The stomatal response to rising CO2 concentration and drought

The stomatal response to rising CO2 concentration and drought is predicted by a hydraulic trait-based optimization model

by Wang Y., Sperry J. S., Venturas M. D., Trugman A. T., Love D. M., Anderegg W. R. L. (2019)

Yujie Wang 1, John S. Sperry1, Martin D. Venturas 1, Anna T. Trugman 1, David M. Love 1,2 and William R. L. Anderegg 1


1 School of Biological Sciences, University of Utah, Salt Lake City, 257S 1400E, UT 84112, USA;

2 Warnell School of Forestry and Natural Resources, University of Georgia, 180 E Green Street, Athens, GA 30602-2152, USA

===

In Tree Physiology 00: 1-12 – https://doi.org/10.1093/treephys/tpz038

http://sperry.biology.utah.edu/publications/Wang_et_al_2019_TPH.pdf

Abstract

Modeling stomatal control is critical for predicting forest responses to the changing environment and hence the global
water and carbon cycles. A trait-based stomatal control model that optimizes carbon gain while avoiding hydraulic
risk has been shown to perform well in response to drought. However, the model’s performance against changes in
atmospheric CO2, which is rising rapidly due to human emissions, has yet to be evaluated. The present study tested
the gain–risk model’s ability to predict the stomatal response to CO2 concentration with potted water birch (Betula
occidentalis Hook.) saplings in a growth chamber. The model’s performance in predicting stomatal response to changes
in atmospheric relative humidity and soil moisture was also assessed. The gain–risk model predicted the photosynthetic
assimilation, transpiration rate and leaf xylem pressure under different CO2 concentrations, having a mean absolute
percentage error (MAPE) of 25%. The model also predicted the responses to relative humidity and soil drought with
a MAPE of 21.9% and 41.9%, respectively. Overall, the gain–risk model had an MAPE of 26.8% compared with the
37.5% MAPE obtained by a standard empirical model of stomatal conductance. Importantly, unlike empirical models,
the optimization model relies on measurable physiological traits as inputs and performs well in predicting responses
to novel environmental conditions without empirical corrections. Incorporating the optimization model in larger scale
models has the potential for improving the simulation of water and carbon cycles.

Advertisements