Estimation of groundwater recharge in sandy till with two different methods using groundwater level fluctuations

Two methods for estimation of groundwater recharge, both based on groundwater level fluctuations, were applied in a moraine area in southeastern Sweden. The first method utilized a onedimensional soil water model which was tested against observed groundwater levels. The boundary conditions were defined by using standard meteorological data and submodels for precipitation, snow dynamics, interception, evapotranspiration and horizontal groundwater outflow. The second method directly transformed groundwater level fluctuations to equivalent amounts of water from a constructed recession curve and the specific yield concept. Conceptually the two methods could be characterized as inflow and response methods respectively.

A good fit between observed and simulated groundwater levels was obtained by the soil water modelling. The results were, however, shown to be rather insensitive to displacement in the water balance between evapotranspiration and groundwater outflow, giving a good fit for a simulated net groundwater recharge ranging between 134 and 197 mm. The results from the attempts to use groundwater level fluctuations directly were discouraging. Compared to the soil water simulations the results were unstable and quite different for different years. It was impossible to use a constant specific yield or even different specific yields depending on depth.

The conclusion was that the possibilities to use groundwater level data for quantitative water balance studies are limited under the studied climatical and hydrogeological conditions. The modelling effort clearly demonstrated the need for a better quantitative knowledge on soil properties if water balance information is to be deduced. The soil water model though, could be a valuable tool studying variations within and between different years as well as processes and single events.