SIMULATING RECHARGE IN HUMID CLIMATES: THE EFFECT OF THE UNSATURATED ZONE IN THE TROUT LAKE WATERSHED, WISCONSIN

Recharge is the most important source of water to groundwater systems, but is often not well quantified. As a result, it is usually treated as a calibration parameter in groundwater flow models, where it is constrained by transient head and baseflow measurements. While the unsaturated zone in humid regions is not as thick as in arid areas, it is conceivable that unsaturated zone processes may also affect the timing and magnitude of recharge.

Recently, USGS researchers have been working toward adding unsaturated zone processes to the groundwater model MODFLOW2000. A one-dimensional, kinematic wave approach is used that maintains conservation of mass in a homogeneous unsaturated zone but does not consider flow due to capillary forces (Niswonger and Prudic, 2004). An early version of this code was applied to the sandy sediments of the Trout Lake watershed in northern Wisconsin. An existing transient simulation was used to test the effects of the unsaturated zone. The model has 30 lakes simulated with the lake and stream flow routing packages, and consists of four layers with a total of 220800 cells. Time was discretized into monthly stress periods, and total length of the simulation was ten years. The test consisted of a comparison of two cases: 1) the standard approach of adding infiltration directly to the water table; and 2) adding the same infiltration below the root zone and subsequently routing it through the unsaturated zone.

Results show that in areas with thin unsaturated zones, very little difference in timing of recharge was observed between methods but exfiltration (rejected recharge) did occur – a process that could not be simulated if the standard approach of adding water to the water table was used. Thus, in areas with thin unsaturated zones, the timing of recharge may be properly simulated using the standard approach, but the ratio of recharge to runoff may be in error. In areas with thicker unsaturated zones, the recharge pulse was more diffuse when routed through the unsaturated zone, which resulted in a better fit to transient head data. Thus, in thicker unsaturated zones, the volume of water infiltrated was properly simulated using the standard approach, but the timing was less representative of field conditions.

Niswonger, R.G., and Prudic, D.E., 2004, Modeling variably saturated flow using kinematic waves in MODFLOW, AGU Water Science and Application 9, p. 101-112.