To gain a deep understanding of the soil-plant-atmosphere continuum, it is crucial for land surface models to consider plant-related processes. This study focuses on the role of leaf stomatal conductance, which is central to photosynthesis and transpiration. Previous research showed that existing stomatal conductance schemes yield divergent results within a land surface model (LSM), implying a knowledge gap. We further the investigation on the performance of five representative stomatal conductance schemes using a full-fledged LSM, VIC+, and conduct a comparison study over a period of one to two years at two U.S. sites, Blodgett and Duke. Our results reveal substantial differences in modeling outcomes for stomatal conductance, leaf water potential, and leaf CO2 concentration. To gain a better understanding of the divergence among the modeling results, we examine the reasonableness of hourly output variables. Boundaries for unreasonable results are first established using published data for similar conditions, combined with physical reasoning, and judgements thus circumventing limitations of the lack of observational data. Based on them, large portions of the output from these schemes are deemed unreasonable. Treating these schemes as equally plausible or of the same level of credence, we present a new approach in which the stomatal conductance is estimated by simultaneously using multiple plausible expressions derived from two different schemes. This results in the introduction of an additional variable that binds two schemes into a robust one. This new scheme gives a significantly lower percentage of unreasonable variable combinations in its outputs, demonstrating its effectiveness.

A modeling framework is presented for hydrological modeling to more accurately describe the water, energy, and carbon cycles and their interactions with participating processes. This framework extends the modeling strategy presented in Luo et al. (2013) by simultaneously using multiple plausible expressions, derived from different perspectives, in representing the same processes, and enforcing them together with an optimality rule and a semi-empirical expression for plant CO2 uptake. The objectives are to reduce unconstrained free variables, mitigate parameter or variable equifinality, reduce result uncertainties, and ultimately increase the model robustness and predictability. For demonstration, the least cost optimality theory from Prentice et al. (2014), after extended to include water-limited conditions, is combined with the updated semi-empirical Ball-Berry-Leuning formulation (Tuzet et al., 2003). These two expressions are combined with other multiple expressions adopted for hydrological modeling. This framework is incorporated into both VIC+ and a modified DHSVM hydrological models with each applied to two different sites. Numerical studies are performed that using three approaches which only differ in the stomatal conductance modeling, namely, one uses the extended Prentice, one the semi-empirical, and the new framework that uses both. Results show that although all three approaches give reasonable estimates of limited measured fluxes, the present modeling framework gives much more reasonable estimates in the stomatal conductance and in other major model variables, and it also results in giving a relationship between carboxylation and transpiration that is consistent with observations. This modeling framework is general and can be adopted for other fields of study.

A modeling framework is presented for hydrological modeling to more accurately describe the water, energy, and carbon cycles and their interactions with participating processes. This framework extends the modeling strategy presented in Luo et al. (2013) by simultaneously using multiple plausible expressions, derived from different perspectives, in representing the same processes, and enforcing them together with an optimality rule and a semi-empirical expression for plant CO2 uptake. The objectives are to reduce unconstrained free variables, mitigate parameter or variable equifinality, reduce result uncertainties, and ultimately increase the model robustness and predictability. For demonstration, the least cost optimality theory from Prentice et al. (2014), after extended to include water-limited conditions, is combined with the updated semi-empirical Ball-Berry-Leuning formulation (Tuzet et al., 2003). These two expressions are combined with other multiple expressions adopted for hydrological modeling. This framework is incorporated into both VIC+ and a modified DHSVM hydrological models with each applied to two different sites. Numerical studies are performed that using three approaches which only differ in the stomatal conductance modeling, namely, one uses the extended Prentice, one the semi-empirical, and the new framework that uses both. Results show that although all three approaches give reasonable estimates of limited measured fluxes, the present modeling framework gives much more reasonable estimates in the stomatal conductance and in other major model variables, and it also results in giving a relationship between carboxylation and transpiration that is consistent with observations. This modeling framework is general and can be adopted for other fields of study.