Extended Data Fig. 3: Estimates of GPP (a, b) and soil respiration (Rsoil) (c, d) based on different methods for both ambient (a, c) and elevated (b, d) CO2 treatment at EucFACE.
From: The fate of carbon in a mature forest under carbon dioxide enrichment
For estimates of GPP, we compared the model simulated total GPP of overstorey and understorey (GPPo and GPPu, respectively), with the sum of data-driven estimates of net primary production (NPP) and autotrophic respiration (Ra), which include NPP of overstorey leaf (NPPol), stem (NPPstem), fine root (NPPfroot), intermediate root (NPPiroot), coarse root (NPPcroot), twigs, barks and seeds (NPPother), understorey aboveground (NPPua), leaf consumption by insects (NPPins), and respiratory fluxes of overstorey leaf (Rol), stem (Rstem), root (Rroot), understorey aboveground (Rua), growth (Rgrow), and volatile carbon emission (VC). For estimates of Rsoil, we compared direct estimates of Rsoil scaled up from soil chamber measurements, with the sum of litterfall and independent estimates of root respiration (Litter + Rroot), assuming no net change in soil carbon stock over time. Here litterfall was inferred from the NPP of overstorey leaf (NPPol), fine root (NPPfroot), intermediate root (NPPiroot), twigs, barks and seeds (NPPother), understorey aboveground (NPPua), and frass production (Frass). These evaluations provide independent mass balance checks of the estimated ecosystem carbon budget. Each colour represents a flux variable. Each point and vertical line represent treatment mean and standard deviation, respectively, based on plot-level estimates of the aggregated flux (n = 3). Values were normalized by a linear mixed model with pre-treatment LAI as a covariate to account for pre-existing differences.
