Published November 28, 2023 | Version v1
Publication Open

Vulnerability of cocoa-based agroforestry systemsto climate change in West Africa

  • 1. Department of Forest Engineering, Laboratory of Dendrochronology, Silviculture and Global Change, DendrodatLab-ERSAF, University of Cordoba, Campus de Rabanales, Crta. IV, km. 396, 14071 Córdoba, Spain
  • 2. Dept. Biodiversity Conservation and Ecosystem Restoration, Pyrenean Institute of Ecology-CSIC. Avd. Montañana, 1005 Zaragoza, Spain
  • 3. CABI, Bakeham Lane, Egham, Surrey, TW20 9TY, UK
  • 4. Mondelez UK R&D Limited, Bournville, B30 2LU, UK
  • 5. Department of Agricultural Sciences, Inland Norway University of Applied Sciences, 2322, Hamar, Norway
  • 6. Digital Inclusion, Bioversity International, Montpellier, France


The rise in temperature and increasing frequency of severe droughts predicted for the near future will make cocoa plantations more vulnerable, ultimately affecting their vigour, yield and long-term persistence. As a result, some important cocoa cultivation areas in West Africa in the 20th century will have essentially become unsuitable for growing cocoa by the near future. This research work has focused on the vulnerability of cocoa potential distribution under climate change as well as the potential impacts of climate change on the habitat suitability for shade tree species commonly used in cocoa-based agroforestry systems across West Africa. Finally, this information was used to estimate the potential of C-AFS across different scenarios of climate change and discuss opportunities and possible risks of cocoa expansion into climatically less vulnerable areas.

Methods: To characterize patterns of habitat suitability at regional level, we compiled presence location points of selected tree species from the Global Biodiversity information Facility (GBIF). We modelled current and future suitable habitats for 38 tree species (including cocoa). Suitability modelling and mapping of all species were done using a consensus method for species distribution modelling (SDM). As environmental predictors in the models, edaphic and climatic variables were used. We considered seven climatic scenarios for 2060 using two contrasting projections of atmospheric CO2 concentration.

Results: The prediction accuracy of the species distribution models for cocoa and shade trees showed values between 0.69 and 0.99. The final ensemble model that was built using all the data had an AUC range from 0.84 to 0.99 for the different species, while TSS value range from 0.66 to 0.96 for the ensemble models. The sum of the contribution rate of the climatic variables reached 85% and that of the soil factors reached 15%. The results show an increase of the potential habitat suitability for Theobroma cacao compared to the current occurrence in West Africa. However, the model projections show a reduction of the highly suitable current areas by 2060. In this sense, the climatic suitability was found to shift from west to east across the coastal border. Regarding shade trees, the distribution range of 49% of the shade tree species will become unsuitable by 2040 in West Africa, and 60% by 2060.

Conclusions: Our results highlight the importance of transforming cocoa-based agroforestry systems by changing shade tree species composition to adapt this production systems for expected climate change scenarios.

Keywords: Habitat suitability, species richness, climate change adaptation, species distribution


Files (3.0 MB)