Neuronal Network Based Modelling of Demand and Competing use of Forestry Commodities for Material and Energy use

A methodology for development of scenarios for multiple forestry commodities quantities and prices through a nonlinear autoregressive neuronal network model with additional exogenous input parameters is presented. By mapping all possible interdependencies between forestry commodities and commodity prices, this approach shall enable to model the demand for different commodities and competing use for these commodities.The presented model performs good in terms of input-output correlation (R=0,99) for all variables combined. The results point to the conclusion that the functional relation between CO2-emission scenarios and biomass use can be captured by the modeling framework.