Projections of where cacao will grow, and measurements of what heat, dry air, drought and heavy rain do to it. A projection is a model of climate suitability, not a forecast of harvests, and two of these models disagree about the direction of change in the same region.
24 studies2004 to 2026Each checked against its published abstract
What these studies add up to
No study here projects cocoa disappearing. Projections show losses near the dry edge of West Africa and in parts of the Amazon, stable areas further south, and in one model a small gain in West Africa by 2060.
Dry air harms cacao as much as dry soil: one harmattan season cost a third of leaf area either way.
Heavy rain may constrain yield more than heat. In Ghanaian records, excess wet-season rain and dry-season drought together explained two-thirds of year-to-year variation.
Drought can be devastating locally: after the 2015 to 2016 El Niño, sampled farms in Bahia had lost 15% of trees and 89% of yield.
Shade is not dependable protection. In one Ghanaian drought all the cocoa under one shade species died while unshaded cocoa survived; elsewhere shade helped.
More carbon dioxide raises photosynthesis and partly offsets heat and drought, in seedlings.
There is genetic variation in drought and temperature response, and breeding with it has not yet reached yield.
Four kinds of cacao differed in the temperature below which stems stop growing
Do cacao genotypes differ in their vegetative response to temperature?
How
Controlled environment glasshouse experiment simulating the temperature conditions of three cacao-growing regions (Bahia, Brazil; Tafo, Ghana; Lower Perak, Malaysia) over a year, on four genotypes.
Finding
Base temperatures calculated from main stem growth varied from 18.6 °C for AMAZ 15/15 to 20.8 °C for SPEC 54/1. The results imply that genetic variability exists in cacao in response to temperature stress.
Does not establish: It does not identify a heat-tolerant variety or an upper temperature limit.
Limits: Young plants and vegetative growth only. The simulated regimes were present-day climates, not future heat.
How does a cacao and Gliricidia agroforest respond to an experimentally imposed drought?
How
Replicated throughfall reduction experiment in a cacao/Gliricidia stand in Sulawesi, Indonesia, over a 13-month period, with roof plots and control plots.
Finding
During roof closure changes were small and non-significant, including cacao sap flux density (-11%) and cacao yield (-10%). In the main harvest following the period of lowest soil water, cacao bean yield in roof plots was substantially lower (-45%). There was evidence of complementary use of soil water through vertical partitioning between cacao and Gliricidia.
Does not establish: It does not show that shade trees protect cocoa yield in drought. Yield was the most drought-sensitive function measured.
Limits: One stand with one shade species in a wet climate. There was no unshaded treatment, so it cannot compare shade with full sun.
How do cacao and Gliricidia shade trees share soil water when rainfall is sharply reduced?
How
Throughfall reduction experiment in Sulawesi, Indonesia, with three control plots and three roof plots in which net precipitation was reduced by 71% for 13 months. Ambient precipitation at the site was 2841 mm per year.
Finding
Average monthly sap flux densities fell by a maximum of 21% in cacao and 29% in Gliricidia compared with control plots. Cacao withdrew water mainly from the topsoil, and competition between the two species for water was low or in favour of cacao. Average daily stand transpiration was 1.3 mm in roof plots and 1.5 mm in control plots.
Does not establish: It does not show that shade trees never compete with cocoa for water. Abdulai and colleagues found strong competition at a dry site in Ghana.
Limits: A very wet site, where small amounts of rain entering gaps in the roof frequently rewetted the topsoil. One shade species.
What is known about the water relations and irrigation needs of cocoa?
How
Narrative review of published research on cocoa water relations, drought tolerance, crop water use and water productivity.
Finding
Partial stomatal closure begins at a leaf water potential of about -1.5 MPa, and stomatal conductance declines as the saturation deficit increases from about 1.0 up to 3.5 kPa. Flowering is inhibited by water stress. Despite cocoa's sensitivity to water stress, there is a paucity of reliable, field-based published data on yield responses to drought or to irrigation.
Does not establish: It does not establish an irrigation requirement or a yield response to irrigation.
Limits: A narrative review reflecting the literature to its date. Not a systematic review.
Which cacao genotypes tolerate soil water deficit, and which traits mark tolerance?
How
Greenhouse experiment on six months-old seedlings of 36 cacao genotypes under two soil water regimes, with drought imposed until leaf water potential reached -2.0 to -2.5 MPa.
Finding
Genotypes MA-15, MO-20 and PA-13 were classed as tolerant and CC-40, C. SUL-4 and SIC-2 as non-tolerant. Leaf and total dry biomass, relative growth rate and leaf magnesium content were the most important variables for classifying genotypes.
Does not establish: It does not show that these genotypes yield more than others in dry years on farms.
Limits: Seedlings in pots. Tolerance at this stage may not predict yield of mature trees in drought.
How vulnerable is cocoa in the West African cocoa belt to projected climate change, and where?
How
Modelling study. A statistical model of climatic suitability (Maxent) plus analysis of individual limiting climate variables, using projections for the 2050s from 19 Global Circulation Models under emissions scenario RCP 6.0.
Finding
Maximum dry season temperatures are projected to become as or more limiting for cocoa as dry season water availability. The most vulnerable areas are near the forest-savanna transition in Nigeria and eastern Côte d'Ivoire, and the least vulnerable in the southern parts of Cameroon, Ghana, Côte d'Ivoire and Liberia. The authors expect shifts in production within the region, with losses and gains partially compensating.
Does not establish: It does not project that cocoa will disappear from West Africa. It identifies areas of low vulnerability and possible gains alongside losses.
Limits: A correlative suitability model for one scenario and one time horizon. It models where today's cocoa climate will be found, not how trees or farmers respond.
How should adaptation be planned across the West African cocoa region given uneven climate vulnerability?
How
Modelling study. A statistical model of relative climatic suitability calibrated on West African cocoa farming areas, with average climate projections for the 2030s and 2050s from 15 and 19 Global Circulation Models respectively.
Finding
The region divides into three adaptation zones: a little affected zone permitting intensification or expansion, a moderately affected zone requiring diversification and agronomic adjustment, and a severely affected zone needing progressive crop change. Local production losses could be compensated through intensification and expansion elsewhere.
Does not establish: It does not say how much cocoa land falls in the severely affected zone, and it does not predict a regional collapse.
Limits: The abstract gives no areas or shares for the three zones. Zoning inherits the assumptions of the suitability model.
Which daily microclimatic variables govern the appearance of Moniliophthora pod rot symptoms and sporulation?
How
Field cohort study on the highly susceptible clone Pound-7: 55 cohorts totalling 2,268 pods tagged weekly and assessed through their lifetime, analysed with generalised linear mixed models. The abstract does not name the country.
Finding
Daily average wetness frequency before tagging was positively linked with symptom appearance, and the effect of daily average maximum temperature after tagging exhibited a maximum at 30°C. Sporulation was negatively linked with minimum temperature and positively with maximum temperature.
Does not establish: It does not project how frosty pod rot will change under climate change.
Limits: One highly susceptible clone at one site. Statistical associations with microclimate.
What did the 2015-16 El Niño drought do to adult cocoa trees in Bahia's agroforests?
How
Field measurement on randomly chosen farms in Bahia, Brazil, after the 2015-16 ENSO drought. The abstract gives no number of farms.
Finding
Drought caused high cocoa tree mortality (15%) and severely decreased cocoa yield (89%). It also increased the infection rate of witches' broom disease.
Does not establish: It does not show that such losses are permanent or that climate change has already raised El Niño frequency. The authors describe an increase as possible.
Limits: One region and one extreme event, measured after the fact. The abstract does not describe the baseline against which the yield loss was calculated.
Does shade protect cocoa during drought, or do shade trees compete with it for water?
How
Field monitoring of micro-climate, sap flux density, throughfall and soil water content from November 2014 to March 2016 at the forest-savannah transition zone of Ghana, comparing cocoa in full sun with cocoa shaded by Albizia ferruginea or Antiaris toxicaria. The abstract gives no number of plots or trees.
Finding
During the extreme drought of 2015/16 all cocoa plants under A. ferruginea died and those under A. toxicaria suffered 77% mortality. Cocoa in full sun maintained a sap flux density of 170 g cm-2 day-1 against 115 under A. toxicaria, and recovered to 163 against 37. Soil water content in full sun was higher than in shaded systems.
Does not establish: It does not show that shade is harmful in general. Schwendenmann and colleagues and Mensah and colleagues in this set report different outcomes under other conditions.
Limits: One site at the dry margin of the cocoa belt, two shade species, one extreme drought. Replication is not stated in the abstract.
How do seasonal climate variables relate to the abundance of cocoa pollinating midges?
How
Monthly sampling with static suction traps along transects on six cocoa farms across three Caribbean islands for a full year. Over 87,000 insects were captured, including more than 1800 cocoa midges.
Finding
The previous month's rainfall positively predicted cocoa midge abundance, and during drought periods midge numbers were very low. Humidity and mean daily maximum and minimum temperatures did not relate to midge abundance. Midges typically comprised less than 2% of insects caught.
Does not establish: It does not show that drought reduces cocoa yield through lost pollination.
Limits: One year on six farms. Midge abundance was measured, not pollination or fruit set.
Are cocoa and common agroforestry trees viable alternatives where climate change threatens coffee in Mesoamerica?
How
Modelling study of potential changes in the distribution of coffee, cocoa and the 100 most common agroforestry trees in Mesoamerica. The abstract does not state the scenario or time horizon.
Finding
Cocoa could potentially become an alternative in most coffee vulnerable areas. Agroforestry with currently preferred tree species is highly vulnerable to future climate change.
Does not establish: It does not show that cocoa is safe from climate change in Mesoamerica, only that it fares better than coffee in the areas coffee loses.
Limits: No scenario, horizon or area figures in the abstract. Distribution models only.
How do elevated carbon dioxide, higher temperature and water deficit interact in young cocoa?
How
Open top chamber experiment on 6-month-old cocoa seedlings: chamber control at 400 ppm, elevated carbon dioxide at 550 and 700 ppm, temperature 3 °C above control, and the combination, each at 100% and 50% of field capacity.
Finding
Increasing carbon dioxide concentration increased photosynthesis by 27% and improved whole plant water use efficiency. A maximum temperature around 39 °C against 36 °C in the control had a quite severe impact on photosynthesis, leaf water potential and biomass. Elevated carbon dioxide ameliorated the negative effects of heat and water deficit to a certain extent.
Does not establish: It does not show that rising carbon dioxide will protect cocoa yields in the field.
Limits: Seedlings in chambers for a short period. The control maximum of 36 °C is already hot, and mature bearing trees may respond differently.
How do atmospheric drought and soil drought each affect mature cacao during the harmattan?
How
Field monitoring of two sub-plots, with and without irrigation, through one Harmattan season (November 2019 - March 2020) in the Eastern region of Ghana, recording sap flow, light, soil moisture, temperature, humidity and precipitation, analysed with boosted regression tree models.
Finding
The atmospheric component of Harmattan-induced drought affected the canopy to a similar extent as soil water stress, both causing a decline in leaf area index of 33%. Soil and atmospheric water stresses did not have a synergistic effect on transpiration under the studied conditions.
Does not establish: It does not show that irrigation would protect cocoa yield through the harmattan.
Limits: Two sub-plots, one season, no replication stated. No yield was measured.
Can biochemical markers identify drought-tolerant cocoa hybrids?
How
Breeding experiment: hybrids produced by crossing four tolerant genotypes in all possible combinations were subjected to water stress and assayed for proline, nitrate reductase activity, superoxide dismutase and glycine betaine. The abstract gives no number of plants.
Finding
Drought tolerant hybrids had high proline, superoxide dismutase and glycine betaine content, and higher nitrate reductase activity than susceptible hybrids. Logistic regression selected proline and nitrate reductase as the two markers for future breeding.
Does not establish: It does not show that a drought-tolerant cocoa variety is available to farmers.
Limits: Young plants and biochemical indicators. No field performance or yield is reported.
How are molecular adaptations distributed across the natural range of cacao?
How
Population genomic analysis of 79 genomes of Theobroma cacao using multiple selection scans to infer selective sweeps.
Finding
The majority of molecular adaptations are not shared among populations. About 71.5% of genes under selection also show significant associations with changes in environmental variables, and disease resistance genes made up 6.5% of selective sweeps.
Does not establish: It does not identify climate-resilient varieties ready for planting.
Limits: Statistical signatures in genomes, not tested performance of trees under heat or drought.
How might climate change alter where cocoa can be grown in the Brazilian Amazon biome?
How
Modelling study. An ensemble of 10 correlative models run in the biomod2 library and projected to two future climate scenarios (RCPs 4.5 and 8.5) by 2050, combined with soil suitability and installed infrastructure.
Finding
The authors found a 37.05% and 73.15% decrease in the areas suitable for intensification and expansion zones under RCP 4.5 and 8.5, respectively, compared with the current scenario.
Does not establish: It does not project that cocoa growing in the Amazon ends by 2050. It projects a reduction and spatial reconfiguration.
Limits: Correlative models of suitability under two scenarios. The zones are defined partly by infrastructure, so the figures are not purely climatic.
Will the shade trees used in cocoa agroforestry remain suitable where cocoa does?
How
Modelling study. Consensus species distribution modelling of current and future habitat suitability for 38 tree species including cocoa, using climatic and soil variables. The abstract does not state the emissions scenario.
Finding
The models projected an increase of up to 6% of the potential suitable area for cocoa by 2060 compared with its current suitable area in West Africa. The suitable area was reduced by 14.5% once only land not contributing to deforestation was considered. 50% of the 37 shade tree species will see a decrease in geographic extent by 2040 and 60% by 2060.
Does not establish: It does not show that cocoa production will rise. It contradicts the idea of a uniform loss of cocoa land, but the gain disappears when forest land is excluded.
Limits: Correlative models. The scenario is not given in the abstract, and suitable area is not the same as yield.
Which soil and terrain features are related to cacao tree mortality in dry years?
How
Observational study of mortality of agroforestry-grown cacao over three periods (1994-1995, 2009-2013 and 2015-2019) on 36 plots of one farm in Bahia, Brazil, in a subhumid to dry climate, using correlation and cluster analysis.
Finding
The soil attributes most related to mortality were physical (granulometry, porosity and density) and mineralogical, mainly in the subsurface layer. The lowest mortality was observed in Oxisols and on slopes facing south and east.
Does not establish: It does not establish how much mortality a given drought causes, only which sites fared better.
Limits: A single farm. The abstract gives no mortality rates.
Does shade alleviate drought stress in mature cocoa trees under field conditions?
How
Split-plot field experiment in Ghana over 33 months on twelve-year cocoa plants: three levels of rainwater suppression (full rainwater, 1/3 and 2/3 suppression) under full sun or 40 % uniform shade.
Finding
Soil moisture increased under shade. Rainwater suppression lowered leaf water potential, reaching -1.5 MPa in full sun. Dry bean yield was higher under shade than under full sun, although fruit abortion and pod damage were also high under shade. There was no interaction between shade and rainwater suppression.
Does not establish: It does not show that shade trees protect cocoa from drought. The authors conclude shade may improve performance but not sufficiently to counteract water stress.
Limits: Uniform artificial shade at one level, so competition from real shade tree roots is absent. The abstract gives no yield figures.
How do higher temperature and elevated carbon dioxide together affect early growth and physiology of cacao?
How
Factorial growth chamber experiment on juvenile plants of two genotypes (SCA 6 and PA 107): two carbon dioxide levels (410 and 700 p.p.m.) and three day/night temperature regimes (31/22 °C, 33.5/24.5 °C and 36/27 °C) at a constant vapour pressure deficit of 0.9 kPa.
Finding
Elevated carbon dioxide increased final dry weight, leaf area, photosynthesis and intrinsic water-use efficiency in both genotypes. Raising temperature above 31/22 °C enhanced photosynthesis, transpiration and stomatal conductance, but dry weight and leaf areas declined, more in SCA 6 than in PA 107.
Does not establish: It does not show that mature cacao will yield more in a warmer, higher carbon dioxide climate.
Limits: Juvenile plants, two genotypes, and air dryness held constant. In the field hotter air is usually drier.
How do pheromone trapping and weather affect mirid populations and damage in cocoa?
How
A 2-yr field study in 11 plantations across Ayos and Konye, Cameroon, recording pheromone trap catches, pod counts, damage and weather.
Finding
Of the recorded weather variables, only relative humidity was correlated, negatively, with damage severity. Cocoa pod counts were higher in 2015 than in 2016 and negatively correlated with temperature and relative humidity. Mirid populations and damage were suppressed where pheromone traps were used.
Does not establish: It does not project mirid pressure under future climates.
Limits: Two years and two localities give little climatic contrast. Correlations only.
How do El Niño and La Niña events affect vegetation productivity in potential coffee and cacao areas of Latin America?
How
Correlational remote sensing study of eight El Niño and La Niña events (1992-2020) over potential coffee and cacao growing areas estimated from species distribution models, using gross primary productivity as a proxy for idealised two-layer agroforestry systems.
Finding
Exposure to El Niño was ubiquitous for both crop zones. Hotspots of productivity decrease during events were mostly in Central America, Northern South America, the Western Amazon and coastal Brazil. The authors raise particular concern for potential lowland cacao areas of the Amazon.
Does not establish: It does not quantify cocoa yield loss from El Niño.
Limits: Vegetation productivity over modelled potential areas, not cocoa yield on actual farms.
Which climate variable best explains year-to-year variation in cocoa yields?
How
Statistical analysis of district-level production records from Ghana combined with daily rainfall and temperature data, with national yield data from Ecuador and Indonesia as a check. The abstract does not give the years covered.
Finding
Excess wet-season rainfall and dry-season drought together explain 68% of interannual yield variability in Ghana. National yield data from Ecuador and Indonesia give a consistent negative response to heavy rainfall. The authors identify heavy rainfall, rather than temperature or mean water availability, as the dominant climatic constraint.
Does not establish: It does not show that warming is harmless to cocoa. It reports that temperature explained less of past year-to-year variation than rainfall extremes did.
Limits: Observational analysis of aggregated production records. The mechanism, which the authors suggest is damage and disease during key stages, is inferred and not measured.
We would like to set two Google Analytics cookies to count page views, and to count them in our own cookie-free measurement as well. They do not identify you and nothing is shared with advertisers. Decline and the site works exactly the same — see cookies for what each one does.