Northern Prairie Wildlife Research Center
University of North Dakota, Department of Geology and Geological Engineering, University Station, Box 8358, Grand Forks, ND 58202
Wetlands in ground-water discharge areas of the Red River Valley of eastern North Dakota provide runoff, flood control, and wildlife habitat. Although essential to the assessment of wetland function and management alternatives, the hydrology of these wetlands is not well understood.
Two years of soil and ground-water measurements have been used to develop a hydrological model of a 600-ha wetland surrounding Lunby and Stewart lakes, about 24 km northwest of Grand Forks. In addition to 20 monitoring wells, instrumentation includes five nests with tensiometers, piezometers, and moisture probe access tubes. Estimates of hydraulic conductivity were made using slug tests, permeameters, and the instantaneous profile method.
Variations in the level of ground-water reveal seasonal uniformity throughout the wetland area. From November through March, upward hydraulic gradients and steady water-level drop occurs in response to the upward movement of water toward frost in the vadose zone. A rapid rise of water levels during April and May results from the infiltration of precipitation, snowmelt, and water held in shallow vadose storage. Because of a thick capillary fringe, sediments have little air-filled porosity and the water table may rise 15-20 times the equivalent amount of infiltrated precipitation, depending on antecedent moisture. Greatest water-level variability occurs during the summer, when recharge from precipitation is rapidly depleted by evapotranspiration (ET). Vertical hydraulic gradients that range between ± 1.2 reverse during the summer and result from the interaction of precipitation and ET. Horizontal gradients generally do not exceed about 0.001; combined with the low hydraulic conductivity of the clayey lacustrine sediments (0.1-1 m/yr), minimal lateral flux of ground-water occurs. Seasonal variations of ground- and soil-water flux are complex in the vadose zone but reveal similar characteristics below the water table at different sites.
In an effort to predict water level variability, the U.S. Geological Survey MODFLOW code was used to simulate one- and two-dimensional saturated flow in the wetland. Climate records provided estimates of rainfall and ET during normal, dry, and wet periods and were applied through the ET and recharge options of the code. Actual ET was estimated from potential ET by using a variable linear function that depended on the average depth to the water table. The recharge option was used to represent upward vadose flow during the winter and recharge of precipitation and water released from vadose storage during the early spring. Water-level response to recharge and ET was very sensitive to storativity. The one-dimensional model provided good correspondence between measured and predicted hydrograls.
Both field and model results suggest that shallow vertical flux of ground- and soil-water dominates the hydrological budget. Hydrological conditions need to be carefully monitored before conceptual or simulation models can be applied, but the results should provide a method to estimate wetland size and permanence during climate variability or following management changes. A more accurate hydrological budget would require meteorological monitoring, calibration of a variably saturated model, and a better method for characterizing evapotranspiration.