Masters Theses

Date of Award

12-1996

Degree Type

Thesis

Degree Name

Master of Science

Major

Environmental Engineering

Major Professor

Bruce A. Tschantz

Committee Members

James Smoot, Terry Miller

Abstract

This study evolved out of the Oak Ridge Y-12 Plant's requirement to comply with the NPDES permit. The permit will limit mercury loads to 5 grams per day for water leaving the Y-12 Plant at Station 17 based on average daily values for a 3-month period in order to reduce the exposure risk to humans, animals and aquatic life. The NPDES permit allows mercury loads produced from flows greater than 15 million gallons per day (MOD) to be excluded from the calculations. Storm event flows frequently exceed 15 MGD and are capable of producing plant-wide mercury loadings of hundreds of grams from NPDES Category I, II, III and other outfall pipes, contaminated sediment in the streambed, and watershed surface area sources. Because of this NPDES requirement, a hydrologic tool is needed to help predict storm event mercury loads from the Y-12 Plant for different assumed management practices and modifications applied to the stream, storage facilities, and watershed areas. For this reason, a hydrologic/mercury transport model was developed and calibrated for the 1.8 square mile UEFPC watershed and 1.6 mile long UEFPC stream above Station 17, the stream gaging and water quality monitoring site at the Plant boundary. The Stormwater Management Model (SWMM) was chosen as an appropriate computer model for simulating both hydrologic flows and mercury transport loading. The model was calibrated using hydrologic and water quality data observed during the passing of the Hurricane Opal storm event in October 1995. The model was then validated using data gathered from a second storm which occurred in March 1996. Based on observations taken during two storm events. Lake Reality is undoubtedly trapping mercury. During the October 1995 storm, observed mercury loading just upstream from Lake Reality was 660 grams, while the loading just downstream from Lake Reality was only 330 grams, a reduction of 50 percent. Simulated loadings and reduction were nearly the same. Lake Reality was determined, through observation and simulation, to reduce the mercury loading for the smaller March 1996 storm event by about 70 percent. Using only rainfall data as input, the current calibrated SWMM model is able to predict water runoff volumes within a 4 percent margin of error, peak runoff flow rates within 8 percent error and mercury loadings within a 22 percent error at reliable stream gaging sites. These uncertainty estimates are based on comparison between observed and simulated data for both the calibration and validation storm events. The model is now capable of simulating both real and hypothetical storm events to predict hydrographs and runoff volume for assumed land use and land management practices in the Y-12 Plant segment of the UEFPC watershed. The SWMM model discussed above is not presently capable of simulating deposition and resuspension of mercury in the channel of UEFPC. Further work is in progress and requires characterization of the UEFPC channel sediments. In order to characterize the sediments, representative streambed sediment samples were obtained approximately every 200 feet in both the 1.4-mile long upper natural and concrete-lined reaches of the channel upstream from Lake Reality and in the lower 0.2-mile reach below Lake Reality, but above Station 17. Grain size analysis of the bed sediment showed that the median (d50) particle size generally decreased from the steeper, upper end of the UEFPC channel at the North/South Pipe (Outfall 200) to the flatter, lower concrete distribution channel to Lake Reality. The inference here is that the finer particles are being transported from the upper to the lower reaches of the stream, where sediment becomes trapped at least temporarily in the distribution channel to Lake Reality. The d50 size (median) ranged from about 7 mm in the upper natural reaches to about 1 mm in the lower concrete distribution charmel. The sediment size in the natural channel between Lake Reality and Station 17 measured approximately the same size as the sediment in the upper reaches of UEFPC. Approximately 20 to 30% of the sediment in the upper natural reach was sand-size (2 mm or 2000 μm) or smaller, but almost 70% of the total sediment in the lower distribution channel to Lake Reality was sand-size or smaller. 100% of the total sediment in Lake Reality was determined to be sand-size or smaller. Laboratory analysis of mercury concentrations in the streambed sediment fraction with particle size less than 2 mm demonstrated the important relationship of sediment size and mercury concentration within the channel bed. Results showed that mercury concentration typically increased with decreasing particle size for the clay/silt/sand sediment classes in UEFPC. Mercury concentrations for the clay size particles (less than 5 μm) ranged from about 125 to 1700 mg/kg (ppm) over the UEFPC streambed reach upstream from Lake Reality, while concentrations for the very coarse sand particles (1000 to 2000 μm) ranged from only 18 to 180 mg/kg over the same reach. Concentration of mercury in sediment samples collected in Lake Reality averaged about 150 mg/kg and ranged from 180 mg/kg for coarse sand to 140 mg/kg for clay particle sizes. Streambed mercury concentrations tended to be much lower in the short reach below Lake Reality and ranged from only 6 mg/kg for very coarse sand to 92 mg/kg for clay. 'Bulk' sediment samples, representing composite clay/silt/sand particles, showed that the composite sample mercury concentration varied from about 45 to 410 mg/kg for the upper UEFPC streambed and only 23 mg/kg for the 0.2 mile reach below Lake Reality. The sediment characterization study has shown that the UEFPC streambed contains sediment with significant concentrations of mercury which is subject to downstream transport.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS