環境変数を用いたトウモロコシ葉の気孔コンダクタンスのモデリング

Abstract

In the present paper, we attempted to use the average photosynthetic photon flux density (Qp), saturation deficit (De), leaf water potential (ΨL) and leaf temperature (TL) in the canopy as environmental variables (X) to optimize Jarvis' environmental variable model for estimating stomatal conductance (Gst) of a leaf in the maize canopy. The data, measured in the field in 1992-1994, were separated into three sets of short time period as one-year and a set of long time period as three-year. Therefore, the optimized models (*G(X)) to the each data set were selected through comparison of the coefficient of determinations (R2) of models with different numbers (from 1 to 4) of variable, respectively. The problems of application of the models on different period scale were also discussed.In the case of a short time period of one year, a relatively high estimated precision of Gst(R2=0.57-0.75) was obtained with the two-variable model *G(Qp, De) or *G(Qp, ΨL). However, in the long time period of three years, it was necessary that the model was constructed with three variables at least. The R2 of three-variable model *G(Qp, De, ΨL) was 0.43, and even if the model was constructed with four variables the R2 of model *G(Qp, De, ΨL, TL) was only 0.45.