WSP03775 - Laserfiche WebLink

Home Browse Search

metal concentrations and pH generally were strong at both sites (AR34.5 and AR31.0). However, at site AR31.0, scatterplots of the data indicated that a linear relation did not exist between pH and the logarithms of most of the metal concentrations; therefore, linear-regression models could not be used for predicting metal loads at site AR31.0. Correlations and scatterplots between the logarithms of metal concentrations and pH at site AR34.5 indicated that a linear relation existed. However, the correlations generally indicated that between 65 and 83 percent of the variation in metal concentrations could be explained with pH; conversely, between 17 and 35 percent of the variation in metal concentrations could not be explained with pH. A variation ofthe standard time-interval method that generally allowed for an estimate of statistical certainty in load calculations and takes into account the large temporal variations that occurred in metal concentrations was used to estimate metal loads in this report. In the modified time-interval method, the data record is divided into several discrete time intervals based on changes in concentration, streamflow, or events. The mean concentration of the values that occurred during each time interval was multiplied by the mean daily streamflow to determine daily metal loads. Estimates of the daily metal loads were summed into flow-period and annual metal loads. The resulting percent standard errors using this method generally were less than 15 percent for dissolved and total aluminum, copper, cadmium, manganese, and zinc. However, at site AR31.0, standard errors for total and dissolved iron were as large as 30 percent and for dissolved zinc, as large as 50 percent. Copper was used as the indicator constituent to determine the divisions for the time intervals for aluminum, iron, cadmium, manganese, and zinc because: (1) Copper was identified as the primary metal of concern by Morrison and Knudsen Corporation and ICF Keiser Engineers (1994); (2) copper was identified as an excellent predictor of the concentrations of iron, cadmium, manganese, and zinc at the Summitville Mine site (Miller and Van Zyl, 1995); and (3) strong cross correlations between concentrations of metals indicate that the predominant source of metals was the same or that similar processes affect the metal concentrations, or both (tables 3 and 4). The time-interval periods were made by analysis of the hydrograph and copper- concentration data, and divisions were made at large changes in copper concentration or where a change in the flow regime occurred. The divisions of time intervals for the other constituents were the same as were determined for copper. Nine time intervals were used to estimate loads at site AR31.0, and 18 time intervals were used to estimate loads at site AR34.5, with each time interval generally containing between 3 and 7 data points. Attempts were made to divide the data so that there were at least three samples in each time interval, allowing for calculation of standard errors of load estimates. However, metal loads associated with storms at site AR34.5 were calculated separately. By removing the metal concentration measured from samples collected during storms from the rest of the data for the time interval, the large metal concentra- tions that were associated with storms did not bias the mean concentration and standard error of the other data. Because the storms frequently contained only one or two samples, standard errors generally could not be determined for these storms. Metal loads varied considerably as a result of changes in streamflow or changes in metal concentrations, or both. Some of the physical and chemical processes that affected metal transport through Terrace Reservoir included mixing, sorption, flocculation, and sedimentation. A summary of metal loads, by flow period, for sites AR34.5 and AR31.0 is included in table 5. Daily load estimates and standard errors for dissolved and total aluminum, iron, copper, cadmium, manganese, and zinc are presented in tables 6 and 7 in the "Supplemental Data" section at the back of this report. Aluminum Large variations in aluminum loads occurred during the study. The largest daily loads of total aluminum were transported into and out of Terrace Reservoir during the peak snowmelt period (fig. 11). During the peak snowmelt period, the maximum daily total-aluminum load that entered the reservoir was about 11 tons (table 7). Large daily total-aluminum loads also were transported into the reservoir during rainstorm runoff during the summer period. Most of the total-aluminum loads that entered the reservoir during rainstorm runoff remained in the reservoir (fig. II). The smallest daily total-aluminum loads occurred during the base-flow period between November 1994 and February 1995 (fig. II). 26 Assessment of Metal Transport Into and Out of Terrace Raaarvolr, Conelos County, Colorado, April 1994 Through March 1995 00318:1