Dynamic coupling of allometric ratios to a process-based forest growth model for estimating the impacts of stand density changes

Abstract

Process-based models are increasingly applied for simulating long-term forest developments in order to capture climate change impacts and to investigate suitable management responses. Regarding dimensional development, however, allometric relations such as the height/diameter ratio, branch and coarse root fractions or the dependency of crown dimension on stem diameter often do not account for environmental influences. While this may be appropriate for even-aged, monospecific forests, serious biases can be expected if stand density or forest structure changes rapidly. Such events occur in particular when forests experience disturbances such as intensive thinning or during early development stages of planted or naturally regenerated trees. We therefore suggest a calculation of allometric relationships that depends primarily on neighbourhood competition. Respective equations have been implemented into a physiology-based ecosystem model that considers asymmetric competition by explicit simulation of resource acquisition and depletion per canopy layer. The new implementation has been tested at two sites in Germany where beech (Fagus sylvatica) saplings have either been planted below a shelterwood of old spruces (Picea abies) or grown under clear-cut conditions. We show that the modified model is able to realistically describe tree development in response to stand density changes and is able to represent regeneration growth beneath a gradually decreasing overstorey of mature trees. In particular, the model could represent the faster crown size development in saplings until full ground coverage is established and a faster height growth afterwards. The effect enhances leaf area and thus assimilation per tree and increases carbon availability for stem growth at early development stages. Finally, the necessity to consider dynamic allometric relations with respect to climate change impacts is discussed, and further improvements are suggested.