Coffee is an important cash crop for small-scale farmers in tropical and subtropical regions, but climate change is expected to increase yield variability and destabilize production. To evaluate climatic effects on coffee yield, it is necessary to understand sensitivity to rainfall and temperature, because their impacts may differ by phenological stage and local environments. Although previous studies examined stage-specific climatic effects in coffee-producing countries, it remains unclear where, when, and under what regional conditions these effects become yield risks. In Bali, such variation is important because wet and dry seasons, elevation, and cultivar composition differ across districts.
Therefore, in this study, stage- and region-specific relationships between coffee yield and climatic factors were analyzed in Bali, Indonesia, using annual yield data and monthly climate data. The analysis covers Bali as a whole and three districts: Bangli, Tabanan, and Buleleng. The crop cycle was divided into flowering, early fruit development, late fruit development, and harvesting stages, and simple regression was conducted by region and month.
The results showed that rainfall during the flowering stage tended to be negatively associated with coffee yield in several regions. In particular, at the island-wide level, September rainfall showed a strong negative correlation with yield (r = -0.80). This suggests that, in Bali, flowering-stage rainfall may not simply function as water supply but may act as excessive rainfall stress, affecting yield formation through flower drop, pollination disturbance, or moisture stress. Temperature showed no uniform relationship with yield at the island scale, while different tendencies were observed among districts, suggesting that temperature effects may vary depending on elevation and cultivar composition. These findings indicate that climatic risks to yield in Bali should be evaluated by considering phenological stages, regional conditions, elevation, and cultivar characteristics. Future research will connect these findings to regional yield risk assessment and adaptation planning.