Apparent differences in agroclimatic requirements for sweet cherry across climatic settings reveal shortcomings in common phenology models.

• PLS regression failed to delineate temperature response phases for multi-location dataset. • Estimated agroclimatic requirements differed between study locations. • Site-specific estimates of agroclimatic requirements indicate need for model revisions. • Phenology response to temperature (and thus to global warming) is location-specific. Temperate fruit trees are widely cultivated across the world's temperate regions. These trees are well-adapted to cold-winter climates through their ability to synchronize their phenology with the seasons. In autumn, they enter a dormant state, which allows them to survive the low winter temperatures and lasts until they resume growth in early spring. We analyzed the agroclimatic requirements (chill accumulation in Chill Portions, CP, and heat accumulation in Growing Degree Hours, GDH) for blooming in three sweet cherry cultivars ('Samba', 'Burlat', and 'Sylvia') grown in distinct climatic settings in Bonn (Germany) and Zaragoza (Spain). We used Partial Least Squares (PLS) regression analysis to relate bloom dates of the three cultivars grown in both locations to local temperatures. In Bonn, the colder location, trees experienced a long period of chill exposure (87–105 CP), which allowed a rapid growth response to warm temperatures (3233–4343 GDH). The flowering dates were mainly driven by conditions during the forcing period. In contrast, in the warmer climate of Zaragoza, chill exposure of the trees was relatively short (48–59 CP). The buds required long exposure to warm conditions (5444–6988 GDH) to subsequently bloom. In this case, flowering dates were influenced more by exposure to chilling than by conditions during the heat accumulation period. Global warming caused opposite effects on flowering dates depending on location. While in Bonn flowering dates have advanced between 3 and 5 days per decade, bloom dates in Zaragoza did not show such a trend, except for minor flowering delays in 'Sylvia', the late-flowering cultivar. Our results show that the response of the flowering dates to temperature appeared to depend on specific local climatic conditions. Although we applied current methodologies to determine the agroclimatic requirements of these cultivars, our methods were unable to derive consistent estimates of agroclimatic needs across the two locations. [ABSTRACT FROM AUTHOR]