Search

This paper examines long-term eddy covariance data from 18 European and 17 North American and Asian forest, wetland, tundra, grassland, and cropland sites under non-water-stressed conditions with an empirical rectangular hyperbolic light response model and a single layer two light-class carboxylase-based model. Relationships according to ecosystem functional type are demonstrated between empirical and physiological parameters, suggesting linkages between easily estimated parameters and those with greater potential for process interpretation. Relatively sparse documentation of leaf area index dynamics at flux tower sites is found to be a major difficulty in model inversion and flux interpretation. Therefore, a simplification of the physiological model is carried out for a subset of European network sites with extensive ancillary data. The results from these selected sites are used to derive a new parameter and means for comparing empirical and physiologically based methods across all sites, regardless of ancillary data. The results from the European analysis are then compared with results from the other Northern Hemisphere sites and similar relationships for the simplified process-based parameter were found to hold for European, North American, and Asian temperate and boreal climate zones. This parameter is useful for bridging between flux network observations and continental scale spatial simulations of vegetation/atmosphere carbon dioxide exchange. [ABSTRACT FROM AUTHOR]

Tree mortality is an important ecological process influencing multiple functions of forest ecosystems. Previous studies have shown two basic size–mortality patterns, including a competition‐driven declining and a disturbance‐driven increasing mortality rate with tree size. Subtropical forests, which have a high species diversity and subject to frequent monsoon disturbances, are widely distributed in eastern Asia. However, the tree size–mortality pattern in the mature subtropical forests remains unclear.Here we analysed patterns of size‐dependent mortality from tree species to forest community using a 5‐year inventory data from 117 species and 163,612 individuals in a 20‐ha forest dynamic plot in a mature subtropical monsoon evergreen forest in eastern China. To explain the spatial variability in mortality patterns, two major biotic drivers (competition and tree size) and multiple local‐scale environmental factors were further analysed.Our results showed that tree size was the best predictor of tree mortality at the scales of both species and community. A species‐level analysis identified four size–mortality patterns that are shaped by species‐specific attributes such as maximum size and life form. For 27 out of 92 species that comprised 59% of tree individuals, the relationship between size and mortality exhibited a U‐shaped pattern of a first decline followed by an increase. An overall community‐scale size‐dependent mortality also showed a U‐shaped pattern.Tree mortality was also influenced by the competition and environmental conditions, but the relative importance varied widely across tree sizes and species. The competition showed significant correlations with the mortality of small trees while the effect of environmental conditions on mortality was strongest for large trees. A principal component analysis showed that a combination of biotic and abiotic factors explained 42.3% of the spatial variation in mortality at large sizes.Synthesis. Our results reveal four identifiable size‐dependent mortality patterns that differ across diverse species, jointly leading to a U‐shaped size–mortality pattern at the community level. This finding calls for the need to establish the details of every potential size–mortality pattern with consideration of the different effects of biotic and abiotic factors on tree mortality of specific size. [ABSTRACT FROM AUTHOR]