Abstract
Arabica coffee is Rwanda’s most economically significant export crop, grown mainly by more than 400,000 smallholder farmers, and is highly sensitive to temperature and rainfall. This study assesses how climate change may affect coffee suitability across Rwanda. An ensemble species distribution model combining Maximum Entropy, Random Forest, Boosted Regression Trees and Support Vector Machines was trained on coffee presence data with climate, soil, and elevation predictors. Climate inputs came from nine CMIP6 models, bias-corrected through ISIMIP and downscaled to about 1 km with the CHELSA method under the HiClAP initiative. Suitability was projected for periods centered on 2030, 2055, and 2080 under SSP1-2.6, SSP3-7.0, and SSP5-8.5. Model performance was acceptable to high (AUC 0.78–0.97), with elevation, number of dry days, and total precipitation as the most influential predictors. About 71% of Rwanda’s land area is currently suitable, mainly in the central, western, and northern highlands. Suitability remains broadly stable under the low-emission pathway but declines under higher-emission pathways over time; by 2080 under SSP5-8.5, losses reach about 3% of the currently suitable area against gains of about 1.2%. Losses concentrate in the eastern and southern lowlands, shifting suitability towards cooler, higher altitudes. Some areas gaining suitability lie within or near protected areas, so any expansion must respect land-use planning, conservation objectives, and the EU Deforestation Regulation. The study also examines gender-specific vulnerabilities: women, central to coffee production, face constraints in land, inputs, and access to climate information. Farm-level adaptation, including agroforestry and soil and water conservation, is recommended.
| Author / Institution | Izere D., Abigaba D., Mazumder A., Niyompuhwe O., Umuhoza J., Rösler K., Undorf S., Gornott C., von Loeben S. (AIMS-RIC; PIK; GIZ; University of Kassel) |
| Year | 2026 |
| Sector | Agriculture |
| Resource Type | Publication |
| Related NDC Theme | Land Use, Land-Use Change & Forestry (LULUCF); Climate-Resilient Agriculture |
| Related SDG | SDG 1; SDG 2; SDG 5; SDG 13; SDG 15 |
| Keywords | Arabica coffee, crop suitability, ensemble species distribution modelling, machine learning, downscaled climate projections, CMIP6, SSP scenarios, gender, climate adaptation, Rwanda |
| Contact Person | Delice Izere (AIMS-RIC) - izeredelice@gmail.com |

