Search

The growing-season length of temperate and boreal trees has a strong effect on the global carbon cycle. Yet, a poor understanding of the drivers of phenological processes, such as autumn leaf senescence in deciduous trees, limits our capacity to estimate growing-season lengths under climate change. While temperature has been shown to be an important driver of autumn leaf senescence, carbon source–sink dynamics have been proposed as a mechanism that could help explain variation of this important process. According to the carbon sink limitation hypothesis, senescence is regulated by the interplay between plant carbon source and sink dynamics, so that senescence occurs later upon low carbon inputs (source) and earlier upon low carbon demand (sink). Here, we manipulated carbon source–sink dynamics in birch saplings (Betula pendula) to test the relevance of carbon sink limitation for autumn leaf senescence and photosynthetic decline in a widespread deciduous tree. Specifically, we conducted a gradient of leaf and bud removal treatments and monitored the effects on autumnal declines in net photosynthesis and the timing of leaf senescence. In line with the carbon sink limitation hypothesis, we observed that leaf removal tended to increase total leaf-level autumn photosynthesis and delayed the timing of senescence. Conversely, we did not observe an effect of bud removal on either photosynthesis or senescence, which was likely caused by the fact that our bud removal treatment did not considerably affect the plant carbon sink. While we cannot fully rule out that the observed effect of leaf removal was influenced by possible treatment-level differences in leaf age or soil resource availability, our results provide support for the hypothesis of carbon sink limitation as a driver of growing-season length and move the scientific field closer to narrowing the uncertainty in climate change predictions.