Water quality in the upper Shoal Creek basin, southwestern Missouri, 1999-2000

Results of a water-quality investigation of the upper Shoal Creek Basin in southwestern Missouri
indicate that concentrations of total nitrite plus nitrate as nitrogen (NO2t+NO3t) in water samples
from Shoal Creek were unusually large [mean of 2.90 mg/L (milligrams per liter), n (sample size)=60] compared to other Missouri streams (mean of 1.02 mg/L, n=1,340). A comparison of instantaneous base-flow loads of NO2t+NO3t indicates
that at base-flow conditions, most NO2t+NO3t discharged by Shoal Creek is from nonpoint sources. Nearly all the base-flow instantaneous
load of total phosphorus as P (Pt) discharged
by Shoal Creek can be attributed to effluent from a municipal wastewater treatment plant. Samples collected from a single runoff event indicate that substantial quantities of Pt can be transported during runoff events compared to base-flow transport. Only minor quantities of NO2t+NO3t are transported during runoff events compared to base-flow transport. Fecal coliform bacteria densities at several locations exceed the Missouri Department of Natural
Resources (MDNR) standard of 200 col/100 mL (colonies per 100 milliliters) for whole-body contact recreation. During 13 months of monitoring
at 13 stream sites, fecal coliform densities (median of 277 and 400 col/100 mL) at two sites (sites 2 and 3) on Shoal Creek exceeded the MDNR standard at base-flow conditions. The maximum fecal coliform density of 120,000 col/100 mL was detected at site 3 (MDNR monitoring
site) during a runoff event in April 1999 at a peak discharge of 1,150 ft3/s (cubic feet per second).
Fecal coliform densities also exceeded the MDNR standard in three tributaries with the largest
densities (median of 580 col/100 mL) detected in Pogue Creek. Results of ribopattern analyses indicate that most Escherichia coli (E. coli) bacteria in water samples from the study area probably are from nonhuman sources. The study area contains about 25,000 cattle, and has an estimated annual production
of 33 million broilers and 300,000 turkeys. Probable nonhuman sources included turkeys, horses, chickens, and cattle; however, wildlife sources such as deer, raccoon, muskrat, and opossum
were not evaluated. Human waste was an important source of E. coli in water samples collected
at the MDNR monitoring site (site 3) on Shoal Creek and at two tributary sites (Joyce Creek and Woodward Creek). In general, the detection of human ribopatterns was consistent with the detection of organic compounds commonly
associated with human wastewater such as caffeine, triclosan, or phenol, and the fecal indicators
cholesterol and 3B-coprostanol. Ribopattern analysis indicate that horses were an important source of E. coli in Woodward Creek, which was consistent with horses being pastured immediately upstream from the sampling site on this creek. Pogue Creek contains a large density of turkey barns and five of eight E. coli isolates from one sample from Pogue Creek were matched to turkeys.
Water samples from Pogue Creek generally did not contain detectable concentrations of human wastewater compounds, but one sample did contain detectable quantities of the antibiotics tylosin and lincomycin (widely used in the animal industry), and sulfamethoxazole (human use only). Although promising, the ability of ribopattern
analyses to positively identify the source of a particular isolate is uncertain because of the small sample size, possible differences between animal source patterns in the study area and database used, lack of native wildlife source patterns, and variation in results depending on the number of possible animal host considered. Results of this study indicate that a trend of increasing fecal coliform densities with increasing time detected by the MDNR is, in part, caused by trends in annual precipitation and stream discharge,
and not necessarily changes in land use or densities of animal operations. A multiple linear regression (MLR) model using specific conductance
and wate