Nonlinear models based on leaf architecture traits explain the variability of mesophyll conductance across plant species.

Mesophyll conductance ( g m ${g}_{{\rm{m}}}$ ) describes the efficiency with which CO 2 ${\mathrm{CO}}_{2}$ moves from substomatal cavities to chloroplasts. Despite the stipulated importance of leaf architecture in affecting g m ${g}_{{\rm{m}}}$ , there remains a considerable ambiguity about how and whether leaf anatomy influences g m ${g}_{{\rm{m}}}$ . Here, we employed nonlinear machine-learning models to assess the relationship between 10 leaf architecture traits and g m ${g}_{{\rm{m}}}$ . These models used leaf architecture traits as predictors and achieved excellent predictability of g m ${g}_{{\rm{m}}}$ . Dissection of the importance of leaf architecture traits in the models indicated that cell wall thickness and chloroplast area exposed to internal airspace have a large impact on interspecific variation in g m ${g}_{{\rm{m}}}$ . Additionally, other leaf architecture traits, such as leaf thickness, leaf density and chloroplast thickness, emerged as important predictors of g m ${g}_{{\rm{m}}}$ . We also found significant differences in the predictability between models trained on different plant functional types. Therefore, by moving beyond simple linear and exponential models, our analyses demonstrated that a larger suite of leaf architecture traits drive differences in g m ${g}_{{\rm{m}}}$ than has been previously acknowledged. These findings pave the way for modulating g m ${g}_{{\rm{m}}}$ by strategies that modify its leaf architecture determinants.