Log reductions, not reassurance#
SLOP4761 · Week 8 · Treatment, reuse and service requirements
Where we are#
First design week. The handover has happened. From here your inputs are the measured year at your home, not scoping estimates.
The water you are treating is the water your findings page reports — including the sample taken after heavy rain.
This week’s question#
Roof water is not clean.
What does each barrier in your treatment train actually remove, expressed as a number somebody else could check, and what does running it cost in energy?
“Filtered” is not an answer. Neither is “we boil it if we are worried”.
By the end you can#
- Assemble a barrier train with a sourced log reduction per pathogen class
- Tell a laboratory figure from a field figure, and choose the right one
- Separate continuous from intermittent energy load and total both
A log reduction is an order of magnitude#
1 log = 90 %. 2 logs = 99 %. 3 logs = 99.9 %. In series, logs add.
WHO rates household water treatment separately for bacteria, viruses and protozoa, because a technology strong against one can be useless against another:
| Rating | Bacteria | Viruses | Protozoa |
|---|---|---|---|
| Three star | ≥4 log | ≥5 log | ≥4 log |
| Two star | ≥2 log | ≥3 log | ≥2 log |
A single “log reduction” figure with no class attached is not usable.
What the barriers actually do#
- First flush and leaf screen — no log rating. It removes the load the rest of the train would otherwise carry
- Sedimentation and cartridge pre-filter — turbidity control. UV’s job is impossible through turbid water
The two that carry numbers#
BioSand filter — about 2.5 log for E. coli once mature. But 1.67 log run intermittently, against 3.71 log continuous, and a field average across studies of about 1.1 log.
UV disinfection — Class A delivers ≥40 mJ/cm² over lamp life, with a sensor and a flow restrictor. Lamps replaced every 9–12 months — an operating cost, not a capital one.
The gap between laboratory and field is this week.
One scope correction#
No Australian retailer page reviewed for this course claimed NSF/ANSI 55 Class A. Australian units cite WaterMark, which is a different scheme answering a different question.
Do not write a Class A claim into a design you have not sourced.
Worked example: Brisbane#
Smallest roof in the set (170 m²), most intense rain (1,146 mm). First flush at 10 L per 50 m²:
170 ÷ 50 × 10 L = 34 L diverted per event
Real water lost, and Brisbane can least afford it — divert less and the pre-filter clogs faster and the UV sees higher turbidity. That is the trade.
Energy:
UV lamp 25 W × 24 h = 0.60 kWh/day
12 % of the 5 kWh/day allowance, running whether or not anyone draws water. The largest constant load in the reference design.
Your tutorial task#
- Quiz 3, twenty minutes, on weeks 6 to 8
- Read your own post-rain turbidity and E. coli first, and write them at the top
- Build the train left to right, a number on every box, by pathogen class. The box with no number is the interesting one — mark it unquantified
- Decide which BioSand figure applies to a household drawing in bursts, and write why
- Total the energy, continuous and intermittent separately
Output: a barrier train with sourced log reductions and a costed running load.
Before you leave#
- Does every barrier carry figures by class, or the word unquantified?
- Does your sheet distinguish laboratory from field where the source does?
- Is every certification one your cited product actually holds?
- Would the train work on your water, or does the UV sit behind something too turbid?
Full sources in the week 8 notes.