Measure how fast your pit fills, instead of guessing

The procedure, in order

1. Wait for real rain. A test run in dry weather measures the water table, which is useful but is not the storm you are sizing for. The number that matters is the peak, so test during sustained heavy rain and take the worst reading you get.
2. Measure the basin, not the lid. Inside diameter of the liner, at the water line. A lid overhangs, and using the lid dimension inflates every gallon figure that follows.
3. Mark the starting water level. A strip of tape on the liner wall beats eyeballing a rough surface. You are going to read a difference of a few inches and a half-inch of parallax is a ten percent error.
4. Unplug the pump. Stay there. This is the part of the procedure where an unattended basement floods, and no measurement is worth that.
5. Time the rise. Sixty seconds or more if the pit fills slowly enough to allow it; the shorter the window, the more your reaction time at each end distorts the result.
6. Plug the pump back in immediately and confirm it starts and clears the basin.
7. Repeat during the next storm. One reading is a data point; three readings across different storms tell you whether the first one was the peak or a lull.

Convert your rise into gallons per minute

The arithmetic, shown in full

There is no black box here, so it is worth writing out. A one-inch-deep slice of a round basin is a cylinder: its volume is pi over four, times the diameter squared, times one inch. The United States liquid gallon is defined as exactly 231 cubic inches, so dividing by 231 converts that slice into gallons and gives a constant of 0.00340 gallons per inch of depth for every squared inch of diameter.

For the 15-inch basin above: 0.00340 multiplied by 15 squared gives 0.765 gallons per inch. A 3-inch rise is therefore 2.29 gallons, and 2.29 gallons in 120 seconds is 1.15 gallons per minute.

Nothing in that chain is estimated. It is geometry and a unit definition, which is precisely why a measurement is worth so much more than the alternative.

Why a long test reads low

There is a bias built into this procedure and it runs in the direction that matters. Water enters the basin because the drain tile around your footing sits at a higher pressure than the basin does, and as the level in the pit climbs, that difference shrinks and the flow into the pit falls with it. Unplug the pump, let the water rise a foot, and you have measured an average taken over a period in which you were steadily throttling the thing you were trying to measure.

So keep the rise small relative to the depth below the tile inlet. Two or three inches in a basin whose drain tile enters a foot or more down costs you almost nothing; a rise that reaches the inlet costs you a great deal, and a basin backed up far enough that water is standing in the tile has stopped measuring inflow altogether. If a slow pit forces a long window, read the first minute of the rise rather than the average across five — the first minute is the one taken at close to the level the pump normally holds, which is the condition you are actually sizing for.

The bias only ever runs one way, which is the useful part: a test that is too long understates your inflow, and nothing about the procedure inflates it. If your reading feels high, it is high because the storm was, not because the method flattered it.

Reading the result honestly

A rate measured in light rain understates your peak, and sizing a pump to it will leave you short on the night that matters. A rate measured while a downspout was discharging next to the foundation overstates it, and will sell you a pump larger than the house needs. Both errors are avoided the same way: test more than once, note the weather each time, and use the highest credible reading.

If your readings vary by less than about twenty percent across several storms, the drainage around the house is behaving consistently and the number is trustworthy. If one storm produces a reading several times the others, something intermittent is contributing — a gutter that overflows only in heavy rain, a neighbour's downspout, a window well without a drain. That is worth finding, because it is almost always cheaper to redirect than to pump.

One more sanity check: a residential footing drain taking more than about sixty gallons a minute is unusual. At that rate, look for a broken supply line or a storm connection before you shop for a larger pump.

Was the storm you measured actually a big one?

A measured rate answers how fast, not how rare, and sizing is a question about rarity. Two inches of rain in an afternoon is ordinary on parts of the Gulf coast and close to remarkable in parts of the Pacific Northwest, so a reading taken in what felt like heavy rain might represent a storm your address sees several times a year or one it sees twice in a lifetime. Those two readings call for very different pumps, and nothing in your pit distinguishes them. A rain gauge standing next to the test does, as long as you look the depth up afterwards.

The National Weather Service publishes the reference for exactly that, free and by coordinates:

The Precipitation Frequency Data Server (PFDS) is a point-and-click interface developed to deliver NOAA Atlas 14 precipitation frequency estimates and associated information.

For your decision, the useful move is this: pick the duration that matches your test window, read across the return periods for your own point on the map, and find where the depth you actually caught falls. A depth near the one-year column means you measured a storm your house sees annually — if the basin was already struggling at that, the problem is drainage rather than pump capacity, and a larger pump will only move the failure. A depth out near the twenty-five or hundred-year column means you caught the storm worth sizing for, and the number in front of you is the one to use.

If you caught something ordinary and want to plan for something rare, scale up: take the ratio of the rarer depth to the depth you measured and multiply your inflow by it. That scaling is crude, because infiltration is not linear in rainfall and saturated ground does not behave like dry ground, and it will usually read low for a long soaking event. It is still a far better footing than guessing what fraction of the roof reaches the drain, because both of its inputs — your measurement and a published depth for your coordinates — are real.

Take the measurement back to the sizing tool

• Sizing calculator — enter 1.15 gpm as a measured rise and get the flow-and-head specification.
• Backup runtime — inflow is what sets the duty cycle, and therefore how long a battery lasts.
• Operating cost — annual run hours come straight from how many gallons have to move.