Revising optimization theory of stomatal behavior to estimate canopy-scale exchanges of mass and energy under increased atmospheric CO₂ and warming

Sari Palmroth, Duke University

Co-investigators: Gabriel Katul and Ram Oren, Duke University

Abstract

A predicted outcome of the increase in atmospheric greenhouse gases by climate models is an unchanging atmospheric relative humidity (RH) but an exponential increase in vapor-pressure deficit (D) due to warming. Predictions of water cycling in future climates will be sensitive to how stomatal conductance (g) responds to increased D under elevated atmospheric CO₂ concentration (c_a) and air temperature (T_a). Our objectives are to (1) develop and test a modeling framework based on optimization theories of leaf-level g that captures the simultaneous responses to elevated c_a and T_a, and (2) scale the results from these leaf-level theories to the canopy in a form usable for coarse-scale models

This proposal seeks to synthesize available results from a large number of experiments to address three inter-related questions pertinent to the two objectives: [1] Can stomatal optimization framework be developed to provide quantitative predictions about the effects of elevated c_a and increase in D on g? [2] Can the optimization results scale from leaf to canopy via coupled radiative transfer and physiological principles, and turbulent transport theories? And [3], if successful, are there emergent functional responses of canopy-level fluxes to D and c_a that can be used in operational ecological and hydrologic models? The proposed research will be conducted at Duke University.

To address the first question, we will extend earlier theories of Cowan & Farquhar (1977) to explicitly treat the time-scale of optimization and derive explicit expressions to predict the responses of g to changes in c_a and D. Gas exchange data available from Free Air CO₂ Enrichment facilities, chamber-based experiments, and NICCR-funded manipulation experiments will be combined to assess the applicability of these revised theories. To address the second and third questions, version D1 of the FLUXNET database, now containing over 16 million half-hourly data points, will be employed. The leaf-to-canopy scaling will utilize an existing model (CANVEG) developed and tested in NIGEC projects. This work aims to replace the empirical formulation of g response to c_a and D with one in which the relationships between g and the climatic drivers are not assumed but are emergent properties of the optimization. As a product from this project, the functions obtained from synthesis and analysis of leaf-level conductance responses will be made available to the community through CDIAC (Carbon Dioxide Information Analysis Center) and through the FLUXNET network.