Search

Phenological variation among trees of different heights provides a small-scale ecological distinction within the forest, allowing the modulation of light interception and, consequently, net carbon gain. While spring phenological variations in temperate forests are well studied, for autumn phenology it is still unclear whether canopy vertical position (exposure), ontogeny or microclimatic factors are more decisive. We observed leaf colouring (LC) and leaf fall (LF) phenology of 166 Fagus sylvatica L. individuals (European beech; Fagaceae), twice/three times a week during autumn 2012 in a mixed forest in southern Germany. We aimed to determine: (1) the extent of variations in leaf senescence among F. sylvatica trees occurring in three different vertical canopy positions in the forest (overstorey, mesostorey and understorey); and (2) whether phenology varies between three different canopy height levels (upper, intermediate and lower). Possible microenvironmental drivers, such as air temperature, air humidity and light, were analysed in relation to autumn phenology. Air temperature and humidity data were obtained from HOBO loggers and light conditions from hemispheric fisheye photographs. Overstorey individuals were the first to start autumn phenology followed by mesostorey and understorey trees. For understorey individuals, the onset of LC and LF were observed 31 and 24 days later than for overstorey trees. The upper canopy parts of individual trees were characterised by the earliest appearance of autumn phases; mean difference in onset date of LC and LF between the upper and lower level was in each case -11 days. As peak dates did not differ, the duration of autumn phases were shorter at the lowest canopy height levels. We find a remarkable phenological avoidance of understorey trees and lower leaves compared to overstorey trees and upper canopy parts in F. sylvatica . We suggest that the observed differences were related to vertical variations in relative humidity and light availability, but also have an ontogenic cue. Since phenological variation in forest stands alters a range of environmental conditions, our study is useful from an ecological and microclimatic viewpoint. Moreover, since phenological development was shown to differ considerably, generalisations are limited when considering trees of different life stages within a forest. Further studies should focus on light conditions to investigate their influence on autumn phenology and importance for phenological avoidance. [ABSTRACT FROM AUTHOR]