Cell Interaction in Semi-Markov Forest Landscape Models

Abstract: This paper describes a model for incorporating both spatial interactions and succession dynamics within a MOSAIC forest landscape model. MOSAIC uses semi-Markov processes to model the succession dynamics of a forest landscape. A landscape is partitioned into cells and the parameters defining the semiMarkov process associated with each cell are derived from individual tree-based models that run in gap-sized plots within the cell. Parameters vary from cell to cell based on the environmental and terrain conditions of the cells. Ecologically, cells and plots influence their neighbors’ development through processes such as seed dispersal and disease transmission. Neighborhood interactions are modeled through influence arrays that are flexible enough to capture a variety of processes operating at the landscape scale. The influence arrays are similar to the transition functions used in cellular automata models. Integrating tree-based and landscape scale models, as well as ideas from cellular automata and semi-Markov processes, further unifies techniques used in forest landscape models. Simulation examples derived from the H. J. Andrews Forest in Oregon are presented.