It is necessary to represent the spatial variability of net CO2 exchange between atmosphere and terrestrial ecosystems in carbon cycle models, for accurately simulating and predicting global carbon budget. However, the current state-of-the-art models still fail to reproduce this variability, largely due to the insufficient understanding on the underlying mechanism determining the spatial variability. Here, we compiled a global dataset of annually averaged gross primary productivity (GPP), ecosystem respiration (RE) and net ecosystem exchange (NEE) from 151 eddy flux sites (data length ≥4 years), and 17 critical factors, to explore the relative contribution of climate, soil and traits to the spatial variability of NEE.
Results/Conclusions
The results showed that compared with soil properties and plant traits, climatic factors always contributed mostly to the spatial variability of NEE in both forests and grasslands. However, the exact mechanisms were differential between forests and grasslands. For example, the divergent interactive effects between climate, soil properties and plant traits on GPP and RE were important contributors to the spatial variability of NEE in forests, but not in grasslands. In grasslands, soil properties had differential effect on GPP and RE, resulting in a large contribution to the spatial variability of NEE, which pattern was not found in forests. Our research suggests the differential mechanisms driving the spatial pattern of NEE in forests and grasslands, and highlighted the critical importance of soil properties in determining the carbon fluxes in grasslands, which should be paid more attention in future studies.