Use of Eddy Covariance Data to Calibrate the PhotoCent Model for Grassland and Forest Ecosystems

Eddy Covariance data sets have been used to calibrate the photosynthesis version of the DayCent ecosystem model (PhotoCent). Data simulation techniques were used to determine the parameters in the photosynthesis submodel for deciduous and coniferous forests, and arid and humid grasslands. A comparison of the seasonal patterns of gross primary production (GPP) and net ecosystem exchange (NEE) for grassland and forest sites showed that maximum NEE and GPP values were observed in the spring (May and June) with NEE and GPP declining until the end of the growing season at different rates for the different biomes. The model results and observed daily GPP and NEE data sets suggest the major factor which controls this seasonal pattern is the decline in maximum photosynthesis rate during the growing season. This analysis suggests that grasslands, coniferous forests and deciduous forests have biome specific patterns in the rate of decline of maximum photosynthesis rates which are supported by observed field data showing similar declines in maximum photosynthesis rates. A comparison of model simulated GPP, NEE and respiration (RESP) rates with observed data shows that the observed biome specific seasonal patterns in these variables are well represented by the PhotoCent model. A comparison of observed vs. simulated daily GPP, NEE , and RESP shows that GPP has the highest correlation with observed data (r2 ranging from 0.70 to 0.90), and somewhat lower correlations for NEE and RESP (0.5 to 0.7).