Identifying and Modelling the Dynamic Response of Leaf Water Content to Water Temperature in Hydroponic Tomato Plant

Abstract

In this paper, we identified and modelled the short-term response of leaf water content to water temperature in hydroponic tomato plants using a neural network. Leaf water content was estimated from leaf thickness using an eddy current-type displacement sensor. Dynamic changes in the leaf water content of the tomato plants, as affected by water temperatures, was identified and modelled using a neural network. A three-layered neural network with optimal system order and hidden neuron number allowed nonlinearity of this system to be successfully identified. The estimated responses obtained from model simulation were correlated closely with the observed responses. Leaf water content increased with water temperature up to 35°C in a short period of several hours. At a water temperature above 35°C, however, leaf water content decreased with increasing water temperature. Leaf water content, including root water uptake, of tomato plants is significantly suppressed by high water temperature. The relationship between water temperature and leaf water content over a short-term of several hours is represented by a hill-shaped curve (nonlinear curve) and reaches maximum value when the water temperature is about 35°C. These dynamic and static relationships between water temperature and leaf water content were successfully confirmed from model simulation.