Development, testing and application of DrainFlow: A fully distributed integrated surface-subsurface flow model for drainage study

- A fully distributed coupled surface-subsurface flow model titled DrainFlow has been developed and is described.
- A smooth transitioning technique is introduced to overcome the non-linearity problem created from switching between wet and dry boundaries.
- DrainFlow has the advantage of estimating infiltration directly from the Richards equation rather than through analytical or empirical infiltration approaches.
- DrainFlow is quite flexible for changing all or part of the model dimensions as required by problem complexity, problem scale, and data availability.

Hydrological and hydrogeological investigation of drained land is a complex and integrated procedure. The scale of drainage studies may vary from a high-resolution small scale project through to comprehensive catchment or regional scale investigations. This wide range of scales and integrated system behaviour poses a significant challenge for the development of suitable drainage models. Toward meeting these requirements, a fully distributed coupled surface-subsurface flow model titled DrainFlow has been developed and is described. DrainFlow includes both the diffusive wave equation for surface flow components (overland flow, open drain, tile drain) and Richard's equation for saturated/unsaturated zones. To overcome the non-linearity problem created from switching between wet and dry boundaries, a smooth transitioning technique is introduced to buffer the model at tile drains and at interfaces between surface and subsurface flow boundaries. This gives a continuous transition between Dirichlet and Neumann boundary conditions. DrainFlow is tested against five well-known integrated surface-subsurface flow benchmarks. DrainFlow as applied to some synthetic drainage study examples is quite flexible for changing all or part of the model dimensions as required by problem complexity, problem scale, and data availability. This flexibility enables DrainFlow to be modified to allow for changes in both scale and boundary conditions, as often encountered in real-world drainage studies. Compared to existing drainage models, DrainFlow has the advantage of estimating actual infiltration directly from the partial differential form of Richard's equation rather than through analytical or empirical infiltration approaches like the Green and Ampt equation.