| Terrestrial ecosystems play a critical role in mitigating the global warming, because they accounted for about 30% of the anthrophic CO2 emissions. The ability of sequestrating CO2 can be characterized by the net ecosystem exchange (NEE) at the ecosystem scale and the carbon (C) sink/source at the regional scale. Whether the global terrestrial ecosystems will continuously absorb CO2 in future can only be predicted by models. However, current state-of-art models generally failed to represent the interannual variability (IAV) in NEE at the ecosystem scale, and resulted in large uncertainty in C sink/source at the regional scale. It is impossible to imporve the model performance without studying the underlying mechanisms of IAV in NEE or the main sources of uncertainty in C sink/source. Therefore, we analysized the flux data from 65 sites all over the world, aiming to evaluate the effects of biotic and climatic drivers on the IAV in NEE, and synthesized the research on net primary productivity (NPP) and C sink of China published in recent 30 years, aiming to investigate the sources of inter-model uncertainty. The main findings were as follows:(1) In order to study the effects of biotic and climatic drivers on IAV in NEE at ecosystem scale, we partitioned the IAV of NEE into biotic (BE) and climatic effects (CE) at a subtropic plantation in Jiangxi Province, China. Then we quantified the relative contributions of BE and CE to the IAV in NEE, and investigated the relationship between BE and CE. The results showed that the contribution of BE to the IAV in NEE increased significantly as the temporal scale got longer, which was 9.2%±10.7%,17.5%±13.3%,30.5%±17.4%,39.3%±16.7% and 47.4% at daily, weekly, monthly, seasonal and annual scales, respectively, while the contribution of CE decreased as the temporal scale got longer, but it still accouted for 52.2% of the IAV in NEE. The CE caused the plantation to absorb less C from 2003 to 2009 with a decrease of 36.2 g C m-2 per year, while the biotic effect offset the climatic one to a large degree. The negative correlation between BE and CE (r2=0.80, P＜0.01) made the IAV in NEE relatively small, suggesting that the plantation might be a stable C sink (-333±47g C m-2 yr-1).(2) In order to investigate the different patterns of BE and CE among ecosystems, we applied additive model, which is more sutiable for multi-site comparison, and model averaging technique based on Akaike weight to the flux data from 65 sites in the world. The results showed that at the annual scale, the relative contributions of BE (CnBENEE) and CE to the IAV in NEE were 57%±4% and 43%±14%, respectively, suggesting that the BE was more important than CE at the ecosystem scale. The inter-site differences of CnBENEE were mainly controlled by the water stress. Among the ecosystems experienced water stress, the ChBENEE decreased 0.18% with every 1 mm decrease of available water, indicating that the ecosystem C cycling might graduately lose the ability of self-regulating when the water stress became more severe. On the contrary, the non-climatic factors such as vegetation type and disturbance regime had no significant effects on CnBENEE.The BE explained 72% of the inter-site varation of IAV in NEE, indicating that it was the major factor determing the spatial differences among ecosystems. There were significant positive correlations between BE and CE in grasslands and dry ecosystems (r>0.45, P<0.05), suggesting that the biotic responses of ecosystems might accelerate the effects of climatic variations in these ecosystems.(3) In order to clarify whether the BE resulted from climatic variation indirectly, we applied the structural equation model to the 65-site’s flux data and investgate the direct and indirect effects of climate on C fluxes at inter-annual scale. The results showed that the direct effect of PAR was only found in the ecosystems with plenty water (W1) in which it was significantly correlated with NEE (r=-0.24, P＜0.05), though it was important in simulating the seasonality of gross primary productivity (GPP). The temperature directly affected NEE by enhancing the ecosystem respiration (RE). The direct effect of water condition became stronger when the ecosystem got drier. The indirect effects of climate were mainly on GPP in evergreen needleleaf forests (ENF), grasslands and croplands, and on RE in deciduous broadleaf forests (DBF). The indirect effect of climate was more significant in grasslands than other biomes, indicating that the climatic and biotic factors were more related in grasslands. The indirect effect of climate was also more significant in ecosystems with moderate water than those with plenty or insufficient water, which pattern might be related to the differentces in lag effects of climate on C fluxes. In general, the direct and indirect effects of climate did not well explained the IAV in NEE (16%-55%), and the correlations between climatic variables and physiological parameters were weak, suggesting that the BE mainly came from the factors other than climate.(4) In order to find the main sources of uncertainty in the estimations of NPP and C sink in China, we synthesized totally 193 estimations of NPP or C sink in China or its main biomes (forests, shrublands, grasslands and croplands). The results showed that the NPP and C sink of China was 3.35±1.25 Gt C yr-1 and 0.14±0.094 Gt C yr-1 respectively. Forests, shrublands, grasslands and croplands accounted for 33%,8.3%, 24% and 27% of total NPP, and 40%,19%,25% and 18% of total C sink, respectively. The discrepancy of model structures might be the main source of inter-model uncertainty. For NPP, the largest uncertiany was ntroduced by light use efficiency (LUE) models (50%), in which the GloPEM model had the largtest uncertainty (45%). This result suggested that the inter-model uncertainty would be relatively large when there lacks the constraints of land observations. The interannual patterns of NPP were consistent among studies, which showed that the annual NPP increased by a rate of 0.01103 Gt C (r2=0.37, P=0.004). For C sink, the process models not only introduced the largest uncertainty (76%), most of them also underestimated the C sink in China, which might be the results of lacking nitrogen addition in the models. In addition, atmospheric CO2 concentration was also an important drivers.of C sink in Cina. Land cover and land use change might have profound effects on C sink, but its magnitude was with large uncertainty (217%). Overall, the productivity, C sink, and the ability of mitigating the increasing atmospheric CO2 concentration of terrestrial ecosystems in China were similar to the global averages.Our research highlighted the importance of BE to IAV in NEE, and found the major factors resulted in the differences of CnBENEE being the water stress. The investgation of the relationships between BE and CE revealed the responses of C cycling to climatic variations. In most sites, the BE did not derive from the climatic variables, suggesting that further non-climatic drivers should be included in the models in order to improve the model performace, which purpose can not be achieved by merely regulating the function forms. The uncertainty analysis of the estimations of NPP and C sink in China revealed the factors causing the discrepancies among studies and is valuable to evaluating regional C fluxes accurately and precisely. The land observations might be critical to the estimation of NPP. As for C sink, further research on underlying mechanisms of dynamics is as necessary as the field observations. Moreover, we should focus on the model structures of heterotrophic respiration in the future, and draw lessons from the study of IAV in NEE at the ecosystem level. |
PDF Full Text Request