An Idea of Concentration Gradient FACE (Free Air CO₂ Enrichment) and a Preliminary Experiment of Its Control Using a Neural Network

Abstract

In existing FACE (Free Air CO₂ Enrichment) systems, efforts have been made to maintain the CO₂ concentration within a FACE ring at the setpoint value. Therefore, effects of only one elevated CO₂ concentration on plants and/or ecosystems can be examined with one ring. The basic idea of the present study was, instead of controlling the CO₂ concentrations within a ring at a constant value, to create and maintain a constant CO₂ concentration gradient within a ring. In this way, examinations of the effects of different CO₂ concentrations on plants and/or ecosystems would be realized using one ring. An artificial neural network was applied to control the CO₂ gradient within a ring of a stretched hexagon with ten CO₂ emitting tubes. The results of the experiments showed that when the wind speed was more than the detectable range by the anemometer (about 30 cm s^-1), a concentration gradient was in most cases observed, though not completely the same as the target gradient. When the wind speed was less than the detectable value, the CO₂ concentrations within a ring rose far above the target values and the gradient was not created.