One millimetre of rain falling on one square metre is one litre of water. That identity is the whole of rainwater harvesting for farms, and everything else is losses and storage. Multiply your roof footprint in square metres by the rain in millimetres, take off about fifteen per cent for what the system loses, and you have litres. The harder question in Uganda is not the annual total. It is how many dry weeks you must bridge.
This page is the supply side: what a catchment yields, how reliable that figure is, and how to get the water off a roof or off the ground and into storage. Where to put it afterwards is covered on farm water storage tanks for manufactured vessels and farm reservoir construction for excavated ones.
The Arithmetic That Turns a Farm Roof Into Litres
Do this calculation before anything else, because it usually changes what people plan to build.
Measure the footprint of the roof, not the sloping surface. This catches almost everybody. The Texas Water Development Board's rainwater manual states it plainly: whatever the pitch, the effective collection surface is the area covered by the roof, measured length by width from eave to eave. A steeply pitched roof has more iron sheet on it than a flat one of the same footprint, and it collects exactly the same rain, because the rain falls vertically. If only one side of the building is guttered, use only the area draining into those gutters.
Then multiply by rainfall in millimetres to get litres, and take off the system losses. The same manual records that most installers assume a collection efficiency of 75 to 90 per cent, and uses 85 per cent in its own worked examples. That allowance covers the first flush you deliberately throw away, water that overshoots the gutter in a heavy storm, evaporation off a hot roof at the start of a shower, and the overflow you lose when the tank is already full.
Now set that against irrigation demand and read the answer honestly. One acre of vegetables applying four millimetres a day drinks about 16,200 litres. A hundred square metre roof's entire annual yield comes to between about seven and eight days of water for a single acre. Roof harvesting will not irrigate a field in Uganda, and any page implying otherwise has not done the multiplication.
What it will do is considerable. Three hundred litres a day, which is what that roof averages, covers a household, or a seedling nursery, or roughly five head of cattle at the middle of published daily intakes, or a greenhouse tunnel for part of the week. Those are the jobs roof harvesting is for, and it does them cheaply because the catchment is already built and paid for.
Runoff Coefficients for Farm Roofs and Bare Soil
Two different numbers get called the same thing, and confusing them is how ground catchment projects get oversold.
For a roof, the figure is a system collection efficiency, and the 75 to 90 per cent band above is reliable because an iron sheet roof sheds almost everything that lands on it. The losses are in the plumbing, not in the surface. One exception worth knowing: the Texas manual notes clay and concrete tiles are porous and may contribute as much as a ten per cent loss through texture, poor flow and evaporation, so a tiled roof sits at the bottom of the band rather than the top.
For the ground, the figure is a runoff coefficient, and it behaves completely differently. FAO's water harvesting manual gives the relationship as runoff in millimetres equals a coefficient times rainfall depth in millimetres, then spends several pages explaining why that coefficient is not a constant. It depends on slope, soil type, vegetation cover, how wet the soil already was, and the intensity and duration of the storm. Two storms of identical total depth produce different runoff if one was twice as intense. For rural catchments the coefficient usually falls between 0.1 and 0.5, which is to say a ground catchment sheds somewhere between a tenth and half of what lands on it, and you cannot know which from a table.
FAO goes further than hedging. It says in terms that runoff coefficients derived for other geographical locations should be avoided for designing a water harvesting scheme, that coefficients for large watersheds should not be applied to small catchments, and that assessment should rest on actual simultaneous measurement of rainfall and runoff in the project area. It cites a recommendation that at least two years be spent measuring before any larger construction programme begins, on the reasoning that a scheme destroyed in its first storm because the design used a borrowed coefficient does lasting damage to the idea.
Three related facts from the same manual make the point concrete. A certain amount of rain always falls before any runoff at all occurs, called the threshold rainfall, and it ranges from about 3 millimetres on sparsely vegetated regular ground to over 12 millimetres where soils are freely draining. Interception on vegetation alone accounts for 1 to 4 millimetres. And the annual or seasonal coefficient is always smaller than the average of coefficients measured on individual storms, because the annual figure includes every shower that produced nothing. So a ground catchment yields less over a season than storm by storm measurements suggest, and that gap is where optimistic designs live.
Sizing Farm Rainwater Harvesting Against Dry Weeks, Not Annual Rainfall
Uganda's annual rainfall is not the constraint. The distribution is.
The World Bank's climate data puts the country's national annual rainfall at 1,283 millimetres on the CRU dataset and 1,577 millimetres on the ERA5 dataset, for the same thirty year period. Both are generous by world standards. But on both datasets the three months of December, January and February deliver only about fourteen per cent of the year's rain in a quarter of the year, and the two dry windows sit either side of the seasons when everything is planted at once.
So the design question is not how much rain your farm gets. It is this: how many consecutive weeks must the stored water last, and how much is being drawn each day over those weeks? Multiply those two and you have your storage volume. The catchment only has to be big enough to fill that volume during the rains, which in Uganda is rarely the binding constraint.
The Texas manual's method is the right one for working this out and it needs no special skill: run a month by month balance where water available equals what was in storage, plus what was collected, minus what was used. Start the balance in the first dry month rather than in January, walk it forward twelve months, and look for the point where storage goes negative. The manual also makes the point that matters here, that storage capacity itself becomes the limiting factor in a wet month, because a tank that is full when the rain comes throws the rest away.
That produces two conclusions people resist. Adding catchment area does nothing once the storage is already full in the rains. And a system that overflows in every storm and runs dry in January is not short of rain, it is short of tank.
Rainfall Across Ugandan Farmland and Why a National Figure Misleads
A national mean is the wrong input for a farm calculation, and it is worth seeing by how much. NASA's long term climatology, taken at points across the country and converted to monthly totals, gives a very different picture in the north from the centre.
| Month | Kampala | Gulu |
|---|---|---|
| January | 65 mm | 13 mm |
| February | 66 mm | 13 mm |
| March | 148 mm | 51 mm |
| April | 194 mm | 120 mm |
| May | 158 mm | 137 mm |
| June | 63 mm | 97 mm |
| July | 57 mm | 102 mm |
| August | 88 mm | 135 mm |
| September | 117 mm | 138 mm |
| October | 158 mm | 152 mm |
| November | 166 mm | 80 mm |
| December | 116 mm | 46 mm |
Read January and February across the two columns. Kampala gets about 131 millimetres over those two months, which is a dry season with rain in it: a roof of any size will collect something most weeks, and storage has to cover days rather than months. Gulu gets about 26 millimetres over the same two months, five times less, which is a dry season with no rain in it at all. A Gulu farm has to store for the whole gap and a Kampala farm does not, and the annual totals of 1,397 and 1,084 millimetres respectively hide that difference completely.
Annual totals at the same five points run from about 723 millimetres at Moroto through 1,084 at Gulu, 1,091 at Mbarara and 1,124 at Kabale to 1,397 at Kampala. That is nearly two to one across one country, and Moroto's December to February total of about 51 millimetres tells you why Karamoja is a different water harvesting problem from anywhere else in Uganda.
Two practical instructions follow. Use the nearest long record you can get to your own farm rather than a national figure, and where you can, put a rain gauge on the farm and start a record, because in five years it will be the most valuable water document you own. And when you do the monthly balance, use the driest plausible year rather than the average, or use FAO's design rainfall idea: pick the rainfall level that is met or exceeded in two years out of three, accept that the system will fall short in the third, and know in advance which year you are planning for.
Gutters and Downpipes on a Farm Rainwater System
The gutter is where most of the missing fifteen per cent goes, and it is the cheapest part of the system to get right.
Size and fall come first. A gutter that overflows in the heaviest storm of the season is losing water exactly when there is most of it, and that storm is the one your annual figure depends on. Run the gutter with a consistent fall towards the outlet, brace it properly along its length so it does not sag into a series of ponds, and set the outer edge slightly lower than the roof line so water sheeting off fast in a downpour drops in rather than shooting over.
Material matters less than fitting, with one safety exception. The Texas manual names half round PVC, vinyl, seamless aluminium and galvanised steel as the common materials, and warns that lead must not be used as gutter solder where the water will be drunk, because slightly acidic rainwater can dissolve it. It also notes copper flashings can discolour fixtures. On a farm system feeding livestock and a household, that lead warning is worth acting on if the guttering is old.
Screen the gutter run. A leaf screen or guard along the length keeps the coarse material out before it reaches the downpipe, and it converts an annual blocked downpipe into an occasional sweep. On a farm with trees over the roof it is not optional.
First Flush Diversion on Farm Rainwater Harvesting
An iron roof accumulates dust, droppings, insects and leaf litter through every dry spell, and the first minutes of a storm wash all of it into your tank unless something stops it. A first flush diverter is a length of pipe that fills before the tank does, holds that dirty water, and lets it drain away afterwards.
The volume to divert is published. The Texas manual recommends one to two gallons for every hundred square feet of collection area, with a rule of thumb of at least ten gallons per thousand square feet, and those two rules work out at the same value. Converted to metric, that is about 0.4 to 0.8 litres for every square metre of roof footprint. So a hundred square metre roof discards roughly 40 to 80 litres at the start of a storm, and a two hundred square metre roof roughly 80 to 160. The manual says the right volume within that range depends on the number of dry days before the rain, the amount and type of debris, tree overhang and the season, so a roof under trees after six dry weeks wants the top of the range.
Ground Catchment Water Harvesting on Farm Soils
When the roofs are all guttered and the volume is still short, the next catchment is the ground. It is far larger and far less reliable, and the reliability problem is the runoff coefficient discussed above.
The surfaces that work are the ones that shed rather than soak: a compacted yard, a road or track, a rock outcrop, a bare crusted slope. FAO notes that crusting, where raindrop impact seals the soil surface, is what allows arid and semi arid areas to produce substantial runoff from short showers, and that soils with high clay or loam content crust most readily while coarse sandy soils barely crust at all.
Which gives the rule that decides whether ground catchment is possible on your farm. FAO names sandy texture as a serious limitation, for a reason worth stating precisely: if the infiltration rate of the soil is higher than the intensity of the rain, no runoff occurs at all. Nothing runs off a sandy field in a gentle shower no matter how much land you have. The comparative infiltration figures in the same manual are about 50 millimetres an hour for sandy soil, 25 for sandy loam, 12.5 for loam and 7.5 for clay loam, and Ugandan rainfall intensities frequently sit below the first of those.
Two further findings shape the layout. Runoff efficiency per unit area rises as the catchment gets smaller, because a short path gives the water less time to soak in or evaporate before it reaches the collection point. And steeper slopes yield more runoff than gentle ones, while runoff falls as the slope gets longer. Together those say: several small catchments beat one large one, and short paths beat long ones. That is the opposite of the intuition that a bigger catchment is better.
Where ground catchment feeds storage, it arrives carrying soil, which is the other cost. A settling stage before the tank or the filter is not optional on ground collected water, and on an excavated reservoir the silt becomes the maintenance problem set out on farm reservoir construction.
Contour Ridges and Bunds: Water Harvesting in the Crop Field
The third form of harvesting does not store water in a vessel at all. It stores it in the soil profile, by stopping runoff where it falls and giving it time to sink in. FAO's water harvesting manual documents the family in detail, and the three that matter on a farm are these.
Three rules apply across all of them. Bund tops should be at least 25 centimetres wide with side slopes of at least one to one, and grass cover is the best erosion protection you can give them. A diversion ditch above the block, aligned at a 0.25 per cent gradient and typically 50 centimetres deep and 1 to 1.5 metres wide, protects the whole system from outside runoff, and FAO insists it is built first so a storm during construction does not destroy the work. And the uncultivated catchment strip between ridges has to be kept clear of vegetation, because vegetation is what stops runoff being generated.
For labour planning, FAO gives earthworks rates of 1.0 to 3.0 cubic metres per person per day, and notes larger structures can take up to fifty per cent more labour per unit volume than small ones because of the extra earth moving. One agronomic point that surprises people: FAO recommends reducing the cereal plant population to about 65 per cent of the conventional rainfed standard in a contour ridge system, so that the fewer plants have more moisture each in a poor year.
Where Contour Harvesting Does Not Suit Ugandan Farmland
This is the section most pages on this subject leave out, and it is the most important one for a Ugandan reader.
FAO specifies contour ridges for crops as suitable in a rainfall range of 350 to 750 millimetres a year. That is a semi arid specification, and it has to be read against Ugandan rainfall. Of the five points measured above, only Moroto, at about 723 millimetres, falls inside that band. Kampala at about 1,397 millimetres is nearly double the top of it. Gulu, Mbarara and Kabale all sit between 1,084 and 1,124, comfortably above it.
What that means in practice is that the FAO contour toolkit was written for conditions that exist in Karamoja and in the drier parts of the cattle corridor, and that do not exist across most of Uganda. In a district receiving 1,100 to 1,400 millimetres a year, concentrating additional runoff into a crop furrow risks waterlogging rather than relieving drought, and the manual's own soil guidance says as much: a very low infiltration rate can be detrimental because of waterlogging in the cultivated area, and the requirements of the cultivated area always take precedence over those of the catchment.
Two more limits from the same source. Water harvesting is not recommended where slopes exceed 5 per cent, because runoff distributes unevenly and the earthworks stop being economic. And the ground must be even: FAO says areas with rills or undulations should be avoided, and that contour bunds are unsuitable on uneven or eroded land because excess water overtops and breaches them at the low spots.
So the honest position for most Ugandan farms is that roof harvesting into storage is the technique that pays, that ground catchment into a reservoir is worth considering where the soil sheds and the site suits, and that FAO's contour field techniques belong in the dry north and east rather than being recommended nationally. Where they do fit, they are cheap and effective. Where they do not, they solve a problem the farm does not have. The wider irrigation picture, including where harvested water sits against pumped supply, is on irrigation farming in Uganda.
Rainwater Harvesting for Farms: Questions Farmers Ask
What does a farm rainwater harvesting system cost in Uganda? The catchment is free because the roof already exists, so almost all the money is in guttering, downpipes, the diverter and above all the storage. Storage dominates the bill, and cost per litre falls as tank size rises, so the volume decision made in the monthly balance is also the budget decision. Get separate quotes for guttering by the metre and for the tank by volume, because bundled quotes hide which one was trimmed. Where this sits in overall irrigation spend is on irrigation system cost per acre.
How much rain can I collect from my roof? Footprint in square metres times annual rainfall in millimetres times 0.85 gives litres a year. Use your nearest long rainfall record rather than a national mean, and use the footprint rather than the sloping area.
Can rainwater harvesting irrigate an acre? No. A hundred square metre roof collects roughly a week's water for one acre over a whole year. Harvested rain irrigates nurseries, greenhouses, seedlings and small high value plots, waters livestock, and bridges dry spells inside the rains. For field scale you need a pumped source, covered on boreholes for agricultural irrigation, or excavated storage.
What size tank do I need? Not what the roof can fill, which is the wrong question. Take your daily draw, multiply by the number of consecutive dry days you must cover, add about a quarter, and check with a month by month balance that the tank both fills in the rains and lasts through the gap. Sizing detail is on farm water storage tanks.
Is water off an iron sheet roof safe to drink? It can be, with screening, a properly sized and self draining first flush diverter, an opaque covered tank and treatment appropriate to the use. What makes it unsafe is skipping those. If the guttering is old, check whether lead solder was used, because slightly acidic rainwater dissolves it. Confirm anything to do with drinking water quality with your district water office rather than with an article.
Why is my tank empty in January when my roof is large? Because the roof is not the constraint, the tank is. A large roof filling a small tank throws away most of the wet season in overflow and then runs out. Measure how often it overflows; if it does so in most storms, the next money goes on storage, not on more guttering.
Should I harvest from the ground as well? Only after the roofs are fully guttered, because roof water is cleaner, more predictable and already paid for. Ground catchment is worth it where the soil sheds water rather than soaking it, where there is a slope to work with, and where you have somewhere silt tolerant to put the result.
Do contour bunds work in Uganda? In the dry north and east, yes, and cheaply. In districts getting well over a thousand millimetres a year, the techniques were specified for a drier rainfall band and concentrating more runoff onto a crop can waterlog it. Judge it on your own district's rainfall and your own soil's drainage, not on a general recommendation.
How do I know my own rainfall? Put a gauge on the farm, read it at the same time each day, and write it down. It costs almost nothing, it settles every argument on this page for your specific site, and after a few seasons it tells you your dry spell lengths directly instead of by inference from a national average.
Start with the multiplication: measure your roof footprint, find the best local rainfall record you can, and work out what you are already losing to the ground every wet season. Then take the daily draw and the dry weeks you need to cover to a supplier and ask for guttering and storage quoted separately. Guttering and tank sellers are reachable through where to buy irrigation equipment, and the crops and structures the collected water will supply run through the irrigation and greenhouses guide. For the pump that moves it once it is stored, see choosing an irrigation pump.
