Define the boundary of efficiency

An efficiency value needs a beginning and an end. Roof runoff relative to rainfall is different from tank inflow relative to roof runoff, and both differ from useful withdrawals relative to annual rainfall. Write the boundary beside every coefficient. Without that description, two apparently conflicting percentages may simply describe different stages of the same system.

Build a loss ledger

List losses once: roof wetting, bypass at gutters, diverted first water, filter rejection, storage overflow and unusable residue. Some are proportional to rain; others depend on events, flow intensity or tank state. A simple combined factor is suitable for early screening. A detailed simulation should represent separate processes when measurements support doing so.

THE EXAMPLE AT A GLANCEThe boundary decides which factor is valid
Separate roof and collection losses720 L
One all-inclusive capture factor800 L

Multiply 0.90 × 0.80 only when the two factors cover independent losses.

100 m² × 10 mm = 1,000 L; two alternative coefficient definitions.

See what double counting does

Take 100 m² and 10 mm rain: gross volume is 1,000 litres. A roof factor of 0.90 followed by an independent collection factor of 0.80 gives 720 litres. But if the 0.80 value already covers all roof and collection losses, multiplying it again by 0.90 incorrectly reduces the result to 720 instead of 800 litres. The definitions decide which calculation is valid.

Event losses can dominate small storms

Suppose an illustrative device diverts a fixed 20 litres for each separate event. Five 100 litre inflows lose 100 litres to diversion; one 500 litre inflow loses only 20 litres. The gross total is identical but retained volume differs. This example explains why an annual percentage cannot precisely reproduce a system whose losses reset after each rainfall event.

Calibrate without hiding faults

Collect paired rainfall and tank-level measurements, allowing for withdrawals and overflow. Estimate a plausible range rather than fitting every storm with a different unexplained coefficient. If performance suddenly changes, inspect the installation. Treating a blocked screen as a new normal efficiency conceals a maintenance problem and can make the next sizing decision unnecessarily expensive.

Coefficient checklist

  • State the start and end of each factor.
  • Keep values between zero and one.
  • Separate assumptions from measurements.
  • Avoid counting roof losses twice.
  • Test low and high scenarios.
  • Record fixed diversion losses separately when known.
  • Explain why annual useful yield can be lower than tank inflow.