Estimation of the effects of land use and groundwater withdrawals on streamflow for the Pomperaug River, Connecticut

Abstract

A precipitation runoff model for the Pomperaug River watershed, Connecticut was developed to address issues of concern including the effect of development on streamflow and groundwater recharge, and the implications of water withdrawals on streamflow. The model was parameterized using a strategy that requires a minimum of calibration and optimization by establishing basic relations between the parameter value and physical characteristics of individual hydrologic response units (HRUs) that comprise the model. The strategy was devised so that the information needed can be obtained from Geographic Information System and other general databases for Connecticut. Simulation of groundwater recharge enabled evaluation of the temporal and spatial mapping of recharge variation across the watershed and the spatial effects of changes in land cover on base flow and surface runoff. The modeling indicated that over the course of a year, groundwater provides between 60 and 70 percent of flow in the Pomperaug River; the remainder is generated by more rapid flow through the shallow subsurface and runoff from impermeable surfaces and saturated ground. Groundwater is recharged primarily during periods of low evapotranspiration in the winter, spring, and fall. The largest amount of recharge occurs in the spring in response to snowmelt. During floods, the Weekeepeemee and Nonnewaug Rivers (tributaries that form the Pomperaug River) respond rapidly with little flood peak attenuation due to flood-plain storage. In the Pomperaug River, flood-plain storage is more important in attenuating floods; abandoned quarry ponds (O&G ponds) adjacent to the river provide substantial flood storage above specific river stages when flow from the river spills over the banks and fills the ponds. Discharge from the ponds also helps to sustain low flows in the Pomperaug River. Similarly, releases from the Bronson-Lockwood reservoir sustain flow in the Nonnewaug River and tend to offset the effect of groundwater withdrawals from a well field adjacent to the river during periods of natural low flow. The model indicated that under the current zoning, future development could reduce low flows by as much as 10 percent at the 99 percent exceedance level (99 percent of flows are greater than or equal to this flow), but would not substantially increase the highest flows. Simulation of projected and hypothetical development in the watershed shows, depending on how stormwater is managed, that between 10 and 20 percent effective impervious area in an HRU results in streamflow becoming dominated by the surface-runoff component. This shift from a groundwater-dominated system would likely result in substantial changes in water quality and instream habitat characteristics of the river. Base flow to streams in the Pomperaug River watershed is reduced by both increased impervious surface and increased groundwater withdrawals. For the watershed as a whole, increasing groundwater withdrawals have the potential for causing greater overall reductions in flow compared to increased development and impervious surfaces. Additionally, on the basis of groundwater-modeling simulations, the projected increase in development across the watershed and, to a lesser extent the increase in groundwater withdrawals, will increase the number of local losing reaches experiencing dry conditions and the duration of these dry periods. The location of the losing reaches tends to be in areas near the transition from the uplands to the valley bottoms that are filled with coarse glacial stratified deposits. The simulated increase in the duration and extent of localized dry stream reaches is most sensitive to local increase in impervious surface. Conversion of land from forest or developed land cover to pasture or agricultural land increases groundwater recharge and discharge to streams, while at the same time increasing overall streamflow (the opposite effect as increased impervious surface). These resu