Modeling Climate Change Impacts on the Phenology of Temperate Fruit Trees

Climate change affects temperate fruit production not only during the growing season but also during winter dormancy. In cool regions, warmer winters may lead to earlier bloom and increase frost exposure, whereas in Mediterranean climates reduced chill accumulation can delay or disrupt flowering and lower fruit quality. These responses depend strongly on local climate and cultivar-specific agroclimatic requirements, making accurate modeling of winter dormancy essential for impact and risk assessment. In this dissertation, I improve and apply the process-based phenology model PhenoFlex to project spring phenology across cultivars, species and climate zones and to link these projections with production risks such as frost damage and unmet chill requirements. A particular focus is placed on advancing calibration strategies to enable robust application under limited data availability and diverse climatic conditions. 1. In Chapter 2, I apply PhenoFlex to three almond cultivars in California under multiple climate scenarios. The model was calibrated using long-term bloom observations and evaluated with respect to both predictive accuracy and biological plausibility. Simulations under multiple climate scenarios indicate different trends: while bloom timing remains stable or slightly delayed under optimistic scenarios, substantial delays emerge under stronger warming. Cultivar-specific responses in the response strength could disrupt pollination synchrony in the orchards. 2. In Chapter 3, I expand the analysis to 110 cultivars of seven temperate fruit and nut tree species (apple, pear, apricot, sweet cherry, plum, almond, pistachio) across Spain, Morocco, Tunisia and Germany. Based on several climate scenarios, I project bloom shifts and probabilities of bloom failure. Results reveal three dominant patterns: advancing bloom in Germany, delayed bloom in southern Mediterranean regions and stable bloom in intermediate climates. Cultivar-specific differences in sensitivity to winter warming highlight opportunities for adaptation through cultivar-choice. 3. In Chapter 4, I extend PhenoFlex to simulate multiple phenological stages (budbreak, first bloom, full bloom) and apply the framework to assess spring frost risk for three apple cultivars in southwestern Germany. Although frost occurrence decreases under future climates, advancing phenology partly offsets this benefit. Consequently, while frost-free seasons become more frequent, the probability of damaging frost events during sensitive stages remains substantial, indicating continued relevance of frost risk management. 4. In Chapter 5, I introduce a novel combined-fitting calibration approach that pools phenological observations from multiple cultivars of the same species, enabling shared chill and heat submodels. This method is designed to overcome the limitation of short datasets. A comparative evaluation using almond, apricot and sweet cherry datasets show prediction accuracy comparable to single-cultivar calibration, while facilitating cross-cultivar comparison. This dissertation improves the calibration and application of the PhenoFlex model for projecting fruit tree phenology under climate change. Integrating phenology projections with risks assessments proved essential for translating model outputs into production-relevant information. Continued collaboration between modelers, plant physiologists, and growers is critical to further improve model realism and support climate adaptation in fruit production.