Temporal constraints on leaf-level trait plasticity for next-generation land surface models

Background and Aims Dynamic global vegetation models (DGVMs) are essential for quantifying the role of terrestrial ecosystems in the Earth's climate system, but struggle with uncertainty and complexity. Eco-evolutionary optimality (EEO) theory provides a promising approach to improve DGVMs based on the premise that leaf carbon gain is optimized with resource costs. However, the timescales at which plant traits can adjust to environmental changes have not yet been systematically incorporated in EEO-based models. Our aims were to identify temporal constraints on key leaf photosynthetic and leaf functional traits, and develop a conceptual framework for incorporation of temporal leaf trait dynamics in EEO-based models.Methods We reviewed the scientific literature on temporal responses of leaf traits associated with stomata and hydraulics, photosynthetic biochemistry, and morphology and lifespan. Subsequent response times were categorized from fast to slow considering physiological, phenotypic (acclimation) and evolutionary (adaptation) mechanisms. We constructed a conceptual framework including several key leaf traits identified from the literature review. We considered temporal separation of dynamics in the leaf interior to atmospheric CO2 concentration (ci:ca) from the optimal ci:ca ratio [chi(optimal)] and dynamics in stomatal conductance within the constraint of the anatomical maximum stomatal conductance (gsmax). A proof-of-concept was provided by modelling temporally separated responses in these trait combinations to CO2 and humidity.Key Results We identified 17 leaf traits crucial for EEO-based modelling and determined their response mechanisms and timescales. Physiological and phenotypic response mechanisms were considered most relevant for modelling EEO-based trait dynamics, while evolutionary constraints limit response ranges. Our conceptual framework demonstrated an approach to separate near-instantaneous physiological responses in ci:ca from week-scale phenotypic responses in chi(optimal), and to separate minute-scale physiological responses in stomatal conductance from annual-scale phenotypic responses in gsmax.Conclusions We highlight an opportunity to constrain leaf trait dynamics in EEO-based models based on physiological, phenotypic and evolutionary response mechanisms.