A modified optimal stomatal conductance model under water-stressed condition

Document Type : Research Paper

Authors

1 Center for Agricultural Water Research in China, China Agricultural University, Beijing 100083, China.

2 College of Agriculture, Guangxi University, Nanning, Guangxi 530005, China.

Abstract

Accurate estimation of leaf stomatal conductance (gs) is important in predicting carbon and
water cycles of terrestrial ecosystem. To estimate gs on field-grown soybean and maize under
water-stressed condition accurately, a modified optimal stomatal conductance (OSCM) model
was established based on the relationship between marginal water cost of carbon gain and soil
water content by introducing a water stress factor (f(θv)). f(θv) had same form with that in
Jarvis and Ball-Berry-Leuning (BBL) models. The OSCM model was evaluated and compared
with the original optimal stomatal conductance (OSC), Jarvis and BBL models by comparing
observed and estimated gs of three-year data on soybean and four-year data on maize in an
arid region of northwest China. Results show that the OSCM and OSC models were more
steady and accurate than the Jarvis and BBL models for estimating gs on soybean and maize at
the different years. Moreover, the OSCM model performed better than the OSC model
because of considering the effect of water stress. Compared with the OSC, Jarvis and BBL
models, the OSCM model improved the accuracy of estimating gs on soybean and maize on
average by 7%, 25% and 35% and reduced the RMSE by 19%, 56% and 43%, respectively.
As for estimating diurnal change of gs on soybean and maize under both well-watered and
water-stressed conditions, the OSCM model also performed better than the OSC, Jarvis and
BBL models. Under water-stressed condition, only the OSCM model is recommended due to
its high accuracy, conservative and accessible parameter, which can provide a more accurate
and convenient tool in predicting water and carbon fluxes of terrestrial ecosystem in the arid
area.

Keywords

Files

Share

How to cite

Statistics