Groundwater tracing studies have shown a well-developed conduit system in the karst aquifer beneath the Buffalo Creek Watershed of Carter County, TN. Groundwater can quickly travel beneath topographic divides, transporting contaminants. Using GIS and field reconnaissance, two sites were selected for a two phase dye trace study. In May-June 2016 groundwater resurgences were located using activated carbon samplers, and in October-December 2016 flow velocities were measured using ISCO automatic water samplers. In study 1, a stream sinking in the back of Carter Saltpeter Cave was traced to its resurgence at Cave Springs Cave. In study 2, the dye mass centroid passed the resurgence 23.2 hours after injection. During study 1, dye injected into a sinkhole near the headwaters of Toll Branch rapidly diffused into multiple conduits. In study 2, dye from the Toll Branch site was not recovered, indicating that groundwater follows different flow paths dependent on water table level.

Horizontal subsurface flow (HSSF) constructed wetlands were evaluated for the treatment of flow-through trout farm effluent, phosphorus sorption affinity of gravel-bed media, and influence on Rhodamine WT (RWT) transport. HSSF wetlands coupled with mechanical pretreatment demonstrated significant (p <0.05) removal of total ammonia-nitrogen (TAN), total phosphorus (TP), total organic carbon (TOC), total suspended solids (TSS), five-day biochemical oxygen demand (BOD5), and turbidity. Treatment occurred predominantly within the wetland cells, with minimal removal of studied water quality parameters by means of sedimentation or microscreen filtration (80 ?m mesh). HSSF wetlands removed 69% of influent TSS, 24% of influent TP, and reduced turbidity by 66%. The removal of organic matter within the wetlands, as measured by BOD5, COD, and TOC was 62%, 50%, and 55%, respectively. After receiving effluent from a flow-through trout farm for about one year, the gravel media exhibited moderate removals of soluble phosphorus in batch and column sorption experiments. Partition coefficients (Kd) from batch sorption tests ranged from 45-90 mL/g. Low (60 mL/min) and high (165 mL/min) flow column experiments removed about 50 and 40% of influent PO4-P, respectively. The conservative nature of RWT in subsurface media has been called into question by many authors. Tracer response curves from tests conducted in pilot-scale HSSF wetlands exhibited elongated tails and dual peaks, in addition to mean tracer retention times far exceeding the theoretical value. Laboratory column testing of RWT and the more conservative NaCl tracer supported field data, indicating that RWT was more reactive within the wetland media. / Master of Science

Water temperature directly affects biological and chemical processes of fresh water ecosystems. Elevated instream temperatures are commonplace in the Virgin River of southwestern Utah during summer due to a hot desert climate and high water demands that result in low stream flows. This is of concern since the Virgin River is home to two endangered species, the Virgin River Chub (Gila seminuda) and Woundfin (Plagopterus argentissimus). Efforts to model instream temperatures within the Virgin River have been undertaken to help mitigate elevated instream temperatures including the development of a two-zone temperature and solute (TZTS) model. This model was developed to approximate the dominant processes that influence instream temperatures and used both temperature and solute data in parameter estimation. Past model applications highlighted two concerns: (1) how to confidently estimate the high number of parameters and (2) whether Rhodamine WT (RhWT) could be used as a conservative solute tracer within the Virgin River. To begin addressing these issues, spatially representative data were collected to facilitate the physical estimation of two previously calibrated parameters: total average channel width (BTOT) and the fraction of channel width associated with dead zones (β). Methods for analyzing multispectral and thermal infrared imagery were developed to provide estimates of these parameters at different resolutions. Three different TZTS model calibration cases were then evaluated to determine how decreasing the calibrated parameters and increasing the resolution and frequency at which these parameters are estimated improved model predictions and/or decreased parameter uncertainty. While temperature predictions did not change significantly in each of the calibrations, parameter uncertainty was reduced. The concern regarding the use of RhWT resulted in a series of studies to quantify the potential losses of RhWT within this system. A batch sorption study resulted in distribution coefficient values lower than those found in literature. A photodegradation study suggested possible photolysis; however, a dual tracer study conducted within the Virgin River comparing Br- (conservative tracer) with RhWT confirmed that there was insignificant RhWT loss within this system.