Modeling gross primary production and ecosystem respiration in a semiarid grassland of Mongolia

We constructed an empirical model in which the rates of gross primary production (GPP) and ecosystem respiration (R-eco) were calculated using meteorological, soil and plant variables, based on our previous measurements of carbon dioxide (CO2) fluxes using a closed-chamber technique in Mongolian grasslands. In the previous studies, we determined GPP and R-eco using transparent and opaque chambers and simultaneously measured the environmental variables inside and beside the chambers. These measurements allowed us to quantify direct and accurate relationships between the rates of GPP and R-eco and their controlling factors. The GPP, R-eco and their difference net ecosystem CO2 exchange (NEE) were computed in a semiarid grassland site of Mongolia for the growing seasons of 2010 and 2011 using observed values of photosynthetically active radiation, air and soil temperatures, vapor pressure deficit, soil water content and estimated values of aboveground biomass from a remotely sensed vegetation index. Model performance was validated by comparing the modeled and observed NEE values using an eddy covariance technique. Results showed that the model successfully reproduced the magnitude and seasonal variations of the observed NEE, and regression analysis showed reasonable agreement in both years. Cumulative rates of GPP and R-eco during the plant-growing season (May-September) were 200.3 and 210.5gcarbon (C) m(-2) in the dry year (2010) and 342.3 and 300.1gC m(-2) in the wet year (2011), respectively. This result indicates that drought was more effective in reducing CO2 uptake by the plant than in reducing ecosystem respiration. Consequently, the grassland ecosystem was a net carbon source during the growing season in 2010 and a net carbon sink in 2011.