Rainwater Harvesting Tanks and Cisterns: Capturing Rain at the Roof

What This Page Covers

Rainwater harvesting from roof catchments is the most accessible water storage investment available to a household, farm, or small institution in almost any jurisdiction. The capital cost is well within reach at every scale; the technology is fully mature; the payback is calculable in years. Despite this, most installations fail to reach their potential because of a single design error: undersized storage. A system that can collect 30,000 litres per year from a roof but has only 2,000 litres of tank capacity will fill and overflow during the first rain event of the wet season, capturing 6.7 percent of its potential yield. The same system with 12,000 litres of storage might capture 70 to 80 percent.

This page covers the complete rainwater harvesting system from catchment to end use: the collection efficiency of different roof materials, the sizing formula, the materials comparison for tanks at household to farm scale, the cost per litre stored for each material type, and the regulatory landscape across the jurisdictions where the question most frequently arises. It connects to the water harvesting pillar, which frames all capture methods by their cost per litre delivered and their applicability by site type.

The four tank materials covered here are polyethylene (poly), corrugated galvanised steel with liner, ferrocement, and concrete in-ground cisterns. Each has a distinct cost profile, construction requirement, and service life. The correct choice depends on the volume required, the local availability of skilled labour and materials, the site's frost exposure, and whether portability matters. No single material is best across all contexts; the cost-per-litre calculation below gives the framework for a rational decision.

The Collection System: Catchment to Storage

A rainwater harvesting system has four components: catchment surface, conveyance, first-flush diverter, and storage. The catchment is the roof. The conveyance is the gutter and downpipe system directing roof runoff toward the storage tank. The first-flush diverter captures and discards the initial portion of each rain event, which carries the highest concentration of dust, bird droppings, leaf debris, and atmospheric pollutants accumulated on the roof surface between rain events. Storage is the tank or cistern itself.

Catchment surface affects both yield volume and water quality. Metal roofs (steel, aluminium, zinc-aluminium) have runoff coefficients of 0.75 to 0.90, meaning 75 to 90 percent of incident rainfall becomes collected water after accounting for retention in surface texture and splash losses at gutters. Fired clay or concrete tile roofs have coefficients of 0.70 to 0.85. Asphalt shingle roofs vary from 0.65 to 0.80, with lower coefficients in older installations where surface texture has roughened. Painted or coated metal roofs in good condition achieve coefficients of 0.85 to 0.92. Green (planted) roofs retain 30 to 60 percent of rainfall for plant use and are unsuitable as catchments for storage systems (Thomas and Martinson, 2007, Roofwater Harvesting: A Handbook for Practitioners, WEDC Loughborough).

First-flush volume calculation: the rule of thumb from the Lancaster rainwater harvesting handbook is to divert the first 0.5 to 1.0 litres per square metre of catchment area at the start of each rain event. For a 100 m2 roof, that is 50 to 100 litres per event diverted to a first-flush chamber. The chamber holds this volume, refills with clean water from mid-event onward, and slowly drains to the garden between rain events via a calibrated drip orifice. First-flush diverters reduce collected water turbidity from 50 to 200 NTU to below 5 NTU and fecal coliform counts from thousands per 100 mL to tens per 100 mL, a 99 percent reduction in biological contamination without filtration (Lancaster 2019; WHO Guidelines for Drinking Water Quality, 4th ed., 2011).

Gutter sizing is frequently the hidden constraint. Standard 100 mm wide gutters are designed for 15 mm/hour rainfall intensity. In tropical or sub-tropical climates where design storm intensity exceeds 50 mm/hour, undersized gutters overflow during the heavy rain events that represent the most valuable collection opportunity. Size gutters for the 10-year design storm intensity of your climate zone, not the average annual intensity. For a 100 m2 roof at 50 mm/hour design storm, peak flow is 5,000 litres per hour. A 100 mm gutter handles 1,500 to 2,500 litres per hour at normal slope. Two 100 mm gutters or one 150 mm gutter handles the peak flow adequately.

Sizing, Yield, and Cost per Litre

The sizing formula: Annual yield (L) = Catchment area (m2) x Annual rainfall (mm) x Runoff coefficient / 1,000. For a 150 m2 metal roof in a location with 700 mm annual rainfall and a 0.85 runoff coefficient: 150 x 700 x 0.85 / 1,000 = 89,250 litres per year. Tank sizing for bridging a dry season: multiply your daily demand by the number of days of dry season you need to cover, then reduce by any expected rain during that period. A household using 300 L/day through a 75-day dry season with no rain needs 22,500 L minimum storage. Add 25 percent safety margin: 28,000 L. The nearest available poly tank size is typically 25,000 L or 30,000 L.

At farm scale, the arithmetic scales but the principle does not change. A 500 m2 market garden irrigating at 4 mm/day (40 mm per m2 = 4 mm depth across 10 m2 = 40 litres per 10 m2) uses 2,000 litres per day. Through a 45-day dry period, that is 90,000 litres. If the farm has a 1,000 m2 total roof area (sheds, polytunnels) with 600 mm rainfall and 0.82 coefficient, annual collection is 492,000 litres. Tank sizing for 90,000 L bridge storage: at minimum one 100,000 L corrugated steel tank, or a combination of smaller poly tanks. Cost comparison at this volume makes the steel tank the clear winner on capital cost per litre.

The payback calculation uses the cost of the next-best water supply alternative as the comparison. In Germany, household mains water costs approximately 2.0 to 3.5 EUR per cubic metre. A 5,000 L poly tank costing 400 EUR (at 0.08 EUR/L, installed) collects 25,000 to 40,000 litres per year from a suitable roof. At 2.50 EUR/m3 mains cost, annual water value saved is 62 to 100 EUR. Payback: 4 to 6.5 years. In Australia at 3.00 AUD/m3 (approximately 1.80 EUR), payback is 3 to 5 years. In arid regions where water is trucked at 10 to 30 EUR/m3, the same tank pays back in under 12 months. The cost per litre stored is a fixed capital number; the value per litre saved scales with the local water price, which makes rainwater harvesting economics highly site-specific and most compelling where supply alternatives are most expensive.

Materials Selection and Construction

Polyethylene tanks are the default choice below 25,000 L in most high-income countries because they are factory-manufactured to consistent quality, require no construction skills, arrive on a flatbed truck, and begin collecting water the same day they are installed. They are UV-stabilised (look for 10-year UV warranty minimum) and food-grade for potable use. Installation requires a level, compacted base: concrete slab, compacted gravel, or compacted sand at minimum 100 mm depth. Poly tanks placed on uneven ground develop stress cracks within 3 to 7 years due to differential point loading on the curved base. Connect tanks in series or parallel using standard poly fittings; 25 mm fittings handle domestic flow rates, 40 mm for gravity-fed farm irrigation flows above 10 L/min.

Corrugated steel tanks are site-assembled from pre-cut corrugated steel sheets bolted to a ring base. A 100,000 L tank is 5.5 metres diameter by 4.2 metres tall. Assembly takes two people two days with bolt wrenches and a level. The steel structure provides compression and hoop strength; the liner (food-grade UV-stable polyurethane or HDPE, 0.5 to 1.0 mm) provides water tightness. The liner is draped into the assembled ring before filling. Key construction points: the base must be level to within 10 mm across the full diameter, otherwise the liner stretches unevenly and fails at the low point within 3 to 5 years. Lay 50 mm of compacted sand inside the ring before fitting the liner, to protect it from any sharp base edges.

Ferrocement cisterns are built from a mesh framework (usually chicken wire or weld mesh) plastered on both sides with a rich cement mortar (typically 1:2 cement:sand by weight, minimum 25 mm thick). They require skilled plasterers familiar with thin-shell construction; the mixing ratio, application technique, and curing regime are not negotiable. A poorly rendered ferrocement tank leaks from the start and cannot be repaired cost-effectively. A well-built one lasts 30 to 50 years and costs 0.06 to 0.12 EUR per litre at volumes above 20,000 L. Ferrocement is most appropriate where: trained artisans are available locally, cement and sand are cheap, and the volume required exceeds what off-shelf poly tanks cover at reasonable cost.

Frost-proofing tanks in cold climates is essential: water expanding as it freezes at 0 degrees Celsius generates approximately 2,000 kPa of pressure, sufficient to split a poly tank or crack a concrete cistern. Options: drain tanks completely before the first frost, insulate the tank with 50 to 100 mm closed-cell foam insulation and a weatherproof cladding, or bury below the frost line if in-ground construction is viable. In-ground concrete cisterns are inherently frost-protected if the top slab is below frost penetration depth for the climate zone. Metal tanks are more frost-tolerant than poly because the steel can deform and return without cracking; the liner is vulnerable and should be checked for cracking after any freeze-thaw cycle.

Integration with the Broader Water Stack

Tanks and cisterns are the first-line storage for roof catchments, but they do not stand alone in a complete water management system. The overflow from a full tank is a high-value resource, not waste. Directing tank overflow to a garden swale, infiltration trench, or spread zone recharges soil moisture and shallow groundwater rather than contributing to runoff. This connection to the earthworks strategy in the water harvesting pillar means that a properly designed household system captures rain twice: first in the tank, then in the landscape as overflow infiltrates.

At farm scale, tank storage pairs most directly with pond storage for different time horizons. Tanks cover daily to seasonal variation (days to months); ponds cover inter-annual variation (months to years). A farm with both a 100,000 L tank array and a well-designed farm pond can bridge dry spells at different timescales: tank water for high-value crop irrigation during a 30-day dry spell, pond water for pasture during a multi-month drought. The tank captures the roof runoff that would otherwise miss the pond catchment entirely.

The connection to managed aquifer recharge is through overflow design. A farm tank system producing 200,000 litres per year with an average overflow of 30,000 litres during wet season peak events can direct that overflow to a dedicated infiltration area sized to receive it, banking the overflow in the local shallow aquifer for recovery during the dry season through bores. This requires 60 to 120 m2 of infiltration area in permeable soil, adding less than 5 percent to the cost of the overall system while doubling its effective water banking capacity.

For farms using water for composting, the nutrient cycle argument parallels the water argument. A composting system requires consistent moisture during the thermophilic phase; a tank system providing 30 to 50 litres per day for pile moisture management from roof runoff reduces the labour and infrastructure cost of a composting operation significantly in dry climates. This is a small volume application that sits well within the first-use tier of a properly sized tank before irrigation demand draws down the main volume.

The regulatory landscape (covered in the grid above) is the primary adoption constraint in some jurisdictions but not in most. The practitioner in the EU, Australia, or eastern US faces no meaningful legal constraint on tank installation for non-potable use at any volume. The practitioner in western US water law states should verify current state statute before investing in large systems: the law is evolving rapidly in Colorado, Nevada, and Utah as drought stress accelerates policy change, with most states having relaxed restrictions since 2010.