One of the biggest climate threats to cocoa may not be heat alone, or even drought, but heavy rainfall arriving at the wrong stage of the tree’s development, according to a new study published in PNAS.
The study found that intense rainfall during the wet season is strongly linked to cocoa yield losses in major tropical producing regions. The finding is significant because cocoa is not only the raw material used to produce chocolate. It is the main source of income for nearly six million smallholder farmers worldwide and supports tens of millions of people whose livelihoods depend on cocoa farming, trade, and processing.
According to Anna Lea Albright of the Department of Earth and Planetary Sciences at Harvard University, the study’s lead author, the research shifts attention away from average climate conditions and towards short, intense weather events. In other words, it is not only how much rain falls in a year that matters, but whether heavy rainfall hits cocoa trees during sensitive stages such as flowering and early pod development.
The researchers began with Ghana, the world’s second-largest cocoa producer after Côte d’Ivoire. They used newly available district-level production records from the Ghana Cocoa Board covering 66 cocoa-producing districts from the 2000/2001 to 2022/2023 seasons. They also estimated the 2023/2024 production anomaly using national data, following a season in which West Africa suffered major harvest losses.
The team then matched these production records with daily rainfall and temperature data. The pattern was striking. In many years, cocoa production rose or fell across much of Ghana at the same time, suggesting that a broad driver, such as regional weather, was affecting yields rather than only local problems in individual districts.
The study found that two rainfall signals together explained about 68% of year-to-year cocoa yield variability in Ghana: excessive rainfall during the main wet season from April to June and drier conditions during the dry season from November to February.
This presents a more complex picture than simply saying that cocoa needs rain. Cocoa trees do require moist conditions, but excessive rainfall at the wrong time can be damaging. In Ghana, the April-June wet season overlaps with flowering and the formation of young pods, which later contribute to the main harvest.
Heavy rain can physically damage flowers and young pods, interfere with pollination, and create favourable conditions for fungal diseases such as black pod disease. Rain splash can also help spread disease from infected plant material or soil to healthy pods.
The researchers then tested whether the same pattern appeared beyond West Africa. Using national cocoa yield data from Ecuador and Indonesia, they found a consistent negative relationship between heavy wet-season rainfall and cocoa yields. The relationship was clearer in Ecuador, while the signal in Indonesia was weaker, partly because cocoa production there has also been affected by non-climatic changes, including farmers shifting from cocoa to other crops such as oil palm and cloves.
Albright and her colleagues argue that the findings could help improve early warning systems for cocoa farmers. Heavy rainfall extremes are partly linked to large-scale climate patterns such as El Niño and changes in sea-surface temperatures. If seasonal forecasts can identify periods when damaging rainfall is more likely, farmers and agricultural advisers may be able to act earlier by improving disease monitoring, adjusting fungicide timing, managing shade and tree density, and protecting vulnerable flowers and pods.
The study also has implications for climate change. Warmer air can hold more moisture, and climate models generally project that heavy rainfall events will intensify in many regions as the planet warms. The study suggests that this could become an increasingly important risk for cocoa, particularly in West Africa, where the analysis found signs of increasing heavy rainfall over recent decades.
However, the authors are careful to note the study’s limitations. In Ghana, they used district-level production data as a proxy for yield because reliable annual harvested-area data were not available at the same level of detail.
The rainfall results were also sensitive to the dataset used. The TAMSAT dataset showed a clearer heavy-rainfall signal than CHIRPS or PERSIANN. In Ecuador and Indonesia, the analysis relied on national-level data, which cannot capture local differences within farming regions.
The study also does not rule out non-climatic pressures, including ageing trees, mining, smuggling, disease, low pollination, changing prices, and farm management practices.
