Trees grow modulated by the ecological memory of their past growth. Consequences for monitoring, modelling, and silvicultural treatment

Abstract

The growth of forest trees under given environmental conditions is assumed to depend mainly on their age, stem and crown size, and competitive position in the stand. The current stem and crown size are commonly recognized as adequate proxy markers for the tree's ecological memory of the past. In contrast, tree ring structure, crown whorl morphology, or other biographical patterns are rarely used for predicting growth. Here, we asked how the latter affect tree growth.Our main hypothesis was that the growth in the longer past significantly co-determines the present growth. To test this hypothesis we derived metrics which quantify the social position, course of growth, and annual variation of trees in their past. We further selected variables for quantifying the trees' present stem and crown size and competitive status. Finally, we selected the approximately 200-years-old thinning experiment in European beech (Fagus sylvatica L.) Fabrikschleichach 15 in South Germany as our study object because it provided all required information.To examine the dependency of the current growth on the present growing conditions and the past tree development more closely, we applied linear mixed models. They revealed that (i) trees with similar age, size, crown and competitive status at present grew better if they were subdominant in the past. (ii) Ceteris paribus, slow starting trees with progressive growth trajectories were associated with higher growth than quick starting trees with degressive growth trajectories. (iii) Trees with lower interannual variations of growth in the past had significantly higher growth rates at present than trees with higher interannual variations of growth in the past. (iv) Including information about the trees' past reduced the RSME of the diameter growth model by 17–27% and increased the R2 by 15–30%. Thus, the diameter growth model could achieve R2 values of 0.76–0.79. (v) The contribution of past information for estimating present growth was higher in periods without thinning.We suggest that in the analysed European beech stands, even at parity of stem diameter or crown size, different courses of growth created different internal stem structures and crown morphologies. Such past structural and morphological formations may affect the tree's light interception and hydraulic conduction. These differences in structure may cause specific differences in the present tree functioning and growth. Of course these findings based on only one long-term experiment should not yet be generalized. However, the revealed relationship between the past and present growth deserves further investigations. We discuss the relevance of the ecological memory embedded in the past growth and in the tree ring pattern. We stress the consequences of the ecological memory for the monitoring, inventory, and modelling of tree growth and its implications for the development of silvicultural prescriptions.