Synchronization in Phenological Time uncovers two-Phase Temperature Control of Leaf Fall Timing

Leaf senescence timing impacts carbon, water and nutrient cycles as well as biotic interactions and microclimatic conditions. Consequently, it is crucial to elucidate drivers and to anticipate future shifts of leaf senescence. One major driver of leaf senescence is widely believed to be temperature. Numerous studies investigated the effect of temperature on leaf senescence, yet most delineate periods of interest based on calendar times (e.g. weeks or months). However, across years or locations the same calendar time does not equal the same point in the seasonal cycle of a plant, i.e. in “phenological time”. Thus, expressing temperature-dependencies in calendar time might mask effects that are more obvious in phenological time.

To elucidate the influence of temperature on leaf senescence synchronized in phenological time we performed three investigations. First, using data on Malus domestica for Germany we examined whether phases of temperature-dependency of leaf senescence manifest, and whether these phases occur consistently across cultivars. Second, we examined whether theses phases occur consistently in Malus domestica over Europe, covering a broad range of climates, altitudes and latitudes. Third, we examined if the same phases occur in another pome fruit (Pyrus communis), stone fruit (Prunus avium) and other deciduous tree species (Fagus sylvatica, Quercus robur) in Europe.

In Malus domestica, we consistently, i.e. independent of cultivar, latitude, climate and for altitudes up to 1,400 m, found a spring phase, occurring around 300 to 100 days before 50 % of leaves fell (DBLF₅₀), and a fall phase, occurring around 80 DBLF₅₀ until the day of 50 % leaf fall (DOLF₅₀). In a multiple linear regression, the mean temperatures of the spring and fall phases combined are a great predictor of DOLF₅₀ in Malus domestica (R² = 0.841 for Germany and 0.718 for Europe). Likewise, these phases were observed in all other species investigated, and the same multiple linear regression performed again well (R² between 0.777 and 0.887). Higher temperatures during the spring or fall phases respectively led to a delay or advancement of DOLF₅₀. This contrasts with the majority of studies which delineated temperature effects on leaf senescence using calendar periods.

Consequently, the novel approach to synchronize observations in phenological time allowed to uncover new insights on the temperature-dependency of leaf senescence timing. Namely, the consistent presence of a spring and a fall phase with inverse temperature effects on leaf senescence, across cultivars, latitudes, altitudes, climates and for a range of species. This finding, crucially, could enable the development of a “unifying” modelling framework for leaf senescence prediction across cultivars and species.