Modeling the Effects of Spatial Variability of Irrigation Parameters on Border Irrigation Performance at a Field Scale

Abstract

The interaction between surface and subsurface water flows plays an important role in surface irrigation systems. This interaction can effectively be simulated by the physical-based models, which have been developed on the basis of the numerical solutions to the Saint-Venant and Richards’ equations. Meanwhile, the spatial variability of field physical properties (such as soil properties, surface micro-topography, and unit discharge) affects the interaction between surface and subsurface water flows and decreases the accuracy of simulating surface irrigation events at large scales. In this study, a new numerical methodology is developed based on the physical-based model of surface irrigation and the Monte Carlo simulation method to improve the modeling accuracy of surface irrigation performance at a field scale. In the proposed numerical methodology, soil properties, unit discharge, surface micro-topography, roughness, border length, and the cutoff time for the unit discharge are used as the stochastic parameters of the physical-based model, while field slope is assumed as the constant value because of the same field tillage and management conditions at a field scale. Monte Carlo simulation is used to obtain the stochastic parameter sample combinations of the physical-based model to represent the spatial variability of field physical properties. The updated stochastic simulation model of surface micro-topography, which is developed to model the spatial distribution of surface elevation differences (SED), is used to obtain the surface micro-topography samples at a field scale. Compared with the distributed-parameter modelling methodology and the field experimental data, the proposed numerical methodology presents the better simulation performance.

Improving border irrigation performance with predesigned varied-discharge

“…In the border irrigation system, the variable measures are soil infiltration properties, roughness coefficient, border dimensions (length and width), slope, inflow rate and cut-off time (or cut-off distance) [7,8]. Although the soil infiltration properties and roughness coefficient have been proven to affect the performance of border irrigation [9][10][11], they are difficult to control artificially, so the soil infiltration properties and roughness coefficient are regarded as input parameters rather than controlled variables in border irrigation design.…”

Modeling Two-Dimensional Infiltration with Constant and Time-Variable Water Depth

“…where ψ is the water pressure potential into soil, expressed as the height of an equivalent water column (L) (positive in saturated zones and negative in unsaturated zones of soil); C(ψ)= dθ/dψ is the humidity specific capacity of soil; θ = θ(ψ) is the water volume per unit volume of soil or water content; (L 3 •L −3 ) is a ψ function, known as the humidity characteristic curve or water retention curve; K = K(ψ) is hydraulic conductivity (L•T −1 ), in a saturated soil as a function of pressure potential; gravitational potential is assigned to spatial coordinate z positively oriented down (L); ∇ = (∂/∂x,∂/∂y,∂/∂z) is the gradient operator; x and y are spatial coordinates (L), and t is time (T).…”

“…In the literature, studies on infiltration modeling in surface irrigation can be found; for example, Liu et al [2] proposed a coupled model in which surface water flow and solute transport are described using the zero-inertia equation and the average cross-sectional convection-dispersion equation, respectively, while the two-dimensional Richards' equation and the convection-dispersion equation are used to simulate water flow and solute transport in soils, respectively. Dong et al [3] developed a new numerical methodology based on the physical-based model of surface irrigation and the Monte Carlo simulation method to improve the modeling accuracy of surface irrigation performance at a field scale. Border irrigation was simulated by coupling the Saint-Venant equations and a one-dimensional Richards' equation.…”