Improved estimates of global terrestrial photosynthesis using information on leaf chlorophyll content

The terrestrial biosphere plays a critical role in mitigating climate change by absorbing anthropogenic CO₂ emissions through photosynthesis. The rate of photosynthesis is determined jointly by environmental variables and the intrinsic photosynthetic capacity of plants (i.e. maximum carboxylation rate; $$ ). A lack of an effective means to derive spatially and temporally explicit $$ has long hampered efforts towards estimating global photosynthesis accurately. Recent work suggests that leaf chlorophyll content (Chl_leaf ) is strongly related to $$ , since Chl_leaf and $$ are both correlated with photosynthetic nitrogen content. We used medium resolution satellite images to derive spatially and temporally explicit Chl_leaf , which we then used to parameterize $$ within a terrestrial biosphere model. Modelled photosynthesis estimates were evaluated against measured photosynthesis at 124 eddy covariance sites. The inclusion of Chl_leaf in a terrestrial biosphere model improved the spatial and temporal variability of photosynthesis estimates, reducing biases at eddy covariance sites by 8% on average, with the largest improvements occurring for croplands (21% bias reduction) and deciduous forests (15% bias reduction). At the global scale, the inclusion of Chl_leaf reduced terrestrial photosynthesis estimates by 9 PgC/year and improved the correlations with a reconstructed solar-induced fluorescence product and a gridded photosynthesis product upscaled from tower measurements. We found positive impacts of Chl_leaf on modelled photosynthesis for deciduous forests, croplands, grasslands, savannas and wetlands, but mixed impacts for shrublands and evergreen broadleaf forests and negative impacts for evergreen needleleaf forests and mixed forests. Our results highlight the potential of Chl_leaf to reduce the uncertainty of global photosynthesis but identify challenges for incorporating Chl_leaf in future terrestrial biosphere models.