Stomatal resistance modelling using the full factorial design: application to the New Guinea impatiens

Abstract

The stomatal resistance Rs appears as a key parameter governing plant transpiration. For that reason, it deserved a particular attention of modelers. Up to now however, most models were multiplicative models established at a given plant growth stage and not valid at another growth stage. In this prospect, the objective of the present work is (1) to develop a dynamic model of Rs based on the full factorial design FFD, and (2) to validate the model at several plant growth stages inside a production greenhouse. The FFD is based on an optimization process which makes it possible to establish a polynomial relationship between Rs and the radiation, humidity and temperature. To implement the model, a set of experiments was conducted inside a 10 m 2 growth chamber equipped with 4 Impatiens New Guinea pots. Rs was directly measured with a porometer on 5 leaves. Nine scenarios were tested with extreme values for the 3 considered parameters: radiation [0-150] W m -2 , humidity [55-75]% and temperature [15-26]°C. These values were chosen from real conditions observed inside a greenhouse and taking account of the operational constraints of growth chambers. The polynomial model deduced from the experiments evidenced its ability to predict Rs from the climatic parameters (r 2 = 0.99). It was then applied at several growing stages of a greenhouse Impatiens crop. Again, the model was able to correctly simulate the stomatal resistance (RMSE[186-309] s m -1 and r 2 [0.50-0.58] for measured Rs [40-2000]s m -1 . It also showed the prevailing role played on Rs by radiation during daytime and by humidity at night. The influence of the temperature however, seemed to be negligible. The FFD therefore appears to be a powerful tool to simulate satisfactorily the stomatal resistance at different growth stages. It could be useful to predict accurately the evolution of the plant transpiration.