What is the simulator of forest biodynamics SIBYLA ?

The name SIBYLA is the acronym formed from the Slovak translation of "Simulator of forest biodynamics"  (Simulátor biodynamiky lesa). The simulator of forest biodynamics (SIBYLA) belongs to the category of tree growth simulators (hereinafter called as a growth simulator). It is a simulator that strives to imitate the behaviour of trees in the context of forest ecosystems. It consists of the set of mathematical models and algorithms that are transformed into an integrated software package SIBYLA Suite. The difference between the growth simulator SIBYLA and a classical forest model (e.g. forest yield tables) can be defined as follows:

A growth simulator is a system, that strives to imitate forest behaviour using the principles of ecosystem and cybernetical modelling. It utilises a very wide range of input conditions and parameters. It simulates different initial forest stand structures starting from even-aged homogeneous stands (pure plantations) of the type of age classes, through differentiated multistoreyed forest, mixed stands and shelterwood systems, up to selection forests. It is able to simulate a wide range of natural conditions defined by ecological (site) classifications in the form of climate, air, and soil characteristics. In addition, it also offers a quite large operating space to make the interventions of a forest manager in the form of various thinning and felling regimes. And besides, a specific economic environment is accounted for inclusive of applied technological techniques. At the same time, growth simulator provides a user with a great variety of output data. Apart from classical production data it also deals with ecological information, such as biodiversity, biomass, fixation of nutrient elements in trees, oxygen production and carbon dioxide consumption. It also covers an economic aspect in the form of assortment structure of produced wood, forest revenues and management costs. To imitate the real forest as faithfully as possible, stochastic principles are applied, i.e. every time the simulation is repeated, the model produces slightly different results. The behaviour of randomness follows the probability principles and functions derived from real forest ecosystems. Thanks to the randomness, the component of theoretical model error can be obtained, and statistical tests of the differences between various scenarios can be performed. The nature of the system is complex, since it utilises a set of various linked models and algorithms of a different nature: allometric equations, regressions, growth curves, mensurational relationships, physical and chemical relationships, production rules, boolean and fuzzy logic, heuristics, planar and spatial geometry, two- and multi-dimensional probability models, etc. Due to this complexity, it is undoubtedly required that the system exists in the form of a computer program. Since the system is characterised by a number of input parameters and a variety of possibilities to define different variants and scenarios, its application is more challenging. Primarily, it is suitable for scientific and educational purposes. 

Yield tables represent a mathematical model, which today describes forest development by the system of mathematical equations. It simulates the development of even-aged homogeneous forest stands (pure plantations) at full density and 100% proportion of a particular tree species in relation to age and site. The site is defined by stand class, or also by stand volume level. Yield tables are resctricted to only one thinning regime, or eventually to a set of pre-defined variant regimes with no possibility to modify them. The outputs are primarily oriented at production aspect of a forest, while usually they are presented in tabular form. Thanks to the model simplicity and to the restricted range of possible variants of a forest, the model does not have to exist as a computer program. It is mainly composed of simple growth curves, or eventually of other mensurational relationships. The model is strictly deterministic, and hence, its character is often normative. Due to the facts that this model does not require many input parameters and is simply applicable, it is primarily used in the forestry practice. 

The differences between the two models are summarised in the following table:

© Copyright doc. Ing. Marek Fabrika, PhD.

© Translated by FORIM - Dr. Ing. Katarína Merganičová