Comparing optimal and empirical stomatal conductance models for application in Earth system models
Peter J. Franks,Gordon B. Bonan,Joseph A. Berry,Danica L. Lombardozzi,N. Michele Holbrook,Nicholas Herold,Keith W. Oleson
DOI: https://doi.org/10.1111/gcb.14445
IF: 11.6
2018-10-19
Global Change Biology
Abstract:Earth system models (ESMs) rely on the calculation of canopy conductance in land surface models (LSMs) to quantify the partitioning of land surface energy, water, and CO<sub>2</sub> fluxes. This is achieved by scaling stomatal conductance, g<sub>w</sub> , determined from physiological models developed for leaves. Traditionally, models for g<sub>w</sub> have been semi-empirical, combining physiological functions with empirically determined calibration constants. More recently, optimization theory has been applied to model g<sub>w</sub> in LSMs under the premise that it has a stronger grounding in physiological theory and might ultimately lead to improved predictive accuracy. However, this premise has not been thoroughly tested. Using original field data from contrasting forest systems, we compare a widely used empirical type and a more recently developed optimization-type g<sub>w</sub> model, termed BB and MED, respectively. Overall, we find no difference between the two models when used to simulate g<sub>w</sub> from photosynthesis data, or leaf gas exchange from a coupled photosynthesis-conductance model, or gross primary productivity and evapotranspiration for a FLUXNET tower site with the CLM5 community LSM. Field measurements reveal that the key fitted parameters for BB and MED, g<sub>1B</sub> and g<sub>1M,</sub> exhibit strong species specificity in magnitude and sensitivity to CO<sub>2</sub> , and CLM5 simulations reveal that failure to include this sensitivity can result in significant overestimates of evapotranspiration for high-CO<sub>2</sub> scenarios. Further, we show that g<sub>1B</sub> and g<sub>1M</sub> can be determined from mean c<sub>i</sub> /c<sub>a</sub> (ratio of leaf intercellular to ambient CO<sub>2</sub> concentration). Applying this relationship with c<sub>i</sub> /c<sub>a</sub> values derived from a leaf δ<sup>13</sup> C database, we obtain a global distribution of g<sub>1B</sub> and g<sub>1M</sub> , and these values correlate significantly with mean annual precipitation. This provides a new methodology for global parameterization of the BB and MED models in LSMs, tied directly to leaf physiology but unconstrained by spatial boundaries separating designated biomes or plant functional types.
environmental sciences,biodiversity conservation,ecology