Association between sap flow-derived and eddy covariance-derived measurements of forest canopy CO2 uptake.

The carbon sink intensity of the biosphere depends on the balance between gross primary productivity (GPP) of forest canopies and ecosystem respiration. GPP, however, cannot be directly measured and estimates are not well constrained. A new approach relying on canopy transpiration flux measured as sap flow, and water-use efficiency inferred from carbon isotope analysis (GPPSF ) has been proposed, but not tested against eddy covariance-based estimates (GPPEC ). Here we take advantage of parallel measurements using the two approaches at a semi-arid pine forest site to compare the GPPSF and GPPEC estimates on diurnal to annual timescales. GPPSF captured the seasonal dynamics of GPPEC (GPPSF  = 0.99 × GPPEC , r(2)  = 0.78, RMSE = 0.82, n = 457 d) with good agreement at the annual timescale (653 vs 670 g C m(-2)  yr(-1) ). Both methods showed that GPP ranged between 1 and 8 g C m(-2)  d(-1) , and the GPPSF /GPPEC ratio was between 0.5 and 2.0 during 82% of the days. Carbon uptake dynamics at the individual tree scale conformed with leaf scale rates of net assimilation. GPPSF can produce robust estimations of tree- and canopy-scale rates of CO2 uptake, providing constraints and greatly extending current GPPEC estimations.