Global Identification of Surface Irrigation Parameters

Abstract

A numerical method is developed for the identification of bed resistance and infiltration parameters in surface irrigation based on global procedures. The method is characterized by global convergence—a unique solution from any feasible set of initial parameter estimates, which is not possible for standard variable metric methods. The algorithm is applied to border irrigation described by the zero-inertia model and is used to produce three-dimensional maps of the convergence history of some well-known field observations. Additional enhancements regarding parameter scaling, gradient modification, and combination with a Gauss-Newton method yield a robust and efficient algorithm that can be applied to a variety of practical problems.