ForestSimMPB: A swarming intelligence and agent-based modeling approach for mountain pine beetle outbreaks

Abstract

The widespread outbreaks of Mountain Pine Beetle (MPB) are responsible for infestations of lodgepole pine forests since 1990 in Canada. In British Columbia, this forest insect disturbance has resulted in losses of more than 13 million hectares of pine trees. The complexity of the MPB emergence, aggregation and attack behaviour is captured by this study, using an intelligent agent-based model (ABM) of beetle outbreaks at a local scale of individual trees. Agent-based approach permits simulation of interactions that describe the ecological context in which insect populations spread. Intelligent reasoning is introduced by a swarm intelligence (SI) algorithm integrated with the ABM that depicts indirect communication, collective behaviour and self-organized aggregation of insects in a forest ecosystem. The objectives of this study are the following: 1) to develop ForestSimMPB model that integrates SI and ABM within a geographic information systems (GIS) framework; 2) to implement the proposed model on real datasets to simulate the MPB aggregation and mass attacks on lodgepole pine trees; and 3) to determine the spatial patterns and extents of these attacks. The ForestSimMPB is calibrated by fine tuning two model parameters, and implemented using data from three sites located in the Cariboo Regional District in the central interior of BC. The obtained results demonstrate the aggregation behaviour of MPB to collectively attack lodgepole pines, as well as portray the spatial clustering of dead trees resulting from infestation. Simulation outputs provide analysis and predictions of spatial patterns in the forest landscape structure as a result of a MPB outbreak. The developed model can be used to assist the improvement of methods for prevention and control of MPB disturbances.