Inverting the maximum carboxylation rate (Vcmax) from the sunlit leaf photosynthesis rate derived from measured light response curves at tower flux sites

- A new model is developed to separate the flux GPP into sunlit GPP and shaded GPP.
- A new approach to invert Vcmax using the sunlit GPP is also proposed.
- The inverted Vcmax is validated at six flux sites in Canada.
- The new inversion approach provides a way to obtain Vcmax at the regional scale.

Obtaining the maximum carboxylation rate (Vcmax) has been tedious and time consuming on the field. An algorithm of inverting Vcmax using the sunlit leaf photosynthesis rate was developed and validated at six Canadian flux sites. The sunlit leaf photosynthesis rate was calculated from the sunlit gross primary productivity (GPPsun) obtained through rectangular hyperbola light response curves constructed using the eddy covariance observations. The ability of the rectangular hyperbola method in GPP simulation and sunlit/shaded GPP separation was tested by 1) comparing the rectangular hyperbola simulated GPP (GPPLRC) to the half-hourly GPP derived from eddy covariance measurements (GPPflux) at six flux sites and 2) comparing the separated sunlit/shaded GPP to their counterparts simulated using the Boreal Ecosystem Productivity Simulator (BEPS). The results exhibited good agreements between the GPPLRC and the GPPflux (R2 = 0.74-0.93, and RMSE = 1.71-3.27 μmol m−2 s−1) for all six sites. The sunlit/shaded GPP separated using rectangular hyperbola method and BEPS also agreed closely in both magnitude and the seasonal and diurnal variations. The sunlit GPP obtained with incoming APAR bigger than 900 μmol m−2 s−1 was selected to invert Vcmax through Farquhar's classic photosynthesis model. The mean Vm25 (Vcmax normalized to 25 °C) values during the peak of growing season (day 152-day 243) agreed well with Vm25 values from previous studies at same sites. The four conifer sites shared similar Vm25 during the peak of growing season (22-29 μmol m−2 s−1) though large difference between two deciduous sites was observed (32 μmol m−2 s−1 and 47 μmol m−2 s−1). The inverted Vm25 showed clear seasonal trends at two deciduous forest sites while little seasonal trend was observed at four conifer sites. Decreasing of Vm25 was observed during the drought period at the conifer site TP74 when volumetric soil water content (SWC) dropped to around 0.05 m3/m3 but the controlling function of soil moisture on Vm25 remained unclear. Further studies on the temperature dependency of Vcmax demonstrated that a same set of parameters was enough in describing the temperature response of Vcmax for all sites. Our analysis revealed that the uncertainty of the inverted Vcmax were less than 25% for all the sites, indicating the algorithm developed in this study can be used to retrieve reliable Vcmax.