Rainwater harvesting in Bali

The arithmetic is simple and we do it before quoting: roof plan area multiplied by rainfall, multiplied by a runoff coefficient of about 0.85 for metal or tile. The complication is timing, not volume. Roughly three-quarters of the island's rain arrives between November and March, and August can deliver almost nothing, so tank sizing is really a question of how many dry days you want to bridge.

A working system is four parts: clean collection with leaf screens and a first-flush diverter that dumps the opening millimetre off the roof along with its dust, salt and gecko droppings; storage sized to the gap you are bridging; filtration matched to the use; and an overflow that goes to a recharge pit rather than to the road. Everything is mosquito-sealed as standard, because standing water and dengue are not a theoretical pairing here.

What a Bali roof is actually worth

Run the numbers on an ordinary house. Two hundred square metres of roof plan, about 1,700 millimetres of usable rainfall once evaporation and first-flush losses are taken off, and a runoff coefficient of 0.85 gives roughly 290 cubic metres a year — close to 800 litres a day averaged out. Toilets, laundry and a garden on a modest property rarely need more than that. Where the average stops helping is the calendar: the wet months deliver most of it, and September delivers next to nothing.

Which is why storage sizing is the real design decision. Ten cubic metres bridges roughly a fortnight of dry weather for a household flushing and irrigating on rainwater; going fully off-well through a Bali dry season needs storage most plots cannot fit. We are straight about the split — harvest for what the volume genuinely covers, and keep the well or mains as backup instead of pretending otherwise.

Clean enough for what you want to use it for

Rainwater arrives fairly clean and gets dirty on the way in. Roof dust, salt carried in coastal air, leaf litter, bird and gecko droppings all wash off in the opening minutes of a storm, which is exactly what the first-flush diverter exists for: discard 20 to 40 litres per hundred square metres and what follows is dramatically better water. After that, treatment is a question of use.

Garden and toilet flushing need screening and settlement, nothing more. Laundry and showers want sediment and carbon filtration. Drinking water needs ultrafiltration or reverse osmosis with UV disinfection, plus a laboratory test to prove it — and a tank that is dark, sealed and screened, because sunlight and open inlets grow algae and mosquitoes in that order. We build the storage right first; filtration is far easier to upgrade later than a buried tank.