Global patterns and predictors of stem CO2 efflux in forest ecosystems

Stem CO2 efflux (E-S) plays an important role in the carbon balance of forest ecosystems. However, its primary controls at the global scale are poorly understood and observation-based global estimates are lacking. We synthesized data from 121 published studies across global forest ecosystems and examined the relationships between annual E-S and biotic and abiotic factors at individual, biome, and global scales, and developed a global gridded estimate of annual E-S. We tested the following hypotheses: (1) Leaf area index (LAI) will be highly correlated with annual E-S at biome and global scales; (2) there will be parallel patterns in stem and root CO2 effluxes (R-A) in all forests; (3) annual E-S will decline with forest age; and (4) LAI coupled with mean annual temperature (MAT) and mean annual precipitation (MAP) will be sufficient to predict annual E-S across forests in different regions. Positive linear relationships were found between E-S and LAI, as well as gross primary production (GPP), net primary production (NPP), wood NPP, soil CO2 efflux (R-S), and R-A. Annual E-S was correlated with R-A in temperate forests after controlling for GPP and MAT, suggesting other additional factors contributed to the relationship. Annual E-S tended to decrease with stand age. Leaf area index, MAT and MAP, predicted 74% of variation in E-S at global scales. Our statistical model estimated a global annual E-S of 6.7 +/- 1.1PgCyr(-1) over the period of 2000-2012 with little interannual variability. Modeled mean annual E-S was 71 +/- 43, 270 +/- 103, and 420 +/- 134gCm(2)yr(-1) for boreal, temperate, and tropical forests, respectively. We recommend that future studies report E-S at a standardized constant temperature, incorporate more manipulative treatments, such as fertilization and drought, and whenever possible, simultaneously measure both aboveground and belowground CO2 fluxes.