Anwendung und evaluierung des waldwachstumssimulators SILVA 2.2 auf bestandes-, Betriebs-, und groṙegion-ebene

The focus of this paper is on the application and validation of the forest growth simulator SILVA 2.2 at the level of forest stands, forestry enterprises and large regions. The Design concept and structure of this distance-dependent and site-sensitive individual tree simulator have been presented in PRETZSCH (1992), KAHN (1994), KAHN and PRETZSCH (1997), PRETZSCH (2001). The simulator owes its flexibility at stand level to the spatially explicit model approach and the parameter estimation, which is based on long-term experimental plots in pure and mixed stands. Some of the included time series date back to the year 1870. The simulator can be applied for various purposes and at different scales, and the results of the scenario calculations include estimates of timber yields as well as ecological and socio-economic indicators. In this paper, the development of a pure and mixed stand are compared by simulation at stand level. Thereafter, different silvicultural policies are investigated at enterprise level. The long-term consequences of forest management decisions at stand level become transparent once the growth projections are carried out at enterprise level. Thus, individual treatment regimes can be modified in order to achieve desired developments at enterprise level. Finally, the site-sensitive approach of SILVA 2.2, which is described in this paper, allows the analysis of the consequences of climate change. This is demonstrated in scenario calculations, which investigate the consequences of climate change on the growth and wood supply of spruce in Bavaria. The model validation was based on long-term experimentalt plots and showed a bias of -1.9% to 4.8%, an accuracy of 19.2% to 38.6% and a precision of 18.5% to 38.6% for the estimate of current stand volume increment. Similar results were obtained for mean diameter increment and, furthermore, for individual tree level. In case of short increment periods and extraordinary climatic conditions the values of bias and accuracy may prove less favourable. In a qualitative comparison with inventory data the prediction behaviour of the site-growth model was proved to be fairly robust. Currently, however, the modelling of regional and temporal peculiarities such as nitrogen immissions or snow-break events has to be improved.