Coarse woody debris in natural beech forests: Generic predictions of the simulation model BEFORE-CWD of quantity, spatial distribution and availability

This aim of this paper is to examine the occurrence of dead wood in a generic virgin beech forest by using a simulation model to determine typical dead wood values (covering amount, availability and spatial distribution). These values should be constant during the development of the forest and robust to small changes in environmental conditions. As we are examining an idealized (modelled) virgin beech forest, the results tend to be qualitative. We show that the invariance of some dead wood values makes them helpful for developing a dead wood concept., However, the reference values must be adapted to the actual stand conditions. We extended the BEFORE model (developed for virgin beech forests) by incorporating both the CWD module (coarse woody debris) for dead wood and an estimation of the volume of the canopy layer trees. The extended model proved to be sensitive only to those volume parameter values which can be taken from the literature. Since the other volume parameter values only have a small influence on the model results, it does not matter that they can only be roughly estimated. As a result, the model is mainly determined by the model parameters and stable with respect to small changes in the estimated values. The parameter values of the dead wood module CWD (probability of fungal decay WahrschPilz and velocity of decomposition Tot-age) also have only a small influence on the target value of calibration (mean total tree volume per ha) and other values such as the mean total tree volume in the various developmental stages. By contrast, the values concerning the coarse woody debris are closely related to the velocity of decomposition. However, due to the coarse time pattern of the model, this parameter can be estimated with sufficient accuracy. The results of the CWD model with standard parameters (Table 2) closely match the empirical results (Tables 5, 6) describing tree volumes in the whole forest area or individual developmental phases', as wen as snags or downed logs. Although CWD is quite simple, significant results not only on coarse woody debris but also of total tree volume can be obtained. Invariant indicators can be identified, i.e. indicators which hardly change during the development of the forest (in the model: during a simulation run) or under slightly varied environmental conditions (in the model: model parameters). These values describe the number, volume and spatial distribution of snags. Although total dead wood volume oscillates in time, the oscillations of dead and live wood augment each other such that the mean total tree volume is (nearly) constant relative to the whole forest area and the individual development stages. A suitable compromise between a managed forest and large unmanaged areas might be to mark out pristine forest islands. Our results show that the typical invariant indicators win even develop on an (inner) forest area of 7.4 ha. Furthermore, they. remain constant despite occasional severe storm events for the whole observation period of 3000 years.