Modeling Canopy Carbon and Water Fluxes Using a Multilayered Model over a Temperate Meadow in Inner Mongolia

To incorporate canopy vertical structure in a process-based model over a temperate meadow, a multilayered model estimated canopy carbon flux (F-c) and water flux (LE) was applied by comparing with eddy covariance measurements in Inner Mongolia, China. Simulations of diurnal, seasonal CO2 and H2O fluxes and model sensitivity to parameters and variables were analyzed. The results showed that the model underestimated F-c and LE by about 0.6% and 5.0%, respectively. It was able to simulate the diurnal and seasonal variation of F-c and LE and performed well during the day and in the growing season, but poorly at night and early in the growing season. F-c was more sensitive to the leaf nitrogen content distribution coefficient and maximum catalytic activity of Rubisco, whereas LE showed greater sensitivity to the stomatal conductance parameter a(1), empirical coefficient of stomatal response to saturated vapor pressure difference V-pds0, and minimum stomatal conductance of CO(2)g(sc0). The response of F-c to environmental factors was ranked as air CO2 concentration (C-a) > air temperature (T-a) > photosynthetically active radiation (PAR) > soil water content (theta(sm)) > vapor pressure deficit (VPD) > wind speed (u(0)). The response of LE to environmental factors was ranked as T-a > VPD > theta(sm)> PAR> C-a> u(0). The response of LE to vegetation characteristic parameters was greater than that of F-c.