Week 7: What a roof can actually deliver
Week 7 — the enHealth runoff formula, a capped bore, and the month-by-month supply curve that every later water decision is built on
19 April 2027
The question#
You now know what the garden wants (week 2) and what the household wants (week 1). How much water does the site actually produce, and in which months does it produce it?
Not the annual total. The months. Annual totals are how people end up with a tank that runs dry in September.
The method: enHealth Appendix B#
runoff (L) = A × (rainfall − B) × roof area- A is collection efficiency, 0.80–0.85 in the source, citing Martin (1980). This course uses 0.80.
- B is initial loss, 2 mm per month, or 24 mm a year.
The source’s own worked example: 0.8 × (750 − 24) × 200 = 116,160 L. If your spreadsheet does not reproduce that line exactly, fix it before going further.
Apply it per month, not to the annual total, or a month with 1 mm of rain contributes a negative number.
Two supporting rules. First flush: divert about 10 L per 50 m² of roof at the start of each event — a figure confirmed verbatim on the Australian Government’s Your Home rainwater page. And the useful identity behind all of it: 1 mm of rain on 1 m² of roof is 1 L.
Worked example: Alice Springs, the hardest site there is#
Alice Springs has the largest roof in the set after Canberra — 350 m² across house and machinery shed — and the least rain: 284 mm a year.
0.80 × (284 − 24) mm × 350 m² = 72,800 L = 73 kL/yearDomestic demand is 219 kL. The roof covers a third of the household, and none of the garden.
Alice Springs is the only site with a second source: a bore allocation capped at 100 kL/year. Add it and the site has 173 kL against a domestic demand of 219 kL and a garden asking for 1,650 mm per m² — which over 400 m² is 660 kL. Total demand approaching 880 kL against a supply of 173 kL.
This is why the excellent Alice Springs reference design reaches 97 % water closure with 16 % garden water satisfaction. It closes the loop by watering almost nothing. Both numbers are published together, always, precisely so that a design cannot claim the first without disclosing the second.
Notice what a capped allocation does to the design problem: it converts a resource question into an allocation question. You are no longer asking how much water exists. You are choosing who gets it.
Five sites: annual roof yield against 219 kL of domestic demand#
| Site | Roof | Rain | Roof yield | As a share of domestic demand | Other source |
|---|---|---|---|---|---|
| Darwin | 300 m² | 1,733 mm | 411 kL | 188 % | — |
| Brisbane | 170 m² | 1,146 mm | 153 kL | 70 % | — |
| Canberra | 300 m² | 616 mm | 142 kL | 65 % | — |
| Adelaide | 230 m² | 525 mm | 92 kL | 42 % | — |
| Alice Springs | 350 m² | 284 mm | 73 kL | 33 % | Bore, 100 kL/yr capped |
Darwin collects nearly twice what the household drinks and still imports water in the dry season. Holding both of those facts at once is what week 9 is for.
What this week hands you#
A twelve-row monthly supply curve for your site, in kilolitres, with the first flush deducted and every source named. Put it beside week 2’s demand curve. The gap between them is week 9’s problem, and it is not a single number.
This is also the week the story turns, and the order matters. Assignment 1 is due at 5 pm on Friday. This lecture, on the Monday, previews what a measurement year at your home produced — enough for you to see why the register you are about to submit was worth writing — and the full Release B handover follows the deadline. Nothing in the preview changes what your plan should have said, and no calculation before Friday is an observation of this house.
From week 8 on, your inputs are findings from the measured year rather than the scoping estimates weeks 2 to 6 produced.
Teaching team
- Teodor Vasilache — tutor