Modeling of incidental cuttings in tree growth model SIBYLA

Bonn

2007

S. 57-67

9 Abb., 5 Tab., 7 Lit. Ang.

Unselbständiges Werk

200141244

The paper deals with modeling of incidental cuttings as consequence of abiotic and biotic injurious agents (windstorm, snow damages, icing damages, bark miners, timber borers, defoliators, wood-destroying fungus, air pollutants, drought, and fires). The model construction is presented in the first part. Software solution is presented in the second part. The model is composed from 2 parts. The first part is based on statistical approach. Data covered all Slovak area has been utilized for mentioned purpose. They are data from stand forest inventory and management records. Data are composed from 3 forest districts for all 47 Slovak Forest Eco-regions. Totally 89707 stands with total area 388830 hectares have been processed. It is approximately 19% of Slovak forest area. Data includes all necessary stand and tree species information, together with information about incidental cuttings (year of cutting, type of injurious agent, amount of cutting). We utilized data from period 2000-2005. We applied factor analysis and regression analysis for estimation of probability of incidental cutting and amount of incidental cutting (together with standard deviation). In first step we select type of injurious agent from frequency table for individual forest eco-regions and tree species. Random selection with probability proportional to frequency is applied. Then we estimate mean probability of selected injurious agent, together with amount of incidental cutting and its standard deviation. Values are selected from statistical classification table on basis of Forest Eco-region, category of site and tree species. The values are modified by relative differences between classification factors and real factors. Factors are derived from factor analysis and they are composed from linear combination of following variables: forest vegetation zone, aspect, slope, age, stand density, percentage of tree species, site index, and relation between mean height and mean diameter. Then modified probability of incidental cutting (risk) is compared with random number (within 0-100%). If probability is higher than random number, then incidental cutting is applied in the simulation. Amount of cutting is generated from Gauss distribution defined by mean amount of cutting and its standard deviation. The second part of the model is based on fuzzy logic rules. The model specifies individual trees for incidental cutting. Three different selection rules are applied: individual (scattered), spotted (expanded from focal point), and local (for specified sub-area). Type of rules depends to type of injurious agent. Individual selection is based on fuzzy functions and their combination (or, and). Fuzzy functions are composed on basis of following tree variables: tree diameter, tree height, coefficient of slightness, crown diameter, crown length, coefficient of crown shape, tree vitality measured by crown surface, bio-sociological tree status, tree competition pressure, and tree quality. Selected trees are defined as trees for incidental cutting.