GAIN-P:: A new strategy to increase furrow irrigation efficiency

The new methodology GAIN-P combines Genetic Algorithms, Artificial INtelligence techniques and rigorous Process modeling for substantially improving irrigation efficiency. The new strategy simultaneously identifies optimal values of both scheduling and irrigation parameters for an entire growing season and can be applied to irrigation systems with adequate or deficit water supply. In this contribution, GAIN-P is applied to furrow irrigation tackling the more difficult subject of the more effective deficit irrigation. A physically-based hydrodynamic irrigation model is iteratively coupled with a 2D subsurface flow model for generating a database containing all realistically feasible scenarios of water application in furrow irrigation. It is used for training a problem-adapted artificial neural network based on self-organized maps, which in turn portrays the inverse solution of the hydrodynamic furrow irrigation model and thus enormously speeds up the overall performance of the complete optimization tool. Global optimization with a genetic algorithm finds the schedule with maximum crop yield for the given water volume. The impact of different irrigation schedules on crop yield is calculated by the coupled furrow irrigation model which also simulates soil evaporation, precipitation, and root water uptake by the plants over the whole growing season, as well as crop growth and yield. First results with the new optimization strategy show that GAIN-P has a high potential to increase irrigation efficiency.