Testing a mechanistic carbon balance model against observed tree growth

Abstract

A forest ecosystem model, FOREST-BGC (Biogeochemical cycles), was used to simulate the 5-year growth increments of 177 interior Douglas-fir (Pseudotsugamenziesii var. glauca (Beissn.) Franco) trees growing in uneven-aged stands. It was neither practical nor possible to develop a complete carbon balance for each tree on first principles, owing to both the cost and difficulty in obtaining carbon balance data for individual trees. Therefore, to calibrate the model carbon allocation parameters, maintenance respiration coefficients of the stems and roots, and the stomatal and mesophyll conductances, were used to maximize the fit between predicted and measured growth for one young, vigorous open-grown tree. The remaining 176 trees were simulated without additional calibration. These parameters were believed to be species specific and were found to be appropriate for a range of sites, tree sizes, and stand densities. A competition factor was added to modify the microclimate of each tree by reducing two driving variables (precipitation and radiation) proportionally to changes in their canopy interception due to the leaf area of the neighboring trees. This method enabled FOREST-BGC to reliably simulate actual growth in uneven-aged stands (R2 = 0.95). Although the model parameters remained constant, the simulated carbon balances were dynamic, with the total maintenance respiration increasing with an increase in tree size and (or) competitive stress.