Global patterns of ecosystem carbon flux in forests: A biometric data-based synthesis

Forest ecosystems function as a significant carbon sink for atmospheric carbon dioxide. However, our understanding of global patterns of forest carbon fluxes remains controversial. Here we examined global patterns and environmental controls of forest carbon balance using biometric measurements derived from 243 sites and synthesized from 81 publications around the world. Our results showed that both production and respiration increased with mean annual temperature and exhibited unimodal patterns along a gradient of precipitation. However, net ecosystem production (NEP) initially increased and subsequently declined along gradients of both temperature and precipitation. Our results also indicated that ecosystem production increased during stand development but eventually leveled off, whereas respiration was significantly higher in mature and old forests than in young forests. The residual variation of carbon flux along climatic and age gradients might be explained by other factors such as atmospheric CO2 elevation and disturbances (e.g., forest fire, storm damage, and selective harvest). Heterotrophic respiration (Rh) was positively associated with net primary production (NPP), but the Rh-NPP relationship differed between natural and planted forests: Rh increased exponentially with NPP in natural forests but tended toward saturation with increased NPP in planted forests. Comparison of biometric measurements with eddy covariance observations revealed that ecosystem carbon balance derived from the latter generated higher overall NEP estimates. These results suggest that the eddy covariance observations may overestimate the strength of carbon sinks, and thus, biometric measurements need to be incorporated into global assessments of the forest carbon balance.