Lawrence Berkeley National Laboratory

A striking incongruity has long persisted in the modeling framework typically used to predict CO₂ and water vapor exchange between land plants and the atmosphere across scales. Generally, photosynthetic CO₂ demand is estimated using process-based models of biochemistry, but the biophysical stomatal constraint on photosynthesis and transpiration (g_sw) is estimated using "black box" empirical or optimization-based models. Empirical models of g_sw can only be parameterized in the domain of the training data, limiting confidence in predictions made outside that domain; optimization-based models rely on eco-evolutionary "goal functions" about which there remains poor consensus. To resolve this incongruity, we present a novel process-based model for g_sw with parameters that all have biophysical meaning, and of which only two require empirical fitting, thus ensuring tractability for application in land-surface models (LSMs). The model successfully reproduces variation in g_sw diurnally, globally, and in relation to soil drought and when drought and heat co-occur. The model also has greater functionality than previous models, by predicting stomatal closure under soil drought, the effect of variations in soil-leaf hydraulic conductance, stomatal closure in response to soil and atmospheric drought in darkness, and stomatal opening at high temperatures in both low and high light. With structure and parameters based on physiological processes, this model can translate continuing improvement in understanding of underlying biophysical and molecular genetic causes into predictions for carbon and water exchange, offering greater confidence for predicting the influence of stomata on land-surface exchanges of mass and energy in future climates.