Global Patterns of Ecosystem C Flux in Forests: a Synthesis of Biometric Measurements

Assessing the carbon budget of terrestrial ecosystems is one of the key issues in current global change analyses. During the past several decades, a number of individual studies have been conducted to examine key parameters of carbon balance in forest ecosystem. These individual measurements offer the possibility to be synthesized to clarify the strength, distribution and mechanisms of carbon sequestration in world's forests. However, most previous syntheses have been derived from eddy-covariance measurements. To the best of our knowledge, key parameters of forest carbon balance obtained from biometric measurements have not yet been synthesized. In this study, we established a global dataset of forest carbon balance by collecting the data from publications which reported the biometric measurements of carbon pool and flux. This dataset contains carbon pool and flux measurements from 243 sites, 304 site/years, and 81 literatures. The main results are summarized as follows: (1) Across all sites, the average Net Ecosystem Productivity (NEP) in global forest ecosystem was 2.47 Mg C ha-1 yr-1, Net Primary Productivity (NPP) was 6.66 Mg C ha-1 yr-1, and Heterotrophic respiration (Rh) was 4.52 Mg C ha-1 yr-1. The order of NEP values was Subtropical > Temperate broadleaves > Temperate conifer > Tropical > Boreal. NPP, GPP, Rh, and Re decreased with increases in latitude. However, NEP exhibited an initial increase and then declined along the latitudinal gradient. The variation in NEP was large in the middle-latitudes. The NPP, GPP, Rh, and Re had positive relationships with mean annual temperature, and increased firstly and declined afterward with mean annual precipitation. NEP peaked in moderate climate areas. Stand age was also very important in determining forest carbon budget. The forests in middle age classes had the highest NEP, while old forests were only a small carbon sink. (2) Our results show a significant correlation between productivity and respiration in forest ecosystems. The relationships between NPP, Rh and NEP had different patterns in natural and planted forests. In natural forests, the relationship between Rh and NPP could be fitted by an exponential function and this curve was a reflection of the carbon balance status for all sites . It indicates that NEP correspond an initial increase with NPP and then decrease. In planted forests, NEP increased linearly with NPP over time, which suggests that planted forests have great potential to serve as a large carbon sink in global carbon cycle. (3) We also compared several key carbon parameters that were derived from biometric and eddy-flux measurements from the sites that both datasets were available. The NEP values from the two approaches exhibited strong correlations, but also great variations when NEP was within the range of 0-5 Mg C ha-1 yr-1. In most forest types, the average NEP based on eddy-covariance measurements was greater than that based on biometric measurements, while the average GPP and Re were smaller. Besides reported here, our study also produced other information useful for estimating the global forest carbon budget based on ecosystem biometrics, instead of conventional country-level inventories.