John E. Tobiason
Wachusett Reservoir, located in central Massachusetts, supplies water to the Boston, Massachusetts metropolitan area. The reservoir receives water from a watershed of 117 square miles, as well as water transferred from the Quabbin Reservoir to the west. Quabbin Reservoir water generally has lower levels of most water quality constituents than water received from the Wachusett Reservoir watershed. The Massachusetts Department of Conservation and Recreation (DCR) manages the watershed and monitors tributary and in-reservoir water quality, while the Massachusetts Water Resources Authority (MWRA) is responsible for treatment and distribution. CE QUAL W2 is a two dimensional, laterally averaged water quantity and quality modeling program. Version 2 of this program, used in this research, provides the ability to model 21 water quality constituents in addition to water surface elevation and temperature. Constituents modeled in this study include: total organic carbon (TOC) consisting oflabile dissolved organic matter (LDOM) refractory dissolved organic matter (RDOM) algae, and detritus; nutrients including nitrate/nitrite, arnmonmm, orthophosphate; and the absorbance of 254 urn ultraviolet light (UV254). This study implemented CE QUAL W2 to study the sources, fate, and transport of these constituents in Wachusett Reservoir. The water quality model was calibrated using data from 2001 and 2002. All required input and initial condition data, including 'inflow and outflow quantities, temperatures, constituent levels, and ambient meteorology were available from DCR, MWRA, the United States Geological Survey (USGS) and the National Oceanic and Atmospheric Administration (NOAA). Values for model parameters were determined to ensure best fit between model predictions and field data for Cosgrove Aqueduct, the main withdrawal of Wachusett Reservoir. The water quality model was then validated with data from 2000 before simulations were run. NOM levels were relatively constant throughout the calibration period, although seasonal variation is significant. Lower NOM levels generally occur when the majority of withdrawn water originated in Quabbin Reservoir, and higher levels occur when water originating in the Wachusett Tributaries dominates. Measured TOC levels varied from 1.8 to 3.3 mglL at Cosgrove during 2001 - 2002. These TOC trends were captured by defining 95% of inflow TOC as dissolved organic carbon (DOC), defining the remaining 5% detritus, and then defining 20% of inflow DOC as LDOM, and the remaining 80% as RDOM. First order LDOM and RDOM decay rates of 0.008 and 0.0008 day"I, as well as a first order RDOM to LDOM decay rate of 0.0008 day"I were most appropriate. Maximum algal growth and respiration rates of 1.9 and 0.1 day"I were used. UV254 varied from 0.03 to 0.08 cm·I during the calibration period. These trends were captured by a first order temperature dependent decay rate of 0.0008 day· I and a value of 2.6E-5 em2 !cal for a constant (a) relating the impact of sunlight irradiance on UV254 decay. The parameter values determined through calibration were successfully used to validate the model using data from 2000, despite the lack of constituent data for Quabbin Transfer for that year (constituent levels were assumed to be the average of 2001 ~ 2002 levels). TOC levels at Cosgrove Aqueduct ranged from 1.7 to 3.4 mglL for that year, while UV254 levels ranged between 0.04 and 0.08 em .1. Two simulations showed that transferring water from Quabbin to Wachusett at 8.7 m3!s during periods of high tributary runoff may reduce TOC levels at Cosgrove by up to 0.2 mglL and UV254 levels by up to 0.008 em·I. A third simulation demonstrated that a large runoff event occurring in late summer! early fall may lead to large increases in TOC and UV254 levels at Cosgrove, and may result in an unusual algal bloom. A fourth simulation was run to evaluate the impact of bypassing Wachusett. Reservoir with Quabbin Transfer; the model predicts that TOC and UV254 levels at Cosgrove will increase, but the increased mean hydraulic residence time within Wachusett Reservoir results in more decay of those constituents. The resulting mixture of Quabbin and Wachusett water in similar proportions to those that actually occurred contains lower NOM levels than would exist if the bypass did not occur. These results show that Wachusett Reservoir constituent levels at the Cosgrove withdrawal are strongly source driven, although in-reservoir processes are also important. The calibration and validation results indicate that CE QUAL W2 can be effectively used to predict NOM levels at Cosgrove. The simulation results suggest that controlled and uncontrolled events may impact water quality at Cosgrove.