ANALYSIS OF BORDER STRIP IRRIGATION AND ESTIMATION OF INFILTRATION PARAMETERS

Abstract

Modelling border strip irrigation system is a complex process since it involves both the overland and subsurface flow. The numerical solution of border strip irrigation involves the solution of both surface and subsurface flows. The present study is concerned with developing a hydrodynamic numerical model for the simulation of overland and subsurface flow for border strip irrigation and estimation of field infiltration parameters. A numerical model is developed by solving the differential equations governing overland flow (Saint Venants equations) and subsurface flow (Richards equation). Explicit MacCormack scheme is used to solve the Saint Venants equations while the Richards equation is solved using a mass conservative fully implicit finite difference method. The model performance is assessed by comparing model predictions by numerically as well as experimentally observed irrigation events such as irrigation advance and recession data reported in literature. The model is validated with the data of surface irrigation experiments on three different soils reported in literature. A detailed surface irrigation field experiments has been conducted involving both overland and subsurface measurements to asses the performance and applicability of the numerical model developed in predicting both overland and subsurface flow variables. In addition to the surface measurements such as advance, recession and flow depths, subsurface measurements such as pressure heads and moisture contents are also measured. The pressure heads are measured using tensiometers and the moisture contents are measured using Time Domain reflectometer (TDR). The model is validated by comparing model predicted irrigation advance, recession and subsurface moisture profiles with experimental data. The accurate prediction of border strip irrigation events such as such as irrigation advance, recession and subsurface wetting front movement mainly depends on the system parameters, like Manning's roughness coefficient n and infiltration parameters: saturated hydraulic conductivity KsaU Van Genuchten water retention parameters (Van Genuchten, 1990), av, ny, 9S and <9r. Among these, the estimation of infiltration parameters at field level is one of the difficult tasks (Walker and Skogerboe, 1987). For a relatively big field, estimation of infiltration parameters using infiltrometers requires that the test be conducted at many places. Further, these parameters maynot represent the infiltration phenomenon at field scale. An alternative to these direct measurement techniques is to employ inverse techniques for parameter estimation. In such an approach, the infiltration parameters are estimated by minimizing the deviations between the model predicted and field observed flow attributes such as irrigation advance, recession, flow depth and wetting front movement. In this study, a parameter estimation model is developed by coupling the numerical model with a Sequential unconstrained minimization technique (SUMT). The issues of identifiability and uniqueness are discussed by estimating the parameters from hypothetical data. The robustness of the model is assessed by varying the number of estimated parameters from 1 to 3. In this study, saturated hydraulic conductivity KsaU water retention parameters av, and /?v, are identified. The irrigation advance and summation of flow depths are used to identify single parameter. It is observed that the parameter estimates using summation of flow depths are in good agreement with their true values. Further, the summation of flow depths is used to identify two and three parameters. The results of simultaneous estimates of two parameters show that the optimization technique converges to the true values. However, simultaneous estimation of all the three infiltration parameters is not possible with flow depth data. Inclusion of moisture content in the objective function does not guarantee unique solutions. in