Modelling non-linear water table dynamics and specific discharge through landscape analysis

A method is presented to derive high resolution maps of water table dynamics and specific discharge through landscape analysis. The approach consists of deriving the parameters of a regionalized non-linear time series model on a detailed grid from information on topography, land use and geology. The resulting model is particularly suited for systems with deep permeable soils, shallow water tables and intensive drainage. The first step in parameterizing the model consists of deriving a topologically correct drainage network from a high resolution DEM in combination with a topographic layer with mapped water courses. The drainage network provides for pixels containing water courses the dimensions of the water course, bottom altitude and, for perennial water courses, their average stage. Next, from the drainage network and a geological map, drainage functions are derived for each pixel relating the local drainage base and the drainage resistance to the height of the water table. The result is a regionalized non-linear time series model of water table dynamics that can be used to generate time series of water table elevation for each pixel, as well a specific discharge, using time series of precipitation surplus (precipitation minus evapotranspiration) as input. The method is tested in a 10 × 10 km area in the North -East of the Netherlands. Predicted water table depth and discharge are compared with observations and with the results from a physically-deterministic groundwater flow model.