A filtering cistern is a water storage tank with an integrated multi-stage filter train — typically a screen, sediment chamber, baffle wall, and calmed inlet — that cleans roof or surface runoff before it enters the main storage volume. The Roman engineer Vitruvius described early two-chamber settling cisterns in De Architectura around 25 BC. Incoming water sheds debris on a coarse screen, drops fines in a quiescent chamber, then rises through a baffle into the clean storage side. The result is potable-grade or irrigation-grade water without active pumping, used today in systems from German Wisy filters to large agricultural tanks.
Filtering Cistern Interactive Calculator
Vary the settling-flow velocity and residence time to see the required quiet flow path and short-circuit risk in a filtering cistern.
Equation Used
This calculator applies the cistern settling rule from the article: keep the sediment chamber velocity at or below 0.01 m/s and provide about 60 to 120 seconds of residence time. The estimated quiet flow path is the simple product of velocity and residence time.
- Uses the article design target of flow velocity at or below 0.01 m/s.
- Residence time is evaluated against the article range of 60 to 120 seconds.
- Flow path length is a plug-flow teaching estimate, not a CFD model.
How the Filtering Cistern Actually Works
A filtering cistern works on gravity and residence time, not pressure. Roof runoff enters through a downpipe, passes a coarse leaf screen — usually 0.4 to 1.0 mm stainless mesh — and drops into a sediment chamber sized so flow velocity falls below 0.01 m/s. At that velocity, sand, pollen, and organic fines settle out within 60 to 120 seconds. The water then rises against a baffle wall, which forces it to change direction twice before spilling into the storage chamber through a calmed inlet that points downward to avoid stirring up the settled sludge layer at the bottom.
Why build it this way? Because the failure modes are all about short-circuiting. If the baffle wall stops 50 mm short of the design level, peak storm flow blows straight across the top of the sediment chamber and dumps unfiltered water into storage — you would be amazed how little baffle gap it takes to ruin the whole system. If the calmed inlet sits too close to the floor, it scours the sludge blanket and resuspends everything you spent six months settling. And if the overflow trap on the far end isn't a proper U-bend with a vermin screen, mosquitoes breed in the storage volume within a week of warm weather.
The first flush diverter is the other critical element. The first 1 to 2 mm of rainfall off a roof carries roughly 80% of the bird droppings, dust, and shingle grit that accumulated since the last storm. A diverter — typically a vertical standpipe with a floating ball valve or a slow-release orifice — captures that first slug and sends it to waste. Skip the first flush and your sediment chamber loads up three to four times faster, and you'll be cleaning it twice a year instead of every three.
Key Components
- Coarse Leaf Screen: Stainless mesh, 0.4 to 1.0 mm aperture, mounted at the downpipe inlet. Catches leaves, twigs, and insects before they enter the sediment chamber. Self-cleaning sloped designs like the Wisy WFF need clearing only twice a year; flat screens clog in weeks.
- First Flush Diverter: A standpipe sized at roughly 1 litre per square metre of roof area, fitted with a ball-float seal or a 3 mm bleed orifice. Captures the contaminated first slug of runoff and discharges it to waste over 4 to 24 hours after the storm.
- Sediment Chamber: A quiescent settling volume sized for flow velocity below 0.01 m/s at design storm intensity. Residence time of 60 to 120 seconds drops particles down to about 50 µm. Floor sloped 1:50 toward a cleanout drain.
- Baffle Wall: A vertical partition that forces water to flow under, over, or around — typically two direction changes. Top of baffle sits at least 100 mm above the static water line to prevent surface short-circuiting during peak inflow.
- Calmed Inlet: A downward-facing 90° elbow or a perforated diffuser at the entry to the storage chamber. Discharges horizontally near mid-depth to avoid disturbing the sludge blanket on the floor and the biofilm on the surface.
- Floating Suction Filter: The pump intake hangs from a float roughly 150 mm below the water surface, drawing from the cleanest layer between sludge and biofilm. Includes a 0.3 mm fine screen and a check valve to keep the pump primed.
- Overflow Trap with Vermin Screen: A U-bend siphon on the overflow line with a 1 mm stainless screen. Skims floating debris off the top during overflow events and blocks rodents, mosquitoes, and frogs from entering the cistern.
Industries That Rely on the Filtering Cistern
Filtering cisterns show up wherever you need clean water from an uncontrolled source without continuous power. The same basic two-chamber-plus-baffle design scales from a 200 litre garden barrel to a 500,000 litre municipal harvesting system. What changes is the screen aperture, the first flush volume, and whether you add UV polishing downstream — the gravity filter train itself stays the same.
- Residential Rainwater Harvesting: A typical German Wisy WFF 150 filter feeding a 6,000 litre underground tank for a single-family home, supplying toilet flushing and laundry.
- Agriculture: Australian Colorbond roof catchment systems on Outback sheep stations, feeding 90,000 litre Pioneer Water Tanks through a leaf eater and first flush diverter for stock and homestead use.
- Heritage Restoration: Restored Roman-era impluvium and cistern at the Casa del Menandro in Pompeii, originally a two-chamber settling system for household water.
- Off-Grid Cabins: Backwoods cabins in coastal British Columbia using a Rain Harvesting Pty leaf eater plus a 4,500 litre Norwesco poly tank with internal baffle and calmed inlet.
- Commercial Buildings: The Bullitt Center in Seattle harvests roof runoff into a 56,000 gallon basement cistern with multi-stage filtration before UV polishing for potable use.
- Disaster Relief and Remote Schools: UNICEF and CAWST community-scale ferrocement cisterns in rural Nepal and Honduras, with built-in screen and baffle and a hand-pump suction filter.
The Formula Behind the Filtering Cistern
The single number that makes or breaks a filtering cistern is the sediment chamber residence time — how long water sits in the quiescent volume before it spills past the baffle. Too short and fines never settle; too long means you've oversized the chamber and starved the storage volume. At the low end of the typical range (around 30 seconds) only sand and large pollen drop out. At the nominal 90 seconds, particles down to roughly 50 µm settle. Push past 180 seconds and you're getting diminishing returns — you'd need a flocculant to drop colloidal clay any further. The sweet spot for most roof catchment is 60 to 120 seconds at design storm intensity.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| tr | Residence time in the sediment chamber at peak design flow | s | s |
| Vchamber | Working volume of the sediment chamber below the baffle overflow level | m³ | ft³ |
| Qpeak | Peak inflow rate from the design storm event after the first flush diverter | m³/s | ft³/s |
| Aroof | Effective horizontal roof catchment area | m² | ft² |
| i | Design rainfall intensity (5-minute peak for the 10-year storm) | mm/hr | in/hr |
Worked Example: Filtering Cistern in a passive house rainwater cistern in coastal Maine
You are sizing the sediment chamber and baffle for a passive house rainwater harvesting system in Camden, Maine. The standing-seam metal roof has an effective catchment area of 180 m², the design rainfall intensity is 75 mm/hr (5-minute peak, 10-year storm for coastal Maine), and the system feeds a 9,000 litre Wisy underground cistern with a WFF 300 vortex pre-filter that passes 90% of inflow. You need to confirm the sediment chamber gives at least 60 seconds of residence time at peak flow.
Given
- Aroof = 180 m²
- i = 75 mm/hr
- Filter pass-through = 0.90°—
- Vchamber (design) = 0.35 m³
Solution
Step 1 — convert design rainfall to peak inflow off the roof:
Step 2 — apply the WFF 300 pass-through ratio of 0.90 to get peak flow into the sediment chamber:
Step 3 — compute residence time at the nominal design storm:
That puts you comfortably in the 60-120 second sweet spot. Particles down to roughly 50 µm — fine sand, pollen, shingle grit — drop out cleanly. Now check the operating-range extremes.
At the low end, a typical 5 mm/hr drizzle gives Qpeak = 0.000225 m³/s and residence time of:
The chamber is glassy-still — even silt fines drop out, and you'd actually start growing a biofilm on the chamber floor if drizzle persisted for days. That's fine; the next storm will scour it.
At the high end, a 150 mm/hr cloudburst (well above the 10-year storm) doubles the inflow:
52 seconds is below the 60 s minimum — you'll see slight turbidity carry-over on the worst storms of the decade. That's an acceptable trade for a system you're not running through a clarifier. If you wanted to hold 60 s even at 150 mm/hr, you'd need to bump Vchamber to 0.41 m³ — about a 17% increase in chamber size.
Result
Nominal residence time is 104 seconds at the 75 mm/hr design storm, which is right in the sweet spot for a residential filtering cistern. The full operating range runs from roughly 26 minutes at light drizzle (over-clean, no harm done) to 52 seconds at a 150 mm/hr cloudburst (slight turbidity bypass on the worst storms — acceptable for non-potable use, marginal if you're polishing for potable). If your measured turbidity downstream is consistently higher than expected, the three failure modes I see most often are: (1) the baffle wall installed 30-50 mm too short, letting peak flow short-circuit across the top — pull a tape measure on the baffle height before blaming the filter; (2) a clogged WFF screen reducing pass-through below the 0.90 spec, which paradoxically forces more flow over the bypass weir and into raw storage; or (3) the calmed inlet pointing upward instead of downward at 30°, which kicks the sludge blanket back into suspension every time a storm hits.
Choosing the Filtering Cistern: Pros and Cons
A filtering cistern is one of three common ways to clean roof or surface runoff before storage. Each handles a different flow regime, particle size range, and maintenance budget — pick wrong and you'll either be cleaning filters every weekend or drinking turbid water.
| Property | Filtering Cistern (two-chamber + baffle) | Inline Cartridge Filter | Slow Sand Filter |
|---|---|---|---|
| Particle removal floor | ~50 µm gravity settling | 5-20 µm depending on cartridge | 1-5 µm with biofilm developed |
| Peak flow capacity | Up to 0.01 m³/s with proper sizing | 0.0005-0.002 m³/s typical | 0.0001-0.0005 m³/s per m² bed |
| Maintenance interval | Sludge cleanout every 12-24 months | Cartridge replacement every 1-3 months | Scrape sand top every 2-6 months |
| Power requirement | None — fully passive gravity | None passive, but pressure drops 0.3-1.0 bar | None passive |
| Installed cost (residential scale) | $400-1,500 added to tank cost | $80-300 plus housing | $2,000-6,000 for purpose-built bed |
| Lifespan of core element | 30-50 years (concrete or HDPE tank) | Cartridge 1-3 months, housing 5-10 yr | Sand bed 8-15 years before full rebuild |
| Best application fit | Roof catchment, 5,000-500,000 L systems | Polishing stage downstream of cistern | Potable polishing for community systems |
Frequently Asked Questions About Filtering Cistern
Almost always it's resuspension, not insufficient settling. When peak inflow hits the chamber faster than the baffle can damp it, the jet stirs the sludge blanket on the floor and lifts fines back into the water column. Check your inlet drop — if water free-falls more than 200 mm into the chamber, the splash energy alone will resuspend everything you settled in the previous week.
The fix is a tranquilising inlet: a vertical pipe terminating in a 90° downward elbow set 100-150 mm above the floor, or a tee with horizontal outlets. Cuts inflow velocity by roughly 4× and the sludge blanket stays put.
Run the standard rule: 1 litre of diverter volume per square metre of roof, then check the slow-release orifice empties the standpipe in 4 to 24 hours after a storm. If the standpipe is still full 48 hours later you'll lose the next storm's first flush entirely — water just rises past it into the cistern. If it empties in under 2 hours you're losing useful captured volume to the bleed.
The 24-hour upper bound exists because the diverter has to reset before the next storm event statistically arrives. In wet climates like the Pacific Northwest, drop that to 12 hours.
For that roof size, a vortex filter is the better engineering choice. Vortex designs use centrifugal action across an angled mesh — water spirals down, debris slides off to waste, and pass-through stays around 90% even when the mesh is partially fouled. Basket leaf eaters drop to 30-40% pass-through once leaves accumulate, and on a wet autumn week you'll be cleaning them every other day.
The cost difference is roughly $200 vs $40, but the labour savings show up within the first year. Below 50 m² of roof a basket is fine; above 100 m² the vortex pays back fast.
Two likely causes I see in the field. First, the float length is wrong — if the suction sits less than 100 mm below the surface, wave action and inflow ripples drag the surface biofilm into the intake. Bump the float tether to 150-200 mm minimum. Second, the suction strainer aperture is too coarse; a 1 mm screen on a floating intake passes anything smaller than fine sand. Specify a 0.3 mm screen if you're feeding a UV polisher or any potable use.
One more check: if the pump cycles on and off during draw, transient pressure swings can collapse the float and dunk the strainer into the sludge layer. A pressure tank or VFD-driven pump fixes that.
In a well-designed two-chamber cistern, the sediment chamber takes 4-8% of total tank volume. So on a 9,000 L system you give up roughly 350-700 L of working storage. That's the trade for passive filtration with no consumables — and the alternative is either a pressurised cartridge train (which has its own pressure-drop and replacement costs) or unfiltered water full of shingle grit destroying your pump impellers within 2 years.
If you're tight on volume, an external pre-filter chamber bolted onto the inlet pipe is an option. It moves the 5% volume penalty outside the main tank, but you've now got an extra concrete or HDPE vessel to maintain.
You can retrofit, and it's usually worth doing. Drop in a polyethylene baffle wall sealed to the tank floor and walls with a marine-grade sealant like 3M 5200, leaving a 100 mm gap top and bottom for flow. Add a calmed inlet on the dirty side and move the suction to the clean side. Cost is typically $200-500 in materials plus a confined-space entry.
The one case where it's not worth it: if your existing tank is under 2,000 litres, you don't have the floor area to give up 400 mm of width to a sediment chamber. Below that size, an external Wisy or 3P filter ahead of the tank is the better solution.
References & Further Reading
- Wikipedia contributors. Rainwater harvesting. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
