Brazilian coffee is predominantly rainfed, with supplementary irrigation used to support flowering and bean filling where rainfall is unreliable. The volume of water applied per hectare therefore varies according to where a farm sits on the rainfed-to-irrigated spectrum – and, critically, the demands of the growing season.
Monitoring water use is essential because irrigation can compete directly with domestic and ecological demand in catchments experiencing periodic water stress, while also carrying associated energy and emissions costs.
Average water application fell sharply from 3,338 m³/ha in 2024 to 1,165 m³/ha in 2025, while the median increased from 368 m³/ha to 843m³/ha.
The season provides a strong explanation for this pattern. 2024 was an exceptional drought year across Brazil's coffee belt – the country's most intense and widespread drought on record according to USDA FAS, and the driest conditions since 1981, according to the national disaster monitoring centre, Cemaden. Soil moisture monitoring by Fundação Procafé in Minas Gerais recorded a water deficit of around 300 mm at the end of the 2024 dry season, compared with roughly 110 mm the year before. Under these conditions, farms with full irrigation infrastructure could run their systems intensively to protect flowering and yield, helping to explain the unusually high 2024 average.
The rise in the median tells the other half of the story. It may seem counterintuitive that the typical farm applied less water in the record drought year, but irrigation is capacity-constrained. Only around 30% of Brazilian coffee farms are equipped with irrigation, so in 2024 the typical farm could not respond to the drought, regardless of need. Without sufficient irrigation equipment and sufficient water availability, many farms applied little water – roughly 37 mm – and absorbed the resulting yield impact, while the minority with full irrigation systems applied substantially more and drove the exceptional average.
The drought changed that dynamic. Its most lasting effect was to spur investment in on-farm water capacity – visible directly in the water sources data below, where rainwater harvesting more than tripled between 2024 and 2025. Harvested and stored rainwater is precisely what allows a previously rainfed farm to apply supplementary irrigation when water is deficient. Industry reporting likewise suggests that after 2024, Brazilian growers began running supplementary irrigation earlier and more routinely as a precaution rather than waiting for visible crop stress. With 2025 rainfall still irregular and below expectations in parts of Minas Gerais – the state's harvest came in around 11% below the previous cycle – newly equipped farms had both the means and the reason to irrigate moderately, lifting the median to roughly 84 mm.
The two movements are therefore the same story seen from opposite ends of the distribution. Conditions short of 2024's record deficit meant that intensive irrigators applied substantially less water, pulling the average down, while the typical farm moved from near-total dependence on rainfall to modest, planned supplementary irrigation, lifting the median. This points to a shift towards resilience rather than greater water intensity: moderate supplementary applications, increasingly supported by harvested rainwater, leave farms less exposed to the next drought than either intensive pumping or unbuffered reliance on rainfall. The values themselves are agronomically plausible for predominantly rainfed systems, while the near-zero minimums are consistent with fully rainfed farms, which still account for the majority of Brazilian coffee production.
Given the dominance of the seasonal signal, we present these figures as a monitoring baseline rather than an efficiency claim in either direction. What they establish is that the programme now holds farm-level data on water application in coffee production, creating a baseline against which genuine changes in irrigation efficiency can be measured across future seasons.
Volume alone is an incomplete measure of water impact: where water comes from matters because it determines whether abstraction is sustainable.
Groundwater drawn from boreholes can be abstracted faster than aquifers recharge, with depletion effectively irreversible on a management timescale. Surface water is renewable, but it is shared with downstream users and ecosystems.
Harvested rainwater, by contrast, imposes no additional pressure on any shared resources by capturing water that would otherwise run off, and buffering farms against increasingly variable precipitation. Moving towards more resilient water sources can therefore reduce water risk even where total irrigation volume remains unchanged.
Farms commonly draw on more than one water source, so the figures represent the percentage of farms reporting each source and therefore do not sum to 100.
The shift towards lower-impact sources is the largest movement in this report. Rainwater harvesting more than tripled, rising from 18.4% to 60.4% of farms (+42.0 percentage points), while reliance on bore and groundwater – the most depletion-sensitive category – fell from 66.3% to 41.5% (−24.8 percentage points). Surface-water use also increased substantially (+22.7 percentage points), consistent with a shift by some farms away from groundwater towards a renewable source.
This pattern indicates that farms are actively restructuring how they source water, rather than simply reallocating use between equivalent options. Rainwater harvesting requires capital investment in catchment and storage infrastructure, so a change of this scale reflects deliberate on-farm investment – a rational response to the drought conditions of 2024 and precisely the kind of resilience-building behaviour the programme is designed to encourage.
USDA FAS Coffee Annual, Brazil
Coffee Intelligence — Brazil's coffee irrigation conundrum
Tea & Coffee Trade Journal — Extreme drought and coffee
Borgen Project — Brazil drought impacts on coffee/orange
StoneX Brazil crop survey 2026/27
DTN — Erratic weather, Brazil coffee 2025
