Bioretention and rain garden field guide for stormwater crews

Bioretention, and why it ends up on the plan

Bioretention is a shallow landscaped depression that catches stormwater runoff, holds it on the surface for a few hours, and filters it down through an engineered soil mix before it soaks into the ground or drains away clean. The plants and mulch and soil biology pull out the pollutants. The depression stores the water. The whole thing is sized to handle the runoff from the roofs and pavement around it.

It goes by a pile of names. Rain garden, bioretention cell, biofilter, bioinfiltration basin, LID cell. They are not all identical, and the difference matters once money and a permit are involved, but they share one job: take dirty, fast runoff off hard surfaces and turn it into slow, clean water the ground can take.

The reason you are building one is usually not landscaping. It is the permit. A site that adds or redevelops impervious area generally has to manage the stormwater that pavement creates, and in a lot of jurisdictions that mandate flows down through the MS4 or NPDES stormwater permit to a required water-quality volume the site has to capture and treat. Bioretention is one of the green-infrastructure BMPs that gets you there. The erosion-control and SWPPP side of stormwater is its own subject; this guide is the permanent cell, not the temporary construction controls.

What's the difference between a rain garden, a bioretention cell, and a bioswale?

All three are planted, soil-based stormwater controls, and the line between them is about engineering and intent, not botany. The residential rain garden is the simplest: a shallow planted dish in a yard, often dug into amended native soil, sized by eye or a rough rule, taking runoff from a roof or a driveway. No underdrain, no engineered media spec, no permit on most lots.

The engineered bioretention cell is the designed version. It has a specified soil media, a defined depth, a gravel storage layer, usually an underdrain, an overflow structure, and a sizing calculation tied to a water-quality volume. This is what shows up on a commercial site plan, what the reviewer checks against the BMP manual, and what carries a maintenance obligation for the life of the site.

A bioswale is the linear cousin. Instead of a dish it is a long, gently sloped channel that conveys runoff while it filters and infiltrates along the way, often running beside a road or a parking lot. Same media idea, different shape, and it does double duty moving water from one point to another. If you need to carry water across a site and treat it, that is a swale. If you need to park and soak a volume in one spot, that is a cell. The grading and conveyance side of that work lives in the drainage and slope guide.

The anatomy of a bioretention cell

A working cell is a stack of layers, each doing one job, built from the bottom up. Read it top to bottom the way water moves through it: it ponds on the surface, soaks through mulch and media, hits a transition layer, collects in stone, and either infiltrates into the native soil or leaves through an underdrain. Big storms that exceed the ponding depth take the overflow instead of drowning the cell.

Get the order and the materials right and the cell breathes between storms. Let the layers contaminate each other, fines migrating down into the stone or native soil smeared up into the media, and you have built a bathtub that holds water instead of a filter that passes it.

What soil goes in a bioretention cell?

The media is sand-dominant, not topsoil and not native dirt. A common default is a sand-dominant mix, with many manuals putting the sand in the 80 to 90 percent range and the compost down around 3 to 5 percent by volume, with only a small fraction of compost and fines. The point of all that sand is speed. The mix has to pass water fast enough to drain the cell in a day or two, and native clay cannot do that.

The numbers a media spec controls are the gradation and the rate. Fines passing the No. 200 sieve are usually held low, commonly in the 2 to 5 percent range, because fines are what clog the mix. Organic matter sits around 5 to 8 percent, pH commonly 5.5 to 7.0, and the field-tested infiltration rate typically lands in the 1 to 8 in per hour band, with various manuals targeting a tighter window inside that. The exact spec is local. The state or municipal BMP manual and the project civil and landscape engineer set the mix, and ASTM F1815 is the test method commonly named for the media's gradation and hydraulic conductivity.

Here is where cells die before they are planted. Somebody substitutes cheaper screened topsoil, or a planting mix heavy on compost, because it grows plants beautifully and it is already on the yard. It also holds water like a sponge, drains slow, and the cell ponds for days. The media is the heart of the thing. If the mix is wrong, nothing downstream of it, not the underdrain, not the plants, not the maintenance, can save the cell. Buy it tested to the spec and get the gradation report before it goes in the hole.

How fast should a rain garden drain?

A bioretention cell should empty its surface ponding within about 24 to 48 hours after a storm. That window is not arbitrary. Faster than that and the soil never got to do much filtering. Slower than that and you have three problems stacking up: the plants drown with their roots in standing water, the cell becomes mosquito habitat, and it has no empty storage left for the next storm that lands on top of it.

Drawdown is measured from the high-water mark down to an inch or two above the cell bottom, not all the way to bone dry. Some manuals tighten it. 24 hours or less is common where the discharge runs to a temperature-sensitive trout stream, because warm water sitting in the cell hurts cold-water fish. The wet-and-dry cycling between storms is the goal. It keeps the soil aerobic and the plants alive.

Standing water past 48 hours is the tell that something is wrong, and it is almost always the media clogging at the surface or the wrong mix going in to begin with. A cell that ponds for days is not a cell. It is a pond, and an illegal one if the permit said it would infiltrate.

How big does a bioretention cell need to be?

The cell surface area is sized as a fraction of the impervious area draining to it, and as a rough starting point that fraction commonly lands somewhere around 4 to 10 percent of the contributing impervious area. The real driver is the water-quality volume, the runoff the permit says the site has to capture and treat, often expressed as the first inch to inch and a half of runoff off the drainage area. The cell has to store and infiltrate that volume.

Designers also watch the loading ratio, the ratio of drainage area to cell area. Push too much pavement at too small a cell and it clogs fast and overflows constantly. Common practice holds the ratio at or below roughly 15 to 1, though field numbers run anywhere from about 5 to 1 up past 20 to 1 depending on the soil and the manual. The storage target is often at least 75 percent of the water-quality volume held across the ponding, the media voids, and the stone.

None of these are numbers to carry in your head as gospel. The sizing comes off the local stormwater BMP manual, the permit's water-quality volume, and the engineer's calculation for the specific drainage area. Build the size on the approved plan, and if field conditions do not match the plan, that is a conversation with the engineer before you dig, not after.

The underdrain, and when you need one

An underdrain is a perforated pipe laid in the gravel storage layer that carries water out of the cell when the native soil underneath cannot take it fast enough. On good sandy ground the cell can infiltrate straight into the earth and needs no pipe. On tight clay it would pond for a week, so the underdrain gives the water a way out while the media still does the filtering.

The rule of thumb a lot of manuals use: when the native infiltration rate falls below roughly half an inch per hour, plan on an underdrain, and when it is down around a tenth of an inch per hour or less, bioretention that relies on infiltration is the wrong tool entirely. Test the in-situ soil before you commit, because the difference between an infiltrating cell and an underdrained one is a different design, a different cost, and a different outlet.

The pipe itself is commonly 4 in perforated PVC or HDPE, laid in washed stone, sloped to drain, with cleanouts so it can be rodded out when it silts up. Put a cleanout at each end and at intervals along a long run. The outlet daylights to a swale or ties into the storm system, and it has to sit at an elevation the receiving system can actually accept. An underdrain you cannot clean is an underdrain that fails quietly, so the cleanouts are not optional.

Pretreatment, the inlet, and stopping the scour

Pretreatment is the single highest-payback feature on the whole cell, and it is the one most likely to get value-engineered out. The idea is to drop the coarse sediment out of the runoff before it ever reaches the media, because sediment on the media surface is the number one way bioretention cells clog and die. Leave it out and the cell silts in and stops draining within a few seasons.

The common forms are a forebay, a small sedimentation pool, often 18 to 30 in deep, where concentrated flow slows and drops its load, or a grass filter strip, frequently called out around 20 ft of dense vegetation, where sheet flow filters before it enters. Use a forebay where the water comes in concentrated through a pipe or a curb cut. Use a filter strip where it sheets in off pavement. The forebay gets cleaned out when it has lost about half its volume, and that is a maintenance line item somebody has to own.

At the inlet itself, fast water scours. A curb cut or a pipe dumping onto bare media digs a hole and undercuts the planting, so you armor the entry with washed rock or a splash pad to break up the energy and spread the flow. The inlet also has to sit low enough to actually let water in. A curb cut set above the ponding line is a cell that never gets fed, which sounds obvious and gets built wrong constantly.

Ponding depth and the overflow

The surface ponding depth, the dish that holds water while it soaks in, is commonly built at 6 to 12 in, with some manuals allowing up to around 18 in. Deeper stores more but takes longer to draw down and stresses the plants. Shallower draws down fast but holds less, so it has to be wider to hit the same volume. The depth on the approved plan balances that against the drawdown requirement.

Every cell needs a defined overflow, because the design storm is not the biggest storm that will ever land on it. When the ponding fills, water has to leave somewhere safe instead of overtopping the low edge and cutting a channel out of the cell. The overflow is a structure: a raised inlet or standpipe, a weir, or an armored spillway set at the maximum ponding elevation and routed to the storm system or a stable outlet.

Set the overflow elevation wrong and you defeat the cell either way. Too low and it short-circuits, dumping water out before the cell treats it. Too high and the big storm overtops the berm and erodes the whole thing. The overflow invert is one of the as-built elevations worth shooting and recording, because it is what protects the cell on the day it matters.

Plants and mulch for the cell

Bioretention plants live a hard life: flooded for a day or two after a storm, then bone dry through a drought, sometimes in the same week. So you plant species that tolerate both, and you zone them by where they sit in the cell. The wet bottom takes the water-tolerant species that can stand periodic inundation. The dry upper edges and side slopes take the drought-tolerant species that would rot in the bottom. Native species are the usual choice because they handle the local climate and feed local pollinators without babysitting.

Plant densely. A thickly planted cell shades out weeds, holds the media against erosion, and keeps roots working through the whole soil profile, which is part of what keeps the cell infiltrating. Sparse planting leaves bare media that crusts, erodes, and weeds in. Establishment is the fragile window. The first two growing seasons need watering through dry spells and weeding, and a cell that gets walked away from right after planting usually fails its first inspection on plant survival.

Mulch goes on at 2 to 3 in of shredded hardwood, not bark nuggets and not wood chips. Shredded hardwood knits together and stays put. Nuggets and chips float, and the first storm rafts them into the overflow and out of the cell. Keep the mulch back off the plant stems so they do not rot. The mulch filters fines, holds moisture during establishment, and feeds the soil biology that does part of the pollutant removal.

Building it, and the rule that kills cells: don't compact the bottom

The fastest way to ruin a bioretention cell is to compact the bottom of it, and on a normal job site that happens by accident before anyone is paying attention. Run an excavator or a loaded truck across the cell footprint, or dig it in the rain and smear the clay, and you seal the infiltrating surface. Compaction can cut the infiltration rate of sandy soil by an order of magnitude and clay soil by something like a factor of 50. The cell then ponds forever no matter how good the media is.

So the cell gets built last, after the contributing drainage area is stabilized, and the footprint gets fenced off from the day grading starts. No equipment, no stockpiles, no traffic on it. Excavate from the side, with the machine sitting outside the footprint where you can. When the hole is open, the bottom gets scarified, raked open with the bucket teeth to break up any smear, never rolled or tamped. A smooth bucket blade glazes the soil and restricts infiltration, so you want it torn up, not slicked.

Building it last also keeps it from silting in. A cell dug early becomes the low point on a bare site, every rain washes the disturbed soil into it, and it is clogged before it is ever planted. If sequencing forces the excavation early, it gets protected, and any sediment-laden media is stripped and replaced before the cell is brought online. The temporary erosion controls that keep that sediment out are the SWPPP's job, and they have to stay up until the drainage area is stable.

Erosion while the cell establishes

A freshly built cell is bare media and small plants, which is exactly the condition that erodes. Until the plants knit in and the mulch settles, concentrated inflow can cut a channel down the cell, rill the side slopes, and carry the media itself into the overflow. The cell is most vulnerable in the same window it is least able to protect itself.

The protections are the ones the erosion-control work uses everywhere else. Keep the contributing area stabilized so clean water, not muddy water, reaches the cell. Armor the inlet so the entering flow cannot scour. Run erosion-control blanket or a temporary cover on the side slopes through establishment, and keep pretreatment functioning so sediment drops out before it reaches the planted surface. The full set of construction stormwater controls, the silt fence, the inlet protection, the stabilization deadlines, lives in the SWPPP and erosion-control side of the work, and a bioretention cell sits downstream of all of it.

Testing the media and proving it meets spec

The media is the part of the cell most worth verifying, and the part most likely to arrive wrong. Get a gradation and infiltration report on the actual mix before it goes in, not a generic data sheet off the supplier's website. The gradation tells you the sand-to-fines ratio matches the spec. The hydraulic conductivity or infiltration test, commonly run per ASTM F1815 for the media, tells you it will actually pass water at the design rate.

Separate from the media, test the native soil at the bottom of the cell. The in-situ infiltration rate decides whether you need an underdrain and how the cell is allowed to be sized, so it is not a formality. A percolation or infiltration test in the actual subgrade, at the actual elevation, is what the design should be built on, and a field-tested infiltration check after construction confirms the finished cell draws down in the required window.

Document the results against the spec, because the reviewer and the long-term owner both need to see that what got built matches what was approved. A cell that fails its drawdown test on day one fails because of a number somebody skipped checking: the media gradation, the in-situ rate, or the compaction of the bottom. Those are the three to verify before you call it done.

Why is my rain garden not draining?

A cell that ponds longer than a day or two has one of a short list of problems, and they rank by how often they show up. The surface clogged with sediment is first. Fines from an unstabilized drainage area or missing pretreatment crust over the media and seal it. You can often see it, a fine silt skin on the surface, and the fix is to strip the top inch or two of clogged media and mulch, replace it, then fix whatever is feeding sediment in.

Wrong media is second, and it is the worst because it is buried. A compost-heavy planting mix or screened topsoil substituted for the spec sand mix drains slow no matter what you do at the surface, and the only real fix is to dig it out and replace it. Compacted bottom is third. The cell was driven on or dug wet, the infiltrating surface is sealed, and unless there is a working underdrain the water has nowhere to go. Fourth is an underdrain that is itself clogged or has no functioning outlet, which is the reason the cleanouts exist.

Diagnose in that order before you start digging. Look at the surface, check the media against the spec and the gradation report, ask whether the footprint was protected during construction, and rod the underdrain. A cell that never drained from the day it was built almost always traces to construction, not to maintenance.

The owner-side maintenance

Bioretention is not install-and-forget, and the maintenance obligation usually runs with the property for the life of the site as a permit condition. Whoever owns the site owns the cell, and most failures in the field are not design problems. They are maintenance nobody did.

The recurring tasks are straightforward and they all serve drawdown. Clean the forebay and pretreatment when sediment builds up, commonly when a forebay has lost about half its volume. Refresh the mulch as it breaks down, replacing rather than just piling on so the cell does not slowly fill in. Replace dead plants and weed through establishment. And check the drawdown after a real storm. If the cell still holds water a couple of days later, something is clogging and it gets dealt with before it gets worse.

The inspection a reviewer looks for centers on the same things: is it draining in the required window, is the pretreatment functioning, is the vegetation surviving, is the overflow clear and the inlet armored. Keep a simple record of inspections and the drawdown after storms, because that record is what shows the cell is doing the job the permit credited it for. A clogged cell that nobody logged is a violation waiting for an inspector to find it.

Cold climates and road salt

In cold climates the cell has two extra enemies: frozen ground and road salt. A frozen media surface infiltrates slowly or not at all, so a midwinter melt can pond on a cell that drains fine in summer, and the design has to account for that reduced winter performance instead of assuming the cell works year-round at full rate.

Road salt is the bigger long-term problem on cells that take runoff from plowed pavement. Chloride from deicing salt washes into the cell, and while the media holds some of it temporarily before later rain flushes it through, the salt loading hammers the plants. Loss of plant cover is common in roadside cells for exactly this reason. Where salt is in the picture, plant salt-tolerant species and expect to replace plants more often.

Do not use the cell as a snow dump. Piling plowed snow on it packs the media down, compacts it, and buries it under the sand, grit, and salt the plow scraped off the road, which is a direct route to a clogged, compacted, dead cell. Stockpile snow somewhere it can be hauled off or melt to a different control, and keep the bioretention footprint clear.

Commercial sites, data centers, and LID compliance

On commercial and large industrial sites, bioretention is rarely a nicety. It is how the project meets its stormwater permit. A big-box roof, acres of parking, a data center's vast building footprint and equipment yards, all of it is impervious area generating runoff the permit says has to be captured and treated, and low impact development, LID, is the design approach that keeps that runoff on site instead of piping it straight to the creek.

Data centers are a sharp case because the impervious ratios are high and the sites are big, so the stormwater volume is large and the cells, or the network of cells and swales, get sized accordingly. The civil design integrates bioretention with the rest of the site stormwater system, and the permit's water-quality volume drives how much treatment area the site has to provide. The same rules apply at scale: pretreatment so the cells do not clog, media that infiltrates, drawdown in the required window, and a maintenance plan that survives the building changing hands.

The compliance reality is that these cells get inspected and the permit has teeth. A cell built wrong, compacted, undersized, or fed by failed pretreatment, is not just a dead plant bed. It is the site out of compliance with its stormwater permit, and that stays the owner's problem, on the owner's record, until it is fixed.

What to document

The record on a bioretention cell is what proves it was built to the approved design and what tells the next crew how it works. Cells outlive the people who build them, change hands with the property, and get inspected for years, so the as-built and the test results have to be findable.

The two numbers that matter most are the field infiltration or drawdown result and the as-built overflow elevation. The first proves the cell drains. The second protects it in the storm that exceeds the design. If you record nothing else, record those two, with the media gradation report behind them.

Common mistakes

• Compacting the bottom by running equipment across the footprint or digging it in the rain.
• Building the cell early so the bare site silts it in before it is ever planted.
• Substituting compost-heavy planting mix or screened topsoil for the spec sand media.
• Skipping or undersizing pretreatment, so sediment clogs the media surface within a few seasons.
• Leaving out a defined overflow, so the first big storm overtops the berm and cuts a channel.
• Sizing the cell too small for the drainage area, so it overflows constantly and clogs fast.
• Mulching with bark nuggets or wood chips that float out the overflow instead of shredded hardwood.
• Leaving the underdrain off tight soil, or building one with no cleanout and no working outlet.
• Walking away after planting, so the cell fails its first inspection on plant survival.

Field checklist

Standards and references

The documents that govern a bioretention cell are local, and that is not a hedge to dodge specifics. It is the actual structure of stormwater regulation. The local or state stormwater BMP manual sets the media spec, the sizing method, the drawdown requirement, and the construction details, and it varies a lot between jurisdictions. Build to the manual the reviewer is using, not to a number from another state's manual that looks close enough.

Above the manual sits the permit. The MS4 or NPDES stormwater permit is what requires the treatment in the first place and defines the water-quality volume the site has to manage. The EPA's green-infrastructure and bioretention guidance frames the national approach, but the enforceable numbers come from the state and local program that implements it. The geotechnical or infiltration test report establishes the native soil rate the design depends on, and ASTM F1815 is the test method commonly named for the media's gradation and infiltration.

The project civil and landscape engineer owns the specific design: the cell size, the media mix, the underdrain decision, the plant schedule, the overflow. When the field does not match the plan, that is a question for the engineer of record, not a field call. Nothing in this guide is a substitute for the adopted BMP manual, the permit conditions, and the stamped design for the specific site.

Units, terms, and conversions

Bioretention picks up terms and acronyms from stormwater regulation, soil science, and the landscape trade, and the same cell can read differently across a permit, a civil plan, and a planting schedule.

Infiltration rates show up in inches per hour in most US manuals and millimeters per hour in metric sources. Media depth and ponding depth are in inches; storage volumes in cubic feet, or the water-quality volume as a depth of runoff over the drainage area. Native soil infiltration capacity gets ranked by hydrologic soil group, A through D, where A drains fast and D is tight clay.

BMP

Best management practice, the physical stormwater control measure on the ground

LID

Low impact development, the design approach that keeps runoff on site

MS4

Municipal separate storm sewer system, the permitted public storm network

WQv

Water quality volume, the runoff the cell is sized to capture and treat

Bioretention soil media (BSM)

The engineered sand-dominant filter mix, not topsoil or native soil

Drawdown

Time for the ponded water to soak away after a storm, commonly 24 to 48 hours

Underdrain

Perforated pipe that drains the cell when native soil cannot infiltrate fast enough

Forebay

A small pretreatment pool that drops sediment before it reaches the media

HSG

Hydrologic soil group A to D, ranking native soil infiltration from fast to tight

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• When do you need a SWPPP?
• How much should the ground slope away from a house?
• What is a detention pond?
• What is a French drain?
• What is pervious concrete?