Parallel Computing of Forest Biotic Dynamics

Abstract

This poster presents a parallel-computing framework developed recently through the collaborative efforts between the College of Food, Agricultural, and Natural Resource Sciences, the Institute of the Environment, and the Supercomputing Institute at the University of Minnesota. The goal of this project is to address the challenges encountered in the modeling of biotic dynamics in a forest region across different time and spatial scales. By implementing the parallel computing framework on Itasca, we have been able to reduce the computing time from 46 days (if the simulation runs on a single desktop) to half a day by using 8 nodes for a study of northern Minnesota, Wisconsin, and Michigan with 1 km grid resolution. This will allow us to attack computationally challenging problems, such as assessing the impact of critical events like the 1999 BWCA blowdown and finer resolution events such as controlled burns on future forest productivity and stability.