Two numbers, multiplied and divided
The whole page is two operations. Storage is the gross volume of the hole times the void ratio of whatever fills it. Drawdown time is that storage divided by the rate at which the soil accepts water, which is the design infiltration rate times the bottom area.
Storage (cu ft) = pit volume × void ratio
Drawdown (hours) = storage ÷ (design rate in ft/hr × bottom area in sq ft)
An 8 by 4 by 4 foot excavation is 128 cubic feet of hole. Filled with clean open-graded stone at a 35 percent void ratio it stores 44.8 cubic feet, or 335 gallons. That is the first surprise: two thirds of a stone-filled pit is stone. The second surprise is the drawdown. With 32 square feet of bottom and a design rate of a quarter inch per hour, the pit accepts 0.667 cubic feet an hour and takes 67 hours to empty. Nearly three days, from a hole that holds about as much as seven bathtubs.
The infiltration rate is not a number anyone can give you
Published tables of infiltration rate by soil texture exist and they are useful for expecting an order of magnitude, but they are not a design input for a pit on your lot. Soils are layered, the layer that governs is whichever one is slowest, that layer might be six inches thick, and a compacted lens left by construction traffic can cut the rate of good soil by an order of magnitude without changing what it looks like in a hand sample.
What you need is a test in a hole at the depth the pit bottom will sit, pre-soaked so you are measuring saturated behaviour rather than the first thirsty hour of dry ground, and preferably more than one hole. Where a design is being reviewed, the reviewer usually specifies the test method they will accept, and that specification is the one to follow rather than any general description.
Then it gets divided. The safety factor on this page defaults to 2, which is a common working figure and not a rule. It covers the gap between a clean test hole and a pit that has been receiving roof grit for five years, between a test in one season and performance in another, and between one hole and the average of the ground under the whole footprint. Where a pit failing would cause water in a building, a larger factor is cheap insurance.
Bottom area does the work, not depth
| Change | Effect on storage | Effect on drawdown rate |
|---|---|---|
| Make the pit deeper | Rises in proportion | No change — same bottom area |
| Make the footprint bigger | Rises in proportion | Rises in proportion |
| Use a chamber instead of stone | Rises sharply | No change |
| Reduce the contributing area | No change | No change, but less to drain |
This is why deep narrow shafts empty so slowly and why a shallow wide bed of the same volume empties far faster. Depth buys storage cheaply, in the sense that the extra excavation is small compared to widening; but it buys no throughput at all, and a pit that cannot empty between storms behaves like a pit that was never built. If the drawdown figure comes out long, the fix is a wider footprint or a smaller contributing area, not a deeper hole.
Reducing what arrives is often the easiest lever and the one people reach for last. Splitting the roof between two pits halves each one. Sending part of the roof to a rainwater harvesting tank takes it off the pit entirely for the storms that matter most. Replacing a sealed surface with a permeable one changes the runoff coefficient itself, which the permeable paving runoff calculator works through.
What the pit is standing in
A dry well works by having somewhere below it to put water. Two conditions remove that. One is a high water table, seasonal or permanent, which means the pit fills from underneath in the wet months and has no capacity left when a storm arrives. The other is rock or a hardpan close below the bottom, which turns the pit into a bathtub that fills to the same level every time. Both are found by digging a hole and looking, and neither shows up in a percolation test run at the wrong time of year.
Distance from buildings matters for the same reason as distance from a downspout: water put into the ground beside a foundation goes down beside the foundation. A pit that solves a puddle in the lawn can create a wet basement wall thirty feet away, and the mechanism is the same one described on the basement water entry organizer. Separation distances from foundations, property lines, wells and septic systems are set locally and this page states none of them.
Before you dig
Before digging anywhere on the property, have the underground utilities located. In much of North America that is a free call-before-you-dig service and using it is a legal requirement, not a courtesy. Separately: excavating alongside a foundation can undermine it. Soil next to a footing is carrying load, a trench cut beside or below the footing can let that soil move, and the failure does not always show up on the day. This page gives no excavation procedure and none should be inferred from it. Work close to a foundation, below the footing line, or in soil that will not hold a face belongs with a contractor or an engineer who is standing on the site.
Where roof water and site water are allowed to go is set locally and it varies enormously. Discharging onto a neighbouring property, into a street gutter, into a storm sewer, into a sanitary sewer, or into a watercourse may be required, permitted, permitted only with approval, or prohibited outright depending on where you are, and some jurisdictions also set minimum setbacks and grading requirements at the building. None of that is stated on this page as fact, because none of it can be. Your local building department or stormwater authority is the source, and it is worth a phone call before anything is dug, because water arriving on someone else's land is one of the more common ways a drainage project turns into a dispute.
Questions people ask
How big should a dry well be for my downspout?
Work it from the volume rather than from a rule of thumb. Take the roof area draining to that downspout, multiply by the runoff coefficient and by the storm depth in feet, and that is the cubic feet you need to store. Divide by the void ratio of the fill to get the size of the hole. A 900 square foot roof section in a one inch storm produces about 71 cubic feet of runoff, which in stone at 35 percent voids needs a 203 cubic foot excavation. That is a footprint of about 51 square feet at 4 feet deep, roughly 8 by 6.5 feet, which is more than most people picture.
What is a good void ratio for the stone?
It depends on the gradation and the supplier is the source. Clean open-graded stone with the fines washed out is commonly quoted in the low to mid thirties as a percentage, but the value depends on the size distribution and on how much it is compacted going in. What matters more than the exact number is that the stone is open-graded. Dense-graded base, the kind that packs to a hard surface, is designed to have almost no voids and stores essentially nothing.
How long should a dry well take to drain?
Many jurisdictions set a maximum drawdown time as a review criterion, and the figure differs between them, so the honest answer is that yours sets it. The reasoning behind whatever number they use is the same everywhere: a pit still holding the last storm has no capacity for the next one, and standing water in a pit for long periods raises other problems. If your arithmetic gives a drawdown measured in days rather than hours, treat that as a signal to widen the footprint or reduce what drains into it.
Can I put a dry well next to the house?
The physics argues against it and the local rules may prohibit it. A dry well deliberately puts a large volume of water into the ground and holds it there for a day or more; doing that close to a foundation loads the soil that the footing drain and the basement wall are already dealing with, particularly in the backfill zone where the soil is loosest. Required separations from foundations, property lines, wells and septic fields are set locally and are worth confirming before anything is laid out, not after.
Do I need a percolation test, or can I use a soil type?
You need a test. Soil texture gives you an expectation, not a design value, because the rate is governed by whichever layer is slowest and that layer is often something a texture description does not capture: a compacted construction lens, a clay band, a plough pan. Two holes twenty feet apart on the same lot routinely differ by a factor of several. If the pit is going in anyway, the test costs an afternoon and a bucket of water, and it is the only input on this page that cannot be replaced by a reasonable guess.