A Bayesian information theoretic approach to assessing stomatal conductance dynamics in a perennial agricultural crop

Le résumé fourni par la source

Perennial grain crops have been proposed as a more sustainable alternative to annual counterparts given their potential to replicate natural ecosystem services when managed correctly while requiring less external input than annual agriculture. In order to better understand the potential benefits of this transition, we investigated the water use dynamics of perennial intermediate wheatgrass under a variety of environmental conditions. We focused on the bulk stomatal conductance calculated from the Ball–Berry method using 6 years of eddy covariance measurements collected in central Kansas, USA. Bayesian inference was used to construct posterior distributions of the stomatal conductance. Kohonen self-organizing maps were used to assess stomatal dynamics under varying environmental conditions such as drought and wet periods. The relationships between the environmental variables and the stomatal conductance were quantified using an information theoretical approach. We note restricted stomatal distributions for water-stressed periods, while wet and normal conditions exhibit similar distributions. Mutual information content demonstrated that the functional dependence between stomatal conductance and environmental variables changes under different environmental conditions with soil water content and incoming solar radiation dominating in wet conditions and vapor pressure deficit and incoming solar radiation contributing the most in dry conditions. Relative entropy metrics illustrated the importance of soil water content, vapor pressure deficit, wind speed, and incoming shortwave radiation on stomatal conductance estimates. This study demonstrates the dynamics of a key land-atmosphere coupling variable, stomatal conductance, giving further insight into the functioning and potential environmental benefits of a perennial intermediate wheatgrass.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.