One thousand millimeters of rain per year on a hectare equals ten thousand cubic meters of water. It’s worth writing out the full number because it changes the perception of the problem, as ten million liters per hectare is a figure that most farms never consider as a available resource, but rather as something that passes overhead and is lost. A good portion of that water must indeed leave, as it feeds aquifers, streams, and ecosystems that do not belong to anyone in particular. However, the fraction that drains from roofs, patios, internal roads, and slopes, as well as the water that exits your processes after being used once, is water that is already within your perimeter and is often lost without any management decisions.
This season focused on the water entering the irrigation system, calculating how much is needed, how much reaches the roots, when it is advisable to restrict it, and how to measure what is available. It’s time to scale up. Up to this point, we looked at the plot; now we need to look at the entire farm, because the decisions that remain to be made are not about irrigation scheduling but about infrastructure and resource accounting, and these determine whether you have water to irrigate in the next dry period or just a good irrigation system with nothing to supply it.
The Farm as a Hydric System: The Accounting That Almost No One Keeps
There is a formal discipline for this called water accounting. The FAO defines it as the systematic recording of water flows in a system, including inputs, outputs, consumption, and storage, with the aim of understanding where the resource is and what real margin of maneuver exists. The usual scale of application is the watershed or irrigation district, but the logic applies equally to a farm, and that is where it is almost never used.
The exercise is simpler than it sounds. The inputs to your farm are the rain on your surface, plus what you capture from a surface or underground source, plus any water you receive from a district. The outputs are the evapotranspiration of the crop, which is the only consumption that produces value, plus the runoff that leaves your property, plus deep percolation, plus the process waters that you discharge. The storage is what you have retained in the soil, in tanks, and in reservoirs. When you put approximate numbers in each box, two things often surprise: the magnitude of the rain that leaves the farm without being used, and the amount of process water that is discharged after a single use.
This diagnosis is the basis for everything that follows, as it defines where it is worth investing. A farm that loses its water mainly through surface runoff has a problem with capture and storage; one that loses it through percolation has a problem with irrigation scheduling that we already know how to solve; and one that discharges large volumes of process water has a reuse problem. All three look the same from the outside, and all three are resolved differently.
Harvesting What Already Falls: Roofs, Runoff, and the Right Size of the Reservoir
Rainwater harvesting has two pathways, and it’s important not to confuse them because their costs and water quality are very different.
The first is capturing water from hard surfaces, mainly roofs of warehouses, processing plants, sheds, and homes. The calculation is straightforward: multiply the roof area in horizontal projection by the annual rainfall in millimeters and by a runoff coefficient that represents losses due to evaporation, initial wetting, and diversion of the first wash. This coefficient is around 0.8 to 0.9 for metal roofs and slightly less for tiles and other rough materials. A roof of three hundred square meters in an area with one thousand five hundred millimeters of annual rainfall and a coefficient of 0.85 yields about three hundred eighty cubic meters per year. In farm terms, this is water of excellent quality, with very low sediment load, available without pumping from the source and at a capture cost reduced to gutters, a first flush diverter, and storage.
The second pathway is capturing surface runoff, which collects the water that flows over the land, internal roads, and drains into a reservoir or infiltration lagoon. It contributes much larger volumes than roofs, but the water arrives laden with sediments and requires sedimentation and filtration structures before entering a localized irrigation system. The FAO and the International Fund for Agricultural Development reference materials for small scale describe the typologies, selection criteria, and design elements of these systems, and agree that the selection should start from the local rainfall regime and the intended use of the water, not from the size of the farm.
The point where design often goes wrong is in sizing. The temptation is to calculate the reservoir to store as much annual rainfall as possible, and that is the fastest route to an oversized structure that never fills and is not cost-effective. The correct criterion is to size against the deficit of the critical window, that is, to calculate how many millimeters your crop lacks during the weeks of the dry period when the deficit truly harms yield, convert that to cubic meters based on the area you want to protect, and build to cover that volume. Research on optimization schemes that combine water harvesting with hydrological balance simulation in small rain-fed systems shows precisely that the optimal storage volume arises from the interaction between rainfall distribution and crop demand at the time of year when it is needed. As a scale reference, the FAO documents that infiltration ponds with capacities ranging from eight thousand five hundred to fourteen thousand cubic meters can expect two or three seasonal fills per year, with enough volume to irrigate between four and six hectares of rain-fed crops.
Recent work has also evaluated flexible storage systems, with modular structures that allow capturing excesses from extreme rainfall events and shifting rain-fed production to production with supplemental irrigation, an approach that is interesting for farms that cannot undertake a permanent civil work.
Storing Without Losing: The Evaporation That Takes Away What You Captured
Here appears a cost that almost no design accounts for, which is that an open reservoir loses water every day from its surface. In warm climates with high atmospheric demand, the evaporated layer during a prolonged dry period can represent a very significant fraction of the stored volume, precisely at the time when that volume is most valuable.
Research published in 2024 on evaporation loss from small agricultural reservoirs in a warming climate argues that this component has been systematically underestimated in water accounting, and that its relative weight increases as temperatures rise. Read it in terms of the farm, because it means that the reservoir you sized with criteria from twenty years ago may be delivering less useful water than its nominal volume suggests.
Reduction measures exist, and their cost is reasonable compared to what they protect. A study that evaluated shade net covers over agricultural ponds, calibrated with field data from 2024 and validated with measurements from 2025, found that a fifty percent shade net reduced the water surface temperature compared to the uncovered pond and achieved an average evaporation suppression of twenty percent, equivalent to about two and a half cubic meters per day of reduced loss and an accumulated saving of close to one hundred ten cubic meters during four months of monitoring. That saved volume is not theoretical; it is water that remains available just at the time of year when there is no other source.
In addition to the cover, the geometry of the reservoir matters more than it seems. For the same volume, a deeper vessel with less exposed surface loses less to evaporation than a shallow, spread-out one, which makes the shape of the storage a design decision rather than just a consequence of the available land. Waterproofing, on the other hand, addresses the other loss, infiltration through the bottom and walls, which in coarse-textured soils can easily exceed evaporation.
The Water That Is Already Inside Your Farm
Do you have any idea how many cubic meters of rain fell on your farm last year, and how many of them you managed to use? Most producers can accurately answer how much fertilizer they bought and have no estimate of this figure, even though water is the most limiting input in the coming years. Harvis by TECDE helps you keep track of rainfall, irrigation, and consumption by plot, build the water balance of your entire farm, and size with your own data how much storage you really need to cover your critical window. Message us on WhatsApp, and we’ll do that accounting together.
Using Water Twice: Drainage and Process Waters
The second half of closing the cycle consists of reusing water that has already passed through the system. There are two main sources, and their treatment is different.
The first is agricultural drainage water. In areas where irrigation water is scarce, reusing drainage is a recognized strategy to complement the available resource, and it has the additional benefit of reducing the volume that needs to be disposed of. Its central limitation is salinity, because the water that passes through the profile drags salts and exits more concentrated than it entered, so reusing it without control leads to progressive accumulation in the soil. Responsible management requires measuring electrical conductivity before reuse, mixing with better quality water to dilute, directing the more saline water to more tolerant crops or stages, and maintaining a washing fraction that allows salts to be removed from the profile.
The second source, and the most relevant for coffee, is process water from wet processing. The literature on wet processing reports consumptions that, depending on the technology used, range from very low values in modern low-consumption equipment to orders of fifty to two hundred liters of water per kilogram of green coffee in traditional facilities, with depulping, washing, and fermentation concentrating practically all of that consumption. This water exits loaded with pectins, sugars, and organic compounds, with very high chemical and biochemical oxygen demand and low pH, and discharging it untreated severely affects the quality of receiving sources, an effect widely documented in studies on water bodies receiving effluents from multiple processing plants.
The interesting thing is that the available solutions are accessible and low-tech. Research from 2025 on reducing carbon and water footprints in wet coffee processing evaluated a set of measures that includes more efficient fermentation methods, reuse of treated water within the process itself, biogas capture, and automation of processing stages, and documents that the integration of constructed wetlands has the potential to reduce the chemical and biochemical oxygen demand load by around eighty percent. Analyses of constructed vertical up-flow wetlands in tropical conditions, with selected macrophyte species, confirm that this is a robust technology, with low operational costs and well adapted to tropical temperatures, where biodegradation occurs at a good pace.
In farm terms, a processing plant that recirculates washing water within the same processing lot and treats the effluent before discharging it simultaneously reduces the volume it captures from the source, the regulatory risk of the discharge, and the impact on the stream from which it likely also depends for irrigation.
The Savings That Are Not Always Savings, and Other Limits of Reuse
There is a concept that is worth understanding well because it avoids costly enthusiasm. The FAO distinguishes between apparent savings and real water savings, noting that the former records reductions in extractions without considering changes in consumption, while the latter records reductions in effective consumption and in non-recoverable return flows. The practical consequence is counterintuitive and worth stating plainly, because part of what was perceived as loss on the farm, deep percolation and runoff leaving the property, actually returned to the system and was being used downstream by someone else, or recharged the aquifer from which the farm itself extracts. Capturing all that flow does not create new water; it redistributes it.
This does not invalidate water harvesting or reuse, but it does change the way to justify them. The solid argument is not that new water is generated, but that control is gained over when the water is available, shifting excesses from the wet season to the dry window where the crop does not forgive. That temporal shift is the real value, and it is enormous.
The other limits are concrete. Process water and gray waters are not suitable for application on edible parts of a crop without treatment that guarantees sanitary quality, and responsibility towards an export market or certification does not allow for improvisation at this point. The salinity accumulated from reusing drainage is a slow damage that is detected when it is already expensive to reverse, and requires monitoring of conductivity in water and soil. Relevant-sized reservoirs and runoff capture are often subject to permits and environmental regulation, and building first to ask later is a route with difficult-to-anticipate costs. And there is an honest scale restriction, because in small farms with well-distributed rainfall throughout the year, the investment in storage may simply not pay off, and the same money yields better by improving the uniformity of the existing irrigation system.
How to Start Without Major Works
The practical version of all this does not start with machinery. It starts with a rain gauge and a notebook, because you need the actual rainfall of your farm and not the regional average, and just one year of records is enough to have a much more reliable working base than any external estimate.
The second step is to identify your critical window, that is, the weeks of the year when the water deficit coincides with a phenological stage that does not tolerate restriction, and calculate the volume you specifically lack during those weeks for the area you want to protect. That number, and not the annual rainfall volume, is what sizes any investment.
The third step is to inventory your already constructed catchment surfaces, adding square meters of existing roofs, because that water is of the best quality and the cheapest to capture, and in many farms it is enough to cover uses that are currently being met with water from the main source. The fourth is to start with a small module, with enough storage for part of the critical window and for a manageable area, measure during a season how much it filled, how much evaporated, and how much you actually used, and only then decide whether to scale up. It’s the same prudence of the pilot plot that we apply to deficit irrigation, transferred to infrastructure.
As we already addressed in our blog Between Rain and Drought, which you can read here, the distribution of rainfall has stopped behaving as it did in historical averages, and that is precisely why storage has ceased to be an optional work to become a tool for stability; when the same amount of annual water arrives concentrated in fewer events and with longer dry periods between them, the ability to store is what turns intense rainfall into a resource instead of an erosion problem.
When Water Stops Being Something You Expect from the Sky
This season started with an account and ends with infrastructure. We calculated how much water your crop demands, reviewed what fraction effectively reaches the root, saw that there are windows where deliberate restriction improves the margin, and learned to measure in real-time what the soil has and what the plant feels. All of this manages the water that enters your system. Closing the cycle is what decides how much water there will be to manage.
The difference between a farm that suffers every dry period and one that goes through it without shocks rarely lies in how much rain each receives, because they usually receive practically the same. It lies in how much of that rain managed to stay within the property, in what conditions it was stored, and how many times it was used before leaving. Start by measuring your rain, calculate the deficit of your critical window, look at the roofs you already have built, and treat the water you are currently discharging. None of those four things require major works, and together they change the fundamental question, because you stop asking when it will rain and start asking how much water you have.
Consulted References
- Food and Agriculture Organization of the United Nations (FAO). (n.d.). Water Accounting and Auditing. https://www.fao.org/land-water/water/water-management/water-accounting/en/
- Food and Agriculture Organization of the United Nations (FAO). (n.d.). Rainwater Harvesting. https://www.fao.org/4/i1861e/i1861e06.pdf
- International Fund for Agricultural Development (IFAD). (n.d.). Water harvesting systems for smallholders, tips for selection and design. https://www.ifad.org/documents/48415603/49788193/waterharvestingsystems_e.pdf
- Food and Agriculture Organization of the United Nations (FAO). (n.d.). Compendium on Rainwater Harvesting for Agriculture in the Caribbean Sub-region. https://openknowledge.fao.org/server/api/core/bitstreams/9e1e6970-26d5-4d07-94ce-7202552e4988/content
- Rainwater harvesting and water balance simulation-optimization scheme to plan sustainable second crop in small rain-fed systems. (n.d.). Science of the Total Environment / ScienceDirect. https://www.sciencedirect.com/science/article/abs/pii/S030147972201708X
- Food and Agriculture Organization of the United Nations (FAO). (n.d.). Water Harvesting and Storage, Land and Water. https://www.fao.org/land-water/water/water-management/water-storage/en/
- Rainwater, Harvesting and Storage through a Flexible Storage System to Enhance Agricultural Resilience. (n.d.). Agriculture, 13(12), 2289. MDPI. https://doi.org/10.3390/agriculture13122289
- Aminzadeh, M., et al. (2024). Evaporation Loss From Small Agricultural Reservoirs in a Warming Climate, An Overlooked Component of Water Accounting. Earths Future, Wiley. https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2023EF004050
- Evaluating evaporation losses from agricultural ponds with shelter net covers by integrated Dalton model. (n.d.). Agricultural Water Management / ScienceDirect. https://www.sciencedirect.com/science/article/pii/S0378377426001186
- Food and Agriculture Organization of the United Nations (FAO). (n.d.). Chapter 6, Drainage water reuse. https://www.fao.org/4/y4263e/y4263e09.htm
- Reduction of carbon and water footprints in wet coffee processing and optimization of wastewater treatment at the wet mill factory. (2025). ScienceDirect. https://www.sciencedirect.com/science/article/pii/S2666790825000904
- Waste Water Management in Wet Coffee Processing Mills and their Impact on Water Quality Status. (n.d.). PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC11181887/
- Uncovering the Effect of Multiple Wet Coffee Processing Industries Effluents on River Water Quality Parameters. (2025). Taylor & Francis Online. https://www.tandfonline.com/doi/full/10.1080/15275922.2025.2490477
- Wastewater treatment in tropical, vertical up-flow constructed wetlands with selected macrophytes species. (n.d.). PMC. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11714395/
- Food and Agriculture Organization of the United Nations (FAO). (n.d.). Real Water Savings in Agriculture. https://www.fao.org/platforms/water-scarcity/Knowledge/knowledge-products/detail/real-water-savings-in-agriculture/en