Masters Theses

Date of Award

5-2005

Degree Type

Thesis

Degree Name

Master of Science

Major

Environmental Engineering

Major Professor

Bruce Robinson

Committee Members

John Buchanan, Randy Gentry

Abstract

Scheduled road construction prompted intensive water quality monitoring of the adjacent stream, the Little Pigeon River. Three phases of monitoring was planned to fully assess any impacts construction may have: pre, during, and post construction monitoring. One year of pre-construction monitoring has been completed. Three monitoring sites were installed. Site 1 was below, Site 2 in the middle, and Site 3 above all road construction. Each site had a YSI sonde that measured 15-minute pH, turbidity, conductivity, temperature, and stage. Storm samples were also captured through use of an Isco auto-sampler. Additionally, bi-weekly grab samples were taken at each site. All collected storm and grab samples were measured for pH, conductivity, acid neutralizing capacity, chloride, nitrate, sulfate, and ten metals including aluminum. In addition to the three main stream sites, Site 4 was located on Ramsey Prong in the south-east side of the watershed. It contained a YSI sonde that measured 15-minute pH, conductivity, temperature, and stage. Four grab samples were taken during base flow at Site 4. Two precipitation stations were also operated that contained bulk and sequential precipitation collectors in addition to tipping bucket rain gauges. Acid deposition is a major water quality driving force. pH and acid neutralizing capacity (ANC) are often used as indicators of acid deposition effects during base and storm flow. Descriptive statistics of sonde and base flow grab sample stream indicates acceptable base flow stream pH for Sites 1 – 3. However, Site 4 mean stream pH of 5.76 indicates unsuitable conditions for aquatic ecosystems. Sonde data and storm samples showed numerous occasions of episodic acidification at Sites 1 – 4 when using stream pH 5.5 as the criteria for acidity (Lachance and Bobee 1991). Similar to pH, base flow ANC iv was acceptable at Sites 1 and 2. However, Sites 3 and 4 mean base flow ANC indicated high sensitivity to acid deposition. Site 4 was likely affected by acid deposition but Anakeesta geology is also believed to play a role in the depressed stream pH and ANC. Increased stream concentrations of aluminum, sulfate, and nitrate are also indicators of acid deposition impacts. Mean stream aluminum, sulfate, and nitrate concentrations increased 25 to 59%, 12 to 24%, and 33 to 39%, respectively compared to base flow concentrations. Notably, mean concentrations of sulfate and nitrate found in precipitation samples were comparable to mean concentrations observed during storm flow. Mean concentrations of aluminum found in precipitation are not comparable to base or storm flow stream concentrations. This indicates soil leaching of aluminum caused from acid deposition. Using EPA criteria, numerous occasions of metal exceedances occurred during base and storm flow at Sites 1 – 3. Data for calcium and magnesium indicates low hardness during base and storm flow. This exacerbates problems associated with metal toxicity. Exceedances increased during storm flow as expected. Metal exceedance of EPA criteria for aluminum, copper, and zinc endured for at least 20.75 hours during a storm event on July, 25, 2004. However, several studies showed that metals concentrations must exceed EPA criteria by three fold to reach lethal concentrations that kill 50% (LC50) of brook and bull trout (Kazalauskiene et al. 2003 and Hansen et al. 2001.) LPR exceedances were not as severe and the duration for the LC50 values were also longer than observed continuous metal exceedances in the LPR. Turbidity and pH are perhaps two of the most commonly measured stream water quality parameters due to their overall importance on aquatic ecosystems. A large v quantity of sonde pH and turbidity was collected during this study. The data were used throughout this document to describe pH and turbidity. Multiple linear regression models were also constructed using sonde data. Storm event data were isolated to build a storm event pH and turbidity multiple linear regression models. Ultimately, lagged stage, sine and cosine functions of the day fraction, and total rain volume were used to predict storm event pH. Similarly, stage, stage2, stage intensity, and sine and cosine functions of the year fraction were used to predict storm event turbidity. Both models had several deficiencies that possibly prevented the best possible linear regression parameters from being estimated. Error residuals were not normally distributed and were serially correlated. However, high r2 values and application to a validation storm event data set showed the models were able to reasonably predict storm event pH and turbidity. The developed storm event linear regression models for pH and turbidity serve several purposes. The models show association between the response and predictor variables. From the associations developed causes for increased turbidity and decreased pH may be inferred (i.e. precipitation input depresses pH and increases turbidity). Additionally, a major objective was to form a basis of comparison of pre, during, and post construction water quality. In order to do this major water quality drivers must be understood in order to fully assess construction impacts. An understanding is particularly important for storm events because mean or median base flow water quality can be compared for pre, during, and post construction, comparing mean or median storm events is likely to be very misleading. The uniqueness of each storm event and the relative infrequency of significant storm events can cause mean or median to be misleading. Regression models allow the water quality to essentially be normalized for the size of the vi storm event and seasonal and diel patterns. The developed preconstruction storm event models should fit the during and post construction water quality. If not, construction impacts can be inferred. As stated above, a major objective was to form a basis of comparison for water quality during the various stages of this study. Once data from other phases of this study have been collected independent comparisons from the water quality sites may be completed. With these data it was prudent to first establish that the three main stream sites’ water quality were significantly different. Sonde and base flow pH was chosen to ascertain significant differences between stream sites. All sites had significantly different sonde and base flow pH with one exception. Mean base flow pH at Sites 1 and 2 were not found to differ significantly. A number of interesting water quality attributes were noticed during this study. Sonde base flow pH, temperature, and dissolved oxygen exhibited distinct diel cycles. Temperature and dissolved oxygen were caused by air temperature variation. Interestingly, pH diel cycles are believed to be due to biological photosynthesis and respiration. A slight diel cycle in conductivity also corresponded to biological action. Another unexpected water quality attribute was upward pH spikes at the onset of storm events. A 2.07 unit increase in pH was observed at Site 1 in response to a storm event on September 3, 2003. However, increases were generally one pH unit in magnitude. Examination of storm event water quality data showed these pH increases were due to increased stream alkalinity. Precipitation also showed similar alkalinity increases which indicate antecedent accumulation in the watershed of alkalinity from an unknown source.

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

Included in

Engineering Commons

Share

COinS