Calibrating a process-based simulation model for the Acadian forest region

Abstract

Climate change is projected to have profound impacts on Canada’s Acadian Forest Region (AFR). However, large uncertainties arising from climate change and increasing disturbance activity pose challenges for forest management decisions. Process-based (mechanistic) simulation models offer a means by which vulnerabilities and different management strategies can be tested under multiple climate and disturbance regimes. However, applying these complex models can be dauting, especially for novice modelers and forest practitioners; nonetheless, this complexity is increasingly necessary to more realistically project changes in forest growth and composition, ecosystem services, biodiversity, disturbance regimes, and the spread of forest pests. Here, we present a methodology for calibrating and validating iLand (v1.1.1), a landscape-scale, process-based forest model that offers a novel approach for assessing the feedback between individual trees and their environment (ecosystem processes, climate, and disturbance). For the first time, 18 tree species were parameterized and calibrated for the AFR and model performance was evaluated against independent field observations at the tree population and stand level. iLand was able to accurately emulate the dynamics of individual tree species populations as well as the succession of mixed-species forest stands across a range of soil conditions and is now ready to be used to simulate future forest dynamics of the AFR. We also discuss calibration method selection and the potential impacts of model and project structure in relation to our project. As the accessibility and usability of process-based forests models increases, our work provides a unique case study for forest managers looking to expand their toolbox.