation gets coarser as you go down, and so does the storage. A joint pattern that reads as solid paving from ten feet away is passing water through roughly five to twelve percent of its plan area, depending on unit shape and joint width. Below the units, clean single-size stone typically runs somewhere in the thirties as a fraction of total volume. That percentage, multiplied by depth, multiplied by area, is the whole storage budget — there is no other reservoir in the section.

Whatever the stone actually holds, the number you are allowed to claim is set elsewhere. Some jurisdictions credit forty percent void for a coarse open-graded subbase; others cap the design value at thirty or thirty-two percent regardless of gradation; a few want a measured value from a submitted sample before they will approve the depth. Settle that question with the reviewing authority before anyone digs, because the gap between a thirty and a forty percent credit is roughly a third of your excavation depth and every truckload of stone that goes with it.

Contamination is the other half of the rating. A No. 57 base at thirty-eight percent void becomes a No. 57 base at eighteen percent void the moment silty runoff from an unstabilised slope sheets across it, and no amount of vacuuming recovers stone that has been infiltrated to depth. Voids are the product; protecting them is the trade.

The Open Area at the Top

Openings between units are the throat of the whole system, and they are smaller than most people assume. A rectangular paver with a nominal joint, a shaped unit with spacer bars, and a grid unit with large drainage cells all present very different open areas, and the manufacturer's published figure is what the hydraulic submittal will hang on. Whatever the figure, the pavement is only as permeable as those apertures on the day they stop passing water.

New work tests absurdly fast. Pour water on freshly filled joints and it disappears; measured surface infiltration on a new permeable unit pavement is typically orders of magnitude above any design storm intensity, which is precisely why design values carry a heavy factor of safety and why an installer's site demonstration proves nothing about year five. ASTM C1781/C1781M covers the surface infiltration test for permeable unit pavement systems, and ASTM C1701/C1701M covers the equivalent test for in-place pervious concrete. Establish a baseline with the appropriate one at handover, and record it, because every later maintenance argument turns on that first number.

Joint fill is a wearing part. It works down under traffic, it gets swept out by aggressive brushing, and it gets replaced with the wrong stone by whoever shows up two winters later with a bag of masonry sand. Sand in the joints of a permeable pavement is a failure, not a repair — it fills the aperture with a graded fine that will not vacuum out and turns a reservoir pavement into a slow, weeping conventional one.

Edge restraints, curb transitions and the joint against conventional asphalt deserve the same attention. Water that enters at the edge and travels laterally along a smeared interface finds the subgrade in one place instead of across the footprint, and that concentrated point is the one that pumps, rutts and shows up as a depression.

Two Inches of Bedding, Screeded and Left Alone

Bedding stone carries a disproportionate share of the risk for a layer nobody talks about. A nominal 50 mm (2 in) course of No. 8, screeded to a consistent thickness and never compacted before the units go down, is the reference detail in ASCE/T&DI 68-18, Permeable Interlocking Concrete Pavement. Thickness tolerance here is tight because bedding depth translates directly into surface elevation after compaction, and a thick spot compacts more than a thin one.

Screed rails laid on the base and pulled as you go is still the method. Wheelbarrow ruts and boot prints in the screeded bedding are refilled by hand with loose stone, not smeared flat with a trowel — a troweled patch densifies locally and shows as a low unit once the plate compactor comes through. Nobody walks the screeded surface. Everything works forward off the laid pavers.

Units are set, joints filled with the specified small stone, then compacted, then the joints topped up and recompacted until they stop taking material. Leave the surface a few millimetres proud of final grade before compaction; the section will come down as the bedding densifies under the units. Getting that allowance wrong at a curb line or a drainage structure means lifting and relaying, and there is no way to correct it from above.

Aggregate for these layers is bought by gradation designation, and ASTM D448, Standard Classification for Sizes of Aggregate for Road and Bridge Construction, is what the numbers refer to. Specify by designation, require a gradation on delivery, and reject anything with a fines tail. A quarry that supplies mostly conventional base work will ship you material that is technically No. 57 and functionally half-choked with crusher dust unless the order says washed.

The Base Course as Transition and Working Platform

Between the small bedding stone and the large subbase stone sits a course of No. 57, and its job is mechanical as much as hydraulic. It stops the fine bedding stone migrating down into the coarse voids below, and it gives the compaction plant something to work on without crushing the subbase into its own fines. Skip it, and No. 8 finds its way into a No. 2 subbase — the storage is still nominally there, but you have paid for a reservoir and delivered a filter.

Compaction of the base is a different exercise from compacting a dense-graded aggregate. Open-graded stone does not have a moisture-density curve to hit; it is compacted in lifts until the stone locks and the plate stops walking. Over-rolling breaks particle edges and manufactures the fines you spent money to wash out. Under-rolling shows up as settlement at the wheel paths a season later.

Ordering is where this section leaks money on most jobs. Depth times area times a loss allowance, converted to tonnes at the right density, per layer, per gradation — and the numbers differ from a conventional paver build because open-graded stone is lighter in place and the bedding is stone rather than sand.

Take the deliveries seriously. Stone that arrives at the wrong gradation and gets placed anyway is not a paperwork problem; it is an excavate-and-replace problem, because the voids you signed for are the deliverable and there is no way to inspect them once the units are down.

The Subbase: Where the Design Storm Actually Lives

Most of the volume in any permeable section sits in the coarse subbase — No. 2 or No. 3 stone, placed in lifts, holding somewhere near a third to two fifths of its own volume as air. Depth of this layer is not chosen by structure; it is chosen by whichever governs of the storage volume the jurisdiction requires and the depth at which the section still drains down in the allowed time. Structural capacity is usually satisfied long before the hydraulic depth is.

Two constraints work against each other. Deeper stone stores more, but a deeper excavation reaches closer to seasonal high groundwater and to bedrock, and most manuals impose a minimum separation between the bottom of the reservoir and both. When separation is the binding constraint, the answer is a wider, shallower footprint, or a section that stores less and drains through an underdrain rather than into the ground.

Run the storage arithmetic before the layout is fixed rather than after, because the result changes the footprint, not just the depth.

Contributing area matters as much as the pavement area. A permeable bay taking run-on from adjacent roof or asphalt is being asked to store several times its own rainfall, and the run-on ratio is capped in most manuals precisely because the sediment load scales with it. Sizing the reservoir for the pavement alone and then quietly draining the parking aisle onto it is the most common way these systems end up undersized and clogged at once.

The Bottom Interface and the Drawdown Clock

A reservoir that does not empty is a pond with stone in it. Design drawdown time — the period within which the stored volume must infiltrate — is set by the local manual and commonly falls between one and three days, chosen to protect against mosquito breeding, to have the storage available for the next storm, and to keep the subgrade from staying saturated. Which end of that range applies to you is a jurisdictional question, not an engineering preference.

Field infiltration rate of the subgrade is the input everything hangs on, and it should be measured at the elevation of the finished excavation, not at the surface. ASTM D3385, Standard Test Method for Infiltration Rate of Soils in Field Using Double-Ring Infiltrometer, is one accepted method; some jurisdictions specify a different test or require a soil classification per ASTM D2487 alongside it. Design rates are almost always the measured rate divided by a safety factor, and that factor is also written into the local manual.

Protecting the subgrade is a construction problem with no fix. Track the bottom of an excavation with a loaded truck in wet weather and the surface smears; the infiltration rate you tested no longer exists, and no amount of scarifying fully restores it. Excavate with the machine working from outside the footprint where geometry allows, keep rubber-tyred plant off the bottom, and scarify lightly if the surface has been trafficked before stone goes in.

Compaction is the other trap. Conventional pavement practice compacts the subgrade; full-infiltration permeable practice does not, beyond what is needed for stability, and a proof-roll requirement written by someone who has only built dense-graded sections will destroy the design. If the geotechnical report calls for a compacted subgrade under a pavement intended to infiltrate, that contradiction gets resolved on paper before anyone starts, usually by converting the section to partial infiltration with an underdrain.

Voids That Must Stay Empty

Not every void in the section is storage. An underdrain sitting near the bottom of the reservoir converts the layer below it into dead storage and everything above it into detention; raising the underdrain invert creates a deliberate sump that infiltrates while the pipe handles the excess. Where the pipe sits vertically is a design decision with a large volumetric consequence, and it is worth confirming on the shop drawing rather than assuming.

Overflow has to exist. Storms larger than the design event will fill the stone and surcharge to the surface, and the section needs a defined route out — an outlet structure, a raised inlet grate, or a controlled surface path to a downstream conveyance — that does not run across a building slab or a neighbouring property. Systems built with no overflow simply flood at the low unit, and the low unit is usually the one closest to the door.

Sloping sites subdivide. Above roughly a few percent longitudinal grade, water in a continuous stone reservoir migrates downhill and stacks up at the low end, filling the last cell and leaving the high end dry. Impermeable check dams or berms across the subgrade break the reservoir into stepped cells, each holding its own volume. Miss them and the storage you paid for is concentrated where you least want it.

Observation wells earn their keep. A perforated riser set into the subbase, capped at the surface, is how anyone later confirms whether the reservoir is draining or has been holding water since March. Adding one costs almost nothing during construction and is effectively impossible afterwards.

A Sequence Built Around Not Losing Voids

Sediment is the enemy, and construction is the dirtiest phase the pavement will ever see. Permeable areas should be among the last things built, kept out of haul routes, and never used to dewater an active site. Where that is unavoidable, the reservoir is built with a sacrificial layer or a temporary barrier over the subgrade, and construction runoff is routed to a settling facility rather than allowed to sheet across the stone.

Upstream catchment must be stabilised and vegetated before the pavement is opened to it. A single storm across bare soil delivers more fines than years of ordinary service, and it deposits them below the surface where sweeping cannot reach.

Silt fence and inlet protection stay in place until the last disturbed area is stabilised, and the discharge permit conditions governing that are typically the construction general permit administered by the state or the EPA, depending on jurisdiction.

1. Stabilise and vegetate the contributing catchment before any permeable area is exposed to run-on.
2. Excavate to subgrade working from outside the footprint; keep loaded plant off the bottom.
3. Verify subgrade elevation and infiltration assumptions, and install the observation well and any underdrain before stone.
4. Place the subbase in lifts, then the base course, compacting until the stone locks rather than to a density target.
5. Screed the bedding course to a uniform nominal thickness; do not compact it and do not walk it.
6. Lay units off the laid surface, fill joints, compact, top up the joints and recompact until they stop taking stone.
7. Test surface infiltration to the applicable ASTM method at handover and record the result as the maintenance baseline.

Voids Lost Over Time, and How the Loss Is Measured

Clogging is progressive and it starts at the surface. Fines arrive on tyres, from adjacent landscape beds, off roofs, and in winter grit, and they lodge in the top inch of joint fill. That concentration is good news operationally, because a surface-bound blockage is recoverable by regenerative-air or vacuum sweeping that lifts the contaminated stone; deeper contamination is not recoverable at all.

Sweeping schedules that are written by area rather than by exposure fail. A loading dock apron, a bay under a tree line and an open plaza accumulate at completely different rates, and the honest way to set frequency is by retesting surface infiltration against the handover baseline. When the measured rate approaches the design value plus its safety factor, service the surface; if servicing does not recover it, the fines have moved past the joint and remediation means removing and replacing joint and bedding material.

Winter maintenance requires its own instruction to the operator. Sand is prohibited on permeable surfaces because it is exactly the material clogging is made of, and deicer application rates are usually reduced because the pavement drains rather than holding brine on the surface. A plough blade set to skim rather than scrape protects the joint fill and the unit chamfers.

Handover documentation is the last defence. The owner needs the surface infiltration baseline, the sweeping method and prohibition on sand, the observation well location, the underdrain and overflow arrangement, and the name of the standard the pavement was built to — ASCE/T&DI 68-18 for interlocking units, ACI 522.1, Specification for Construction of Pervious Concrete Pavement, where the surface is pervious concrete. Systems fail far more often from being maintained as ordinary paving than from being designed wrong.