Water In & Out of a Building

The single most useful thing you can do to a building, for water, is to draw a boundary around it and treat it as a system you must account for. Picture a dashed line around the whole property. Pipes cross that line: some carry water in, some carry water out. The building water balance simply says that, over any stretch of time, the water crossing inward equals the water crossing outward, plus or minus the change in water stored inside (in tanks, cisterns, the soil of the garden). In steady use, in equals out. It is the same water balance from lesson 2.1 - rain in equals evapotranspiration plus infiltration plus runoff - now applied across a building's own boundary, with taps and drains instead of soil and sky.

This framing sounds trivial and is quietly revolutionary, because the conventional way of thinking about building water is not a balance at all - it is a one-way habit. Water 'comes from the tap' (we rarely ask from where, or at what cost to a distant aquifer) and 'goes down the drain' (we never ask to where, or in what state). There is no ledger, no sense that the inputs and outputs are connected, no notion that the litres leaving might have served again or the litres arriving might have been caught for free. The balance view refuses that amnesia. It forces three questions that structure this entire lesson and, really, this entire course: what comes in (and from which source, at what grade, at what cost and reliability)? What is it used for (and does each use actually need the grade of water it gets)? And what leaves (in how many separate streams, in what condition, and could any of it have been reused)?

Why insist on the ledger before any technology? Because regenerative water is a set of interventions on these flows - reduce a flow, catch a flow, redirect a flow from waste to reuse - and you cannot intervene sensibly on flows you have not mapped. A greywater system is a redirection of the greywater output back to the flushing input; rainwater harvesting is a new, free input substituting for mains; demand reduction shrinks the whole balance. Every one of them is a change to this diagram. So before we specify anything, we learn to draw it: the building as an accountable water system, with an honest balance of in, used, and out. The rest of the lesson fills in each side.

Start with the inputs: every litre a building uses arrives from one of a handful of sources, and a regenerative designer treats the mix as a deliberate choice, not an accident of what the plot happens to have. The most common input is mains supply - treated, drinking-quality water piped from a municipal network. It is convenient and (usually) reliable and safe, but it is drawn and treated at real energy and cost from a distant, often stressed source, and in much of India it is intermittent, delivered only for a few hours a day. The second common input, especially where mains are weak or absent, is a borewell drawing groundwater from beneath the site. It feels free and private, but it is the very abstraction lesson 2.1 warned about: across much of India borewells are pumping a falling water table, and 'free' groundwater is often the least sustainable input of all.

Then come the inputs regenerative design wants to grow. Harvested rainwater - collected from roofs and stored - is naturally clean, free after the plumbing, and (crucially) fits the local water cycle instead of overdrawing a distant source; it is the flagship regenerative input and the subject of Module 3. Recycled water - treated greywater or blackwater produced on site - is really an output looped back to become an input, the heart of Module 4. And where supply fails entirely, buildings fall back on tanker water trucked in - expensive, carbon-heavy, of uncertain quality, and a sign that the balance has broken down (Indian cities in shortage run on tankers). Some sites also have specialised inputs (a well, a stream abstraction, or in rare cases desalinated or atmospheric water), but these are edge cases.

The designer's job on the supply side is to see this as a portfolio to be shifted, not a given to be accepted. The conventional building leans wholly on mains and borewell - distant abstraction and groundwater overdraft - and treats every drop, whatever the source, to a single drinking grade. The regenerative building deliberately grows the local, low-impact inputs (rain first, then safely-treated reuse) so it leans less on distant mains and falling groundwater, improving both sustainability and resilience: a building with its own harvested and recycled water is far less exposed when the mains fail or the tanker does not come. But - the discipline holds - a new input never comes before reducing the demand it would serve, and any recycled or harvested input is safe only to the grade it is treated and verified for, which the specialists and codes decide.

Once inside, the water is used - and this is where the linear model's central flaw becomes visible and quantifiable. List the uses in a typical building and sort them by how much water they take: in a home, toilet flushing is one of the largest single uses, followed by bathing and showers, then laundry, outdoor and garden watering, cleaning, and finally the small but vital sliver for drinking and cooking. In offices and institutions the mix shifts (flushing and cooling loom even larger, showers shrink), and cooling towers, landscape irrigation and process uses can dominate a large commercial building. The exact shares vary enormously by climate, culture and building type - the numbers here are illustrative, not a specification - but one pattern holds almost everywhere and is the single most important fact in this lesson.

That pattern is the quality mismatch: the overwhelming majority of the water a building uses does not need to be drinking quality, yet the conventional building supplies drinking-quality water to all of it. Only drinking and cooking genuinely require potable water, and that is a small fraction of total demand. Flushing a toilet with treated, chlorinated, energy-intensive drinking water is, put plainly, absurd - the toilet does not care; nor does the garden, the washing machine, the mop bucket or the cooling tower. We are using filtered mineral water to flush, wash and irrigate. Seeing the end-use breakdown makes this waste concrete: it shows, in litres, exactly how much drinking-grade water is being spent on tasks that a lower grade would serve perfectly well.

This is the opening that regenerative water exploits, and it is worth stating as the strategic pivot of the module. Because most demand is non-potable, most demand can in principle be met by non-potable sources - harvested rainwater and treated greywater - reserving precious potable supply for the small potable need. This is fit-for-purpose matching (introduced in 0.1, detailed in 4.4 and 2.3): match the quality of water to the quality the task requires. Map the uses, mark which need potable water (few) and which do not (most), and you have simultaneously found the biggest target for demand reduction (the large flushing and bathing uses) and the biggest opportunity for reuse (the large non-potable demand a recycled source could feed). The use map is where the strategy is born - but only after the discipline: cut the wasteful demand first, and let specialists and codes confirm what quality is truly safe for each use.

Now the output side - and here regenerative design makes its most important observation: what leaves a building is not one undifferentiated 'wastewater' but three distinct streams, of very different quality and very different reuse potential, and the whole game is to stop mixing them. Greywater is the relatively lightly-used water from basins, showers, baths and (usually) laundry - it carries soap, hair, skin, some grease and lint, but not the heavy contamination of sewage. It is the largest and cleanest waste stream, and therefore the easiest and most valuable to reuse: treated appropriately, greywater can serve flushing and irrigation, closing much of the loop. Blackwater is the heavily contaminated stream from toilets (and, because of food waste and grease, usually the kitchen) - it carries pathogens and high organic and nutrient loads, demands serious treatment, and is the stream where the health discipline is most acute. Stormwater is the rain that runs off the building's roofs and paved surfaces - relatively clean (bar the dirty first flush and whatever the surfaces add), and best captured for harvesting or infiltrated to recharge groundwater rather than piped away.

The tragedy of the conventional building is that it combines these three the moment they leave the fixture: greywater, blackwater and often stormwater are plumbed into a single drain and become, all together, 'sewage' - a large volume of uniformly contaminated water that must now be treated (or, too often, discharged untreated) as if all of it were as dirty as the toilet water. Mixing a little blackwater into a lot of greywater does not clean the blackwater; it contaminates the greywater, destroying the reuse value of the cleanest and largest stream. It is the single most wasteful move in building plumbing.

Regenerative design does the opposite: it keeps the streams separate at source, because separated they are resources and combined they are waste. Kept apart, greywater can be lightly treated and reused for non-potable demand, stormwater can be harvested or recharged, and only the small, genuinely contaminated blackwater stream needs heavy treatment (and can even be handled with resource-recovery approaches, Module 4.2). This is why 'seeing the flows' is not a preliminary but the strategy itself: the moment you draw a building's outputs as three streams rather than one drain, the loop's closure points become obvious - reroute greywater to flushing, harvest the stormwater, isolate and properly treat the blackwater. And every reuse, of course, stays bound by the absolute rules: rigorous separation of non-potable from potable, treatment to a genuinely safe grade, and the binding water-quality, plumbing and public-health decisions left to qualified specialists and the governing codes.