Dynamics of a mixed beech (Fagus silvatica L.) and hornbeam (Carpinus betulus L.) natural regeneration in North-East France: evolution of stand density and species composition with an insight into tree mortality factors.

Abstract

ContextMixed beech and hornbeam regenerations are quite common in the east of France but their management is based more on general principles linked to ecosystem functions or services than on silvicultural practices. This is why it is so useful to establish stand trajectories to characterizing the evolution of density during stand development, as well as individual tree mortality models to predict how each species may react, particularly in a changing environment. AimsThis study was conducted to establish size-density trajectories of mixed even-aged beech and hornbeam natural regenerations, using an already tested piecewise polynomial function, and develop an individual tree mortality model based on a logit function. Material and methodsThe study took place in a mixed beech-hornbeam naturally regenerated stand in Hesse forest (NE of France), where a square design 42*42m comprising 64 square plots of 5.25m side was inventoried each year since 2001 until 2012. Measurements included tree species and status (dead or alive), girth at breast height for all trees and total height for a sample of living trees (beech and hornbeam). The size-density trajectories of the 64 plots describing the course of the number of living trees in relation with the mean stand girth, in logarithmic scales, were modeled with a piecewise polynomial function fitted with a mixed-effects model. On the other hand, the individual tree mortality model was fitted with a logit function including several independent variables defined at tree, plot and stand levels. ResultsThe size-density trajectory of mixed beech-hornbeam naturally regenerated stands was successfully fitted using the same function as for pure stands, with a plot-level random component that appeared linearly related to site fertility, initial density of trees (N0) and relative initial proportion of beech trees. The mortality onset appeared to occur at a higher density (RDI = 0.5) in mixed beech-hornbeam naturally regenerated stands than in pure beech even-aged stands (RDI = 0.29), while the maximum density (RDI = 1) was reached at a comparable relative number of surviving trees (N/N0). An individual tree mortality model could also be fitted using a logit function. The probability of mortality of trees appeared linked to individual (social status, species), collective (relative density) and site (fertility, water stress) factors. ConclusionThe size-density trajectory model first developed for pure even-aged stands appeared well adapted to mixed beech and hornbeam natural regenerations, and the individual tree mortality model constructed at the same time for these stands revealed that water stress induced mortality was relatively comparable for beech and hornbeam. These two models give the opportunity to simulate the development of mixed beech and hornbeam regenerations with the addition of a growth model.