Use of a watershed-modeling approach to assess hydrologic effects of urbanization, North Fork Pheasant Branch basin near Middleton, Wisconsin

Abstract

The North Fork Pheasant Branch Basin in Dane County, Wisconsin is expected to undergo development. There are concerns that development will adversely affect water resources with increased flood peaks, increased runoff volumes, and increased pollutant loads. To provide a scientific basis for evaluating the hydrologic system response to development the Precipitation Runoff Modeling System (PRMS) was used to model the upper Pheasant Branch Creek watershed with an emphasis on the North Fork Basin. The upper Pheasant Branch Creek (18.3 mi2; 11,700 acres) Basin was represented with 21 Hydrologic Response Units (daily time step) and 50 flow planes (5-minute time steps). Precipitation data from the basin outlet streamflow-gaging station located at Highway 12 and temperature data from a nearby airport were used to drive the model. Continuous discharge records at three gaging stations were used for model calibration. To qualitatively assess model representation of small subbasins, periodic reconnaissance, often including a depth measurement, was made after precipitation to determine the occurrence of flow in ditches and channels from small subbasins. As a further effort to verify the model on a small subbasin scale, continuous-stage sensors (15-minute intervals) measured depth at the outlets of three small subbasins (500 to 1,200 acres). Average annual precipitation for the simulation period from 1993 to 1998 was 35.2 inches. The model simulations showed that, on average, 23.9 inches were intercepted by vegetation, or lost to evapotranspiration, 6.0 inches were infiltrated and moved to the regional ground-water system, and 4.8 inches contributed to the upper Pheasant Branch streamflow. The largest runoff event during the calibration interval was in July 1993 (746 ft3/sec; with a recurrence interval of approximately 25 years). Resulting recharge rates from the calibrated model were subsequently used as input into a ground-water-flow model. Average annual recharge varied spatially from 2.3 inches per year in the highly impervious commercial/industrial area to 9.7 inches per year in the undeveloped North Fork Basin with an average overall recharge rate of 8.1 inches per year. Two development scenarios were examined to assess changes in water-budget fluxes. In scenario A, when development was predominantly low-density residential with 5 to 10 percent commercial development along principal roadways, mean annual streamflow increased by 53 percent, overland flow increased by 84 percent, base flow decreased by 15 percent and annual recharge to the regional ground-water system was reduced by 10 percent. In development scenario B, the entire North Fork and intervening area basins contained 50 percent commercial and 50 percent medium density residential land use. Annual storm runoff increased by over 450 percent. The ground-water model for the Pheasant Branch that used the scenario B recharge rates simulated a lowered water table with zero base flow and that flow from Frederick Springs would be reduced 26 percent from present-day (1993?98) conditions.An additional example application of the model evaluated locations of flood detention ponds and potential recharge areas that may mitigate the changes in flood peaks and ground-water recharge resulting from urbanization. From February 1998 through July 1998, water-quality samples were collected by use of stage-activated automated samplers. Median suspended- sediment concentrations were similar between the North and South Fork Basins (194 and 242 mg/L, respectively); however, for other constituents, North Fork values were considerably higher: median phosphorus concentrations by 4 times (1.5 and 0.35 mg/L), median ammonia concentrations by 13 times (1.9 and 0.14 mg/L), and the phosphorus-to-sediment ratio by more than 6 times (21 and 3.1 mg/g).