Modeling water flow in a Dutch polder network

Abstract

Agragrians in the Dutch Wadden sea region struggles with salinizing farmlands. Acacia Water's project Spaarwater enabled agrarians to measure salt concentrations in the ditches of their land using their mobile phone. All these measurements would ideally be used to evaluate the effects of salinization mitigation measures and to predict water quality over time. This gives rise to this thesis for an inquiry into a fast and simple surface water flow model which eventually could even be run on a mobile phone. The water model is a first necessary step for developing a water quality model with salt concentrations. The main focus of this thesis is on developing finite volume methods for the kinematic wave equation. It is a simplification of the full shallow water equations which are the governing equations for surface water flow. Finite volume methods for the diffusion wave equation and the full shallow water equations are also developed to serve as a comparison to the kinematic wave. It is shown that the kinematic wave and the diffusion wave have comparable features. Even though the kinematic wave is not the closest representation of reality, high computational gains could be made by using the kinematic wave equation instead of the full shallow water equations. For a simple straight ditch, the kinematic wave has shown to be 16 times faster than the full shallow water equations. For a network, however, it is not as straightforward. When comparing a kinematic wave network with a network in SOBEK, a software program that uses the full shallow water equations, the latter still has lower computational times.