Validation of MODIS-GPP product at 10 flux sites in northern China

Gross primary production GPP is an important variable in studies of the carbon cycle and climate change. The Moderate Resolution Imaging Spectroradiometer MODIS-GPP product MOD17 provides global GPP data for terrestrial ecosystems; however, it is not well validated in China. In this study, an eddy covariance EC system observed GPP at 10 sites in northern China and was used to validate MOD17. The results indicated that MOD17 presents a strong bias in the study region due to the meteorological data, MODIS FPAR fraction of absorbed photosynthetically active radiation MOD15, and the model parameters in the MODIS-GPP algorithm, Biome Parameters Look Up Table BPLUT. Maximum light-use efficiency ϵ0 had the strongest impact on the predicted GPP of the MODIS-GPP algorithm. After using the inputs observed in situ and improving parameters in the MODIS-GPP algorithm, the model could explain 85% of the EC-observed GPP of the sites, whereas the MODIS-GPP algorithm without in situ inputs and parameters only explained 26% of EC-observed GPP.