State-based open-loop control of plant growth by means of water stress training

In this paper a first evidence is provided that the targeted control of adaptive plant behavior for irrigation purposes is possible. The objective is aligning plant growth to water availability (and not vice versa) and utilizing training mechanisms to affect the relation between water use and plant growth. The approach is based on the experimental specification of two water deficit-related behavioral patterns: memory of stress and point of no return (damage). Mild stress duration has to be shorter than 2.7 days to avoid irreversible growth rate reduction (maximum stress duration time). Water stress information is stored (memorized) by the plant for three days at most (maximum water stress memory time). Therefore, adequate stress stimuli have to be repeated within this period to maintain training effect. Exceeding maximum memory time without stimulus results in a drop of water-based growth performance (growth[cm]water[g]) back to the level of untrained plants. In control experiments two different plant growth performance ranges were identified: ‘Hydrological time’ performance range without activated memory, and ‘usage-bound’ performance range in memorized states. ‘Usage-bound’ growth performance range shows 47 % higher water-based growth performance than ‘hydrological time’-based. An open-loop control approach is developed to control growth and water consumption using the intended alternation between the two performance ranges. The plant behavior due to water stress is modelled as a state machine (method of a conditioned automaton (control engineering)) representing directly the control algorithm. Based on the statistical validation results it can be concluded that training plants with intended stress sequences allows the control of plant growth and water use.