Uncertainty Source of Modeled Ecosystem Productivity in East Asian Monsoon Region: A Traceability Analysis

East Asian monsoon region, benefits from sufficient water-heat availability and increasing nitrogen deposition, represents significantly higher net ecosystem productivity than the same latitudes of Europe-Africa and North America. A better understanding of major contributions to the uncertainties of terrestrial carbon cycle in this region is greatly important for evaluating the global carbon balance. This study analyzed the key carbon processes and parameters derived from a series of terrestrial biosphere models. A wide range of inter-model disagreement on GPP was found in China's subtropical regions. Then, this large difference was traced to a few traceable components included in terrestrial carbon cycle. The increase in ensemble mean GPP over 1901-2010 was predominantly resulted from increasing atmospheric CO2 concentration and nitrogen deposition, while high frequent land-use change over this region showed a slightly negative effect on GPP. However, inter-model differences of GPP were mainly attributed to the baseline simulations without changes in external forcing. According to the variance decomposition, the large spread in simulated GPP was well explained by the differences in leaf area index (LAI) and specific leaf area (SLA) among models. In addition, the underlying errors in simulated GPP propagate through the model and introduce some additional errors to the simulation of NPP and biomass. By comparing the simulations with satellite-derived, data-oriented and observation-based datasets, we further found that GPP, vegetation carbon turn-over time, aboveground biomass, LAI and SLA were all overestimated in most of the models while biomass distribution in leaves was significantly underestimated. The results of this study indicate that model performance on ecosystem productivity in East Asian monsoon region can be improved by a more realistic representation of leaf functional traits.