Description and test of a simple process-based model of forest growth for mixed-species stands

Process-based models of forest growth have been discussed for decades, but their utility as management tools has only recently begun to increase. Ideally this type of model would be tested by treating it as a complex hypothesis that relates independently measured parameters to predicted responses. This approach provides a test of the structure and the parameterization of the model that is not possible if the model has been calibrated (or tuned). We conduct such a test of a new model of forest production and allocation. The test uses a series of plots located across complex terrain in northern Idaho, USA. The production model scales leaf-level gas exchange to the canopy, and is parameterized with foliar nitrogen, leaf area index (LAI), and canopy structural parameters. New biomass is allocated such that tree allometry is maintained while foliage and branches turn over. A simple approach combines allometric equations across species in mixed-species stands. Predictions of volume increment for a 10-year period were higher than measurements, but the two were significantly correlated. The discrepancy was reduced when leaf area index was estimated from canopy light transmission rather than allometric equations. We argue that one likely reason for the overprediction is the occurrence of soil water deficits in the summer. A sensitivity analysis showed that estimates of production were most sensitive to leaf area index and canopy average foliar nitrogen content, but much less to other parameters. We conclude that comparing the model to observed data reveals shortcomings that might have been hidden if the parameters had been tuned. The mixed-species allometric constraint provides a new tool for modeling biomass allocation in mixed-species stands.