Cacao and climate

The headline is that chocolate will run out by 2050. No study read here says that. What the evidence shows is land becoming less suitable in some places and more in others, a tree that suffers from dry air and heavy rain as well as heat, and real uncertainty about how much shade or breeding can help.

7 questionsModels labelled as modelsDisagreements kept

Two kinds of evidence

Modelled: A projection of where the climate cocoa grows in today will be found later, under a stated emissions scenario. It is not a forecast of harvests and takes no account of prices, breeding, irrigation or disease.
Measured: What happened to real trees: in a drought, under a rain shelter, in a growth chamber. It is certain about one place and time and says little about others.

Will cocoa disappear by 2050?

Nothing read for this page says so. Here is what the sources that mention 2050 do say.

The Intergovernmental Panel on Climate Change says of cocoa and coffee that both crops “are at risk from climate change impacts by 2050”. At risk is its wording. (source 1)
In its chapter on Africa the same assessment gives low confidence to projected impacts on West African cocoa, because for most crop and country combinations there is a single study and limited evidence on the direction of change. (source 2)
The West African projection most often cited finds vulnerability varying sharply inside the cocoa belt: greatest near the boundary between forest and savanna in Nigeria and eastern Côte d'Ivoire, least in the southern parts of Cameroon, Ghana, Côte d'Ivoire and Liberia. (Schroth et al. 2016)
A later model that added soil to climate projects up to 6% more suitable area for cocoa in West Africa by 2060, not less. (Ariza-Salamanca et al. 2023)
The paper usually named as the origin of the claim could not be retrieved, so the trace back to a source is incomplete.

Not shown: That cocoa is safe. Several projections show large losses in particular regions, and one measured drought killed a sixth of the trees on the farms sampled.

Not shown: Where the “2050” headline first appeared.

What the suitability models project, and how they differ

Each of these is a model of where the climate cocoa now grows in will exist in future. They differ in region, in scenario, in horizon and in what else they include.

West Africa, to the 2050s: maximum dry-season temperature is projected to become as limiting as dry-season water, or more so. (Schroth et al. 2016)
The same group turned its projection into three zones for the 2030s and 2050s: little affected, where cocoa could intensify or expand; moderately affected, needing diversification and changed practice; and severely affected, needing a gradual change of crop. It judges that local losses could be made up elsewhere in the region. (Schroth et al. 2017)
West Africa, to 2060, with soil included: a model projects up to 6% more suitable area; 14.5% less once land that would require deforestation is excluded; and half of 37 shade tree species losing range. (Ariza-Salamanca et al. 2023)
Brazilian Amazon, to 2050: the areas suited to intensification and expansion are projected to shrink by 37% under a middle scenario and 73% under a high one. (Igawa et al. 2022)
Mesoamerica: cocoa is projected to become a possible replacement in most areas where coffee becomes vulnerable, while the shade trees farmers now prefer are themselves highly vulnerable. (de Sousa et al. 2019)

Not shown: Future production. Suitable climate is one condition for cocoa; none of these models includes prices, labour, disease, breeding or irrigation.

Not shown: Anything for Asia or the Pacific. No projection for either was read.

Not shown: A consensus. Two models of West Africa reach opposite signs.

How the plant responds: temperature, dry air, carbon dioxide

Water stress first shows in the leaf: stomata begin to close at a leaf water potential of about −1.5 MPa and conductance falls as the air dries, and flowering is inhibited. The same review found few reliable field data on how yield responds to drought or to irrigation. (Carr and Lockwood 2011)
Dry air is a stress of its own. Through one harmattan season in Ghana it reduced leaf area by a third, as much as dry soil did. (Della Sala et al. 2021)
Heat: in seedlings a daily maximum around 39°C against 36°C severely reduced photosynthesis and growth. In chambers kept humid, warming above 31°C by day raised photosynthesis and still reduced growth and leaf area, more in one genotype than another. (Hebbar et al. 2020; Mateus-Rodríguez et al. 2023)
Carbon dioxide works the other way: raising it increased photosynthesis by 27% in seedlings and improved water use, and partly offset heat and drought. (Hebbar et al. 2020; Mateus-Rodríguez et al. 2023)
Genotypes differ. The temperature below which stem growth stops ranged from 18.6 to 20.8°C across four types. (Daymond and Hadley 2004)

Not shown: How mature trees in the field respond to heat or to more carbon dioxide. These are seedlings and young plants in chambers and glasshouses.

Not shown: Yield. None of the chamber experiments ran to pods.

What weather has been seen to do to yields

Drought can be catastrophic locally. After the 2015 to 2016 El Niño, sampled farms in Bahia had lost 15% of their trees and 89% of their yield, and witches' broom infection had risen. (Gateau-Rey et al. 2018)
On one Bahia farm, which trees died in drought years followed the physical properties of the subsoil, the direction the slope faced and the soil type. (Souza Júnior et al. 2023)
Too much rain may matter more than too much heat. In district records from Ghana, excess wet-season rain and dry-season drought together explained 68% of the year-to-year variation in yield, and national data from Ecuador and Indonesia showed the same negative response to heavy rain. (Albright et al. 2026)
Delay matters. Excluding rain from an Indonesian plot for thirteen months changed little at the time and cut the next main harvest by 45%. (Schwendenmann et al. 2010)
Along Ghana's climate gradient, dry-zone farms averaged 288 kg per hectare against 712 and 849 in the wetter zones. (Abdulai et al. 2018)
Satellite measures of plant growth through eight El Niño and La Niña events suggest lowland cacao zones of Latin America are more exposed than mountain coffee zones. (González-González et al. 2025)

Not shown: A yield loss per degree of warming. No study read gives one.

Not shown: That the Bahia losses are typical of an El Niño year. They are one region in one event, and the abstract does not say how many farms.

Does shade protect cocoa from drought?

Not reliably. This is the sharpest disagreement on the page.

Against: at the dry edge of Ghana's cocoa belt, during the drought of 2015 to 2016, all the cocoa under one shade species died and 77% died under another, while cocoa in full sun kept transpiring and the soil under it stayed wetter. (Abdulai et al. 2018)
For: in a 33-month experiment elsewhere in Ghana shade increased soil moisture and bean yield. It did not offset reduced rain, and fruit loss and pod damage were also higher under shade. (Mensah et al. 2023)
For, in a wetter place: with 71% of rain excluded in Sulawesi, cacao and its shade tree drew water from different depths and competition stayed low or favoured the cacao. (Köhler et al. 2010; Schwendenmann et al. 2010)
The international assessment records both: shaded cocoa may be more resilient in some situations and full-sun cocoa in others. It gives high confidence to shade systems providing other services, among them pest regulation, soil fertility, biodiversity and carbon storage. (source 1)

Not shown: A rule. The outcome depended on the shade species, the site's rainfall and how severe the drought was.

Not shown: Which shade species to plant. Two species were compared in one study; the Indonesian work used one.

Flowering, pollinators, pests and disease

Pollinators follow the rain. On Caribbean farms, pollinating midges were very scarce in drought and their numbers tracked the previous month's rainfall; temperature and humidity did not predict them. (Arnold et al. 2018)
Flowering is inhibited by water stress. (Carr and Lockwood 2011)
Frosty pod rot followed wet conditions and peaked at a daily maximum of 30°C on a susceptible clone. (Leandro-Muñoz et al. 2017)
Mirid damage in Cameroon was related to relative humidity alone among the weather variables recorded, and fell as humidity rose. (Mahot et al. 2024)
Drought raised the rate of witches' broom infection in Bahia. (Gateau-Rey et al. 2018)
Climate and black pod, and climate and swollen shoot virus: no study was reached.

Not shown: How flowering, pollination and fruit set respond together to heat or drought. No study measured all three.

Not shown: Whether any disease will become more or less damaging as climate changes. The studies describe present weather.

Adaptation: what there is to work with

There is genetic variation to select from. Screening 36 genotypes as seedlings separated drought-tolerant from intolerant types, and in 79 cacao genomes most signs of adaptation were local to single populations and tied to environment. (Santos et al. 2014; Nelson et al. 2021)
Breeding has begun and has not reached yield: hybrids of four tolerant parents were ranked by leaf chemistry under water stress. (Juby et al. 2021)
Moving and diversifying are the other options the models suggest: one regional plan pairs intensification where climate stays suitable with a gradual change of crop where it does not. (Schroth et al. 2017)
Farmers already adapt by spreading risk: those in Ghana's dry zone relied far less on cocoa than those in wetter zones. (Abdulai et al. 2018)
Irrigation trials with measured yields, and any released drought-tolerant variety tested on farms: none was reached.

Not shown: That any drought-tolerant cacao exists as a planted variety. The work read is at the stage of seedlings and markers.

Not shown: What adaptation costs a farmer.

Sources

Studies are cited by author and year and summarised, with their limits, on the research pages. The official documents are listed here. Each was downloaded and read, and the passage relied on was checked to be in it.

Intergovernmental Panel on Climate Change. IPCC AR6 Working Group II, Chapter 5: Food, Fibre and Other Ecosystem Products (Box 5.8) (2022). www.ipcc.ch/report/ar6/wg2/downloads/report/IPCC_AR6_WGII_Chapter05.pdf
Risk assessment · International
Intergovernmental Panel on Climate Change. IPCC AR6 Working Group II, Chapter 9: Africa (2022). www.ipcc.ch/report/ar6/wg2/downloads/report/IPCC_AR6_WGII_Chapter09.pdf
Risk assessment · Africa

Reviewed 9 October 2026.