|
<br />municipal well field that is operated by the Eagle
<br />River Water and Sanitation District. Ground-water
<br />data were available for field properties, major ions,
<br />nutrients, trace elements, pesticides, DOC, VOCs,
<br />bacteria (total coliforms and Escherichia coli
<br />[E. coli]), methylene blue active substances (MBAS),
<br />and chlorofluorocarbons at the USGS sites.
<br />Aquatic-ecology data (stream and riparian
<br />habitat, and algae, macroinvertebrate, or fish commu-
<br />nity) were available for 36 sites in the Gore Creek
<br />watershed (fig. 6, table 1). Fish-community data at one
<br />site were available for 1995-98. Algae, macroinverte-
<br />brate, and habitat data were available from a synoptic
<br />study done in September 1997. Discussion of histor-
<br />ical and current aquatic ecology in this report focuses
<br />on sites in the main stem of Gore Creek as it flows
<br />from its headwaters through the Town of Vail and sites
<br />in Black Gore Creek as it flows from Vail Pass to its
<br />confluence with Gore Creek along Interstate 70. Most
<br />of the data from this synoptic study were collected by
<br />the USGS for the Town of Vail, as part of a national
<br />study by the U.S. Environmental Protection Agency
<br />(USEPA) to assess the ecological benefits of storm-
<br />water controls such as natural vegetation buffer strips
<br />or swales (Watershed Management Institute, unpub.
<br />data, 1996). Interpretation of the tributary-stream data
<br />is beyond the scope of this report.
<br />
<br />METHODS OF DATA REVIEW AND
<br />ANALYSIS
<br />
<br />Water-quality properties and constituents are
<br />presented graphically and statistically in this report.
<br />The surface- and ground-water-quality data were
<br />quality assured by examining the total cation and total
<br />anion concentrations and the total nitrogen and total
<br />phosphorus concentrations in all samples with avail-
<br />able data. For data used in this report, differences in
<br />total cation and total anion concentrations were less
<br />than 10 percent. Fewer than 5 percent of the available
<br />nutrient samples were discarded for use in this report
<br />because total nitrogen or total phosphorus concentra-
<br />tions were less than the sum of the constituents that
<br />make up the total concentration.
<br />Water-quality data were analyzed using
<br />nonparametric statistical methods. Nonparametric
<br />statistical analyses of rank-transformed data are not
<br />unduly affected by outliers and are not dependent on a
<br />normal distribution of the data. During assessment of
<br />the spatial distribution of individual nutrient species,
<br />greater than one-half of the data for some sites were
<br />
<br />censored (reported below a laboratory reporting limit).
<br />If more than one-half the data for a particular nutrient
<br />constituent were censored at a site, estimates of the
<br />10th, 25th, 50th (median), 75th, and 90th percentiles
<br />were calculated using the maximum likelihood
<br />estimation (MLE) method (Helsel, 1990; Helsel
<br />and Cohn, 1988). These estimated percentiles were
<br />used to construct the boxplots and provide median
<br />concentrations plotted on the maps. Statistical compar-
<br />ison of streamflow relations to nitrate and total phos-
<br />phorus concentrations was performed using the
<br />Mann-Whitney U test, which is a nonparametric
<br />version of the two-group unpaired t-test (Helsel
<br />and Hirsch, 1992). For this test, censored values were
<br />treated as equal to the reporting limit for that sample.
<br />Long-term temporal differences in the concentrations
<br />of ammonia, nitrate, orthophosphate, total phosphorus,
<br />and specific-conductance values were compared statis-
<br />tically for the 1968-97 period by using Tukey's
<br />Significant Difference Test (Tukey test) on the rank-
<br />transformed data (Helsel and Hirsch, 1992). Because
<br />of a high percentage of censored data values during
<br />the spring and summer seasons, and because higher
<br />nutrient concentrations occur typically during low
<br />flow, only the winter season (November-April) was
<br />tested statistically for trends for each constituent. For
<br />the Tukey test, censored values were treated as equal
<br />to the reporting limit. Flow-adjusted concentrations
<br />were not used to evaluate temporal changes in nutrient
<br />concentrations because much of the available data
<br />lacked concurrent streamflow information and also
<br />because a high percentage of the data was censored.
<br />Boxplots were used to display the central
<br />tendency and variability in specific-conductance
<br />values and nutrient concentrations. Boxplots (for
<br />example, fig. 13) graphically show the central
<br />tendency of the data (the median, or 50th percentile
<br />line of the box, marked with a diamond for clarity), the
<br />variation of the data (interquartile range, or the box
<br />height between the 25th and 75th percentiles), the 10th
<br />and 90th percentiles (shown by whiskers below and
<br />above the box), and the skewness (quartile skew, or the
<br />relative size of the box halves divided by the median
<br />line) (Helsel and Hirsch, 1992). Data points plotted
<br />below and above the boxplot whiskers represent data
<br />values that are less than the 10th and greater than the
<br />90th percentile, respectively.
<br />The water-quality data were compared with
<br />drinking-water and stream-quality standards where
<br />applicable. Drinking-water standards are set by the
<br />USEPA and include the primary (MCL), secondary
<br />
<br />18 Gore Creek Watershed, Colorado-Assessment of Historical and Current Water Quantity, Water Quality,
<br />and Aquatic Ecology, 1968-98
<br />
|