Empirical models of the conductance of leaves in apple orchards

Publisher Website
Google Scholar
Abstract
Seasonal data on leaf conductance (gl) for three different apple cultivars grown in four separate orchards with different aged trees was studied between 1979 and 1985. A number of empirical models for predicting leaf conductance from environmental measurements were compared using this data and a general method for adapting such models for the prediction of different data sets is proposed. Although stepwise multiple regression identified relative humidity or vapour pressure as important variables, it frequently did not identify the optimal set of independent variables, which often did not include either of these. There was no advantage in regressing glagainst principal components of the environment, rather than against the raw environmental variables. A simple model involving air vapour pressure deficit, air temperature and a hyperbolic function of irradiance was found to explain between 32 and 62% of the variance in glfor the different data sets. Parameters fitted for one data set led to the effective prediction of glin other years or plots. The model fit could generally be improved significantly by including soil moisture deficit among the independent variables.