A scalable simulator for forest dynamics

Abstract

Models of forest ecosystems are needed to understand how climate and land-use change can impact biodiversity. In this paper we describe an individual-based, spatially-explicit forest simulator with full accounting of both landscape context and the fine-scale processes that influence forest dynamics. Unfortunately, performing realistic forest simulations of such models is computationally infeasible. We design efficient algorithms for computing seed dispersal and light, using a plethora of techniques. These include hierarchical spatial decomposition, monopole approximation and utilizing the graphics hardware for fast geometric computations. These algorithms allow us to simulate large landscapes for long periods of time.