Optimization of selected parameters of the forest growth model Biome-BGC (version ZALF) using HOPSPACK

Biome-BGC model is widely used for estimating C, N, and water budgets in terrestrial ecosystems. Like other complex forest models, Biome-BGC contains many eco-physiological and site specific parameters. Estimation of these parameters is critical to Biome-BGC model applications. Many techniques are currently available to parameter estimation including derivative-free inverse optimization techniques. In this work, we tried to optimize some selected Biome-BGC parameters using a derivative-free optimization tool, HOPSPACK, for 27 ICP Forests Level II plots. Rather than using a single objective function, we used nine frequently used objective functions. We concluded that intensive care is required during parameter optimization using such a derivative-free optimization tool because the outcome might be sensitive to the initial parameter values and the sequence of optimization of parameters. Inclusion of more objective functions increased robustness of the outcome.