Estimation of global vegetation productivity from 1982 to 2015 with remote sensing data.

The vegetation productivity is an important parameter to reflect vegetation activities, which is important to evaluate the ecological carrying capacity of ecosystem and understand the terrestrial carbon cycle. 8-days gross primary production (GPP) and annual net primary production (NPP) are contained in MODerate Resolution Imaging Spectroradiometer (MODIS) products (MOD17), which are considered to be the first operational data sets for monitoring global vegetation productivity. But some researchers suggest that the MODIS NPP and GPP products are responsive to general trends in the magnitude of NPP and GPP associated with local climate and land use, but tend to be overestimated at low productivity sites and underestimated at high productivity sites. In this paper, we estimated global GPP and NPP from 1982 to 2015 with a light use efficiency model, which is driven by Global LAnd Surface Satellite (GLASS) FPAR and LAI remote sensing data sets、ERA-Interim meteorological data、air temperature and the other variables. The spatial resolutions of estimated global GPP/NPP products are 5km for every 8-days. Global FLUXNET GPP and BigFoot NPP and other data were used to verify the product and MODIS GPP/NPP 1km product was also compared with our results in various time series and space scale. Results suggest that our results have better spatial and temporal continuity in terms of global spatial distribution and interannual variation, the NPP estimation of different vegetation types is basically accurate. In different regions, the simulation accuracy is different, which is highest in temperate region.