arising from: Lai, J. et al. Nature https://doi.org/10.1038/s41586-024-08050-3 (2024).
Satellite-based optical observations1,2,3 have converged to show average estimates of global contemporary gross primary production (GPP), the photosynthetic assimilation of atmospheric CO 2 by terrestrial vegetation, at 120–140 petagrams of carbon per year (PgC yr−1). This consensus was challenged by a recent modelling study conducted by Lai et al.4 that posits a higher global estimate of 157 ± 8.5 PgC yr−1, revealing a pronounced discrepancy in the tropics where modelled fluxes of 79 PgC yr−1 exceed satellite-derived values by a factor of 1.33–1.85. We show that the global estimate of Lai et al.4 is extrapolated spatially based on a temporal relationship between the ratio of carbonyl sulfide (OCS) to CO 2 fluxes (hereafter, the leaf relative uptake, LRU) and photosynthetically active radiation (PAR) from a single boreal station and also neglects considerable negative influences of vapour pressure deficit (VPD). This finding underscores the need for accounting for both spatial heterogeneity and climatic constraints when upscaling site-level relationship to the global scale, warranting the necessity of a comprehensive worldwide validation for global estimates.