ABSTRACT

Water withdrawals from surface-water reservoirs and groundwater have affected streamflow in the Sudbury and Assabet River Basins. These effects are particularly evident in the upper Sudbury River Basin, which prompted the need to improve the understanding of water resources and aquatic habitat in these basins. In 2004, the U.S. Geological Survey, in cooperation with the Massachusetts Department of Conservation and Recreation, developed a precipitation-runoff model that uses Hydrologic Simulation Program–FORTRAN (HSPF) to evaluate the effects of water use and projected future water-use and land-use change on streamflow. As part of this study, the aquatic habitat in the basins and the effects of streamflow alteration also were evaluated.

Chapter 1 of the report covers the development of the HSPF model that focuses on the upper Sudbury River Basin (106 square miles) but covers the entire Sudbury and Assabet River Basins (339 square miles). The model was calibrated to an 11-year period (1993–2003) using observed or estimated streamflow at four streamgages. The model was then used to simulate long-term (1960–2004) streamflows to evaluate the effects of average 1993–2003 water use and projected 2030 water-use and land-use change over long-term climatic conditions. Simulations indicate that the average 1993–2003 withdrawals most altered streamflow relative to no withdrawals in small headwater subbasins where the ratios of mean annual withdrawals to mean annual streamflow are the highest. The effects of withdrawals are also appreciable in other parts of the upper Sudbury River Basin as a result of the perpetuation of the effects of large withdrawals in upstream reaches or in subbasins that also have a high ratio of withdrawal to streamflow. The simulated effects of potential 2030 water-use and land-use change indicate small decreases in flows as a result of increased water demands, but these flow alterations were offset as a result of decreased evapotranspiration associated with the loss of deep-rooted vegetation. Simulations of reactivating production wells near the north end of Lake Cochituate indicate pumping could substantially affect lake levels and flows at the lake outlet or in nearby reaches in the Sudbury River during periods of low flow, but the effects vary depending on the source of the water to the wells, which is largely unknown.

Chapter 2 of the report covers the fish-community assessment and comparison of streamflow-setting standards for protecting aquatic habitat. The fish-community assessment indicates the main stems of the Sudbury and Assabet Rivers are dominated by macrohabitat generalists. Water temperatures recorded in seven free-flowing reaches in the upper Sudbury River Basin at three sites unaffected by withdrawals or impoundments are generally suitable for cold-water fish; however, summer temperatures often rose to a level considered critical to long-term survival of brook trout. At four sites downstream from withdrawals or reservoirs, or both, summer water temperatures were often in the upper critical range for brook trout survival.

Physically and statistically based methods for determining streamflows for protecting aquatic habitat were applied at 10 selected riffle sites in the Sudbury and Assabet River Basins. Physically based methods, R2Cross and Wetted-Perimeter, use site-specific physical and hydraulic information and a one-dimensional hydraulics model, HEC-RAS, to determine flows that meet the criteria set forth by the method. The median flow that meets 2-of-3 of the R2Cross hydraulic criteria (percentage of bankfull wetted perimeter, average velocity, and mean depth) ranged from about 0.07 to 0.72 cubic feet per second per square mile (ft³/s/mi²) with an overall median of about 0.24 ft³/s/mi²; the median Wetted-Perimeter target flow ranged from about 0.10 to 0.51 ft³/s/mi² with an overall median of about 0.25 ft³/s/mi². Statistically based methods—Tennant, New England Aquatic Base Flow (ABF), and the Range-of-Variability Approach (RVA)—utilized HSPF simulated long-term streamflow (1960–2004) with no withdrawals to determine target flows or the monthly streamflow variability. The Tennant 30-percent of the mean annual flow (Q_MA) ranged from about 0.56 to 0.61 ft³/s/mi² with an overall median of 0.58 ft³/s/mi²; the ABF low-flow target ranged from about 0.14 to 0.62 ft³/s/mi² with an overall median of 0.32 ft³/s/mi². The frequency and duration target flows were not met were evaluated using HSPF simulated flows with and without water withdrawals. The median annual duration and total number of days flows were below the target were about three times greater for statistically based standards (Tennant 30 Q_MA and ABF) than for physically based targets (Wetted-Perimeter and R2Cross 2-of-3 criteria). Overall, physically based target flows were slightly more sensitive to flow alterations than the statistically based target flows.

First posted August 12, 2010