Water-quality data consisted of nutrient-concentrations from water samples collected during the study in-situ and water-quality measurements. Concentrations of constituents in duplicate samples were within 5 percent of original sample concentrations, and equipment blank sample concentrations indicated there was no significant contamination from the sampling procedure. Nutrient-concentrations and the in-situ water-quality parameters generally exhibited seasonal variations, although short-term variation of concentrations and water-quality measurements also occurred.

Water-Quality Characteristics

Broad, Harney, and Shark Rivers were well mixed, with a difference in specific conductance from top to bottom usually no greater than 500 µS/cm during flood and ebb tides. Lostmans Creek and North River were usually well mixed, but there were occurrences of slightly stratified flows (top-to-bottom difference 2,000 µS/cm) near the beginning of the wet season, or after significant rainfall events. Specific conductance at the Lostmans Creek and Broad River stations remained at 250 to 800 µS/cm from August through December 1999 because of rainfall from the summer season and two late-season tropical systems (39 in. of rainfall). Specific conductance at the Lostmans Creek and Broad River stations was less than 1,000 µS/cm for 30 percent of the study period.

All stations had relatively high total nitrogen concentrations of about 1 to 2 mg/L and very low total phosphorus concentrations, typically less than 0.04 mg/L (tables 5 and 6). Concentrations for total nitrogen and total phosphorus were very similar in magnitude and range at all five stations. Infrequently, a duplicate sample's phosphorus concentration was 10 to 20 percent different than the original sample concentration. Discussions with the USGS laboratory personnel identified possible interference with the low-level phosphorus analysis method, when used on saline water, that could result in erroneous phosphorus concentrations. Nitrite and nitrate concentrations were detected at very low levels and ranged from below the detection limit of 0.002 to 0.070 mg/L.

Total nitrogen concentrations were approximately 75 to 85 percent dissolved nitrogen, and total phosphorus concentrations were approximately 35 to 50 percent dissolved phosphorus. Dissolved nutrients are readily available to plants, and analyses indicated that approximately three-quarters of the total nitrogen and half of the total phosphorus are available in the dissolved form.

Total nitrogen concentrations for the five stations generally were inversely related to specific conductance (saltier water had lower nitrogen concentrations), but the relations were not statistically significant. However, at four of the five stations (all but the Shark River station), total phosphorus concentrations were directly related to specific conductance and the relations were statistically significant (p = 0.0 to 0.11), which means that as the water became saltier (increased specific conductance water from the Gulf of Mexico), the total phosphorus concentrations increased. Two models were tested for estimating total measured phosphorus flux. One model used specific conductance as the independent variable to estimate phosphorus concentration for any given specific conductance. The second model is a more direct method that used discharge as the independent variable to estimate total phosphorus flux for any given discharge. In the second method, nutrient-concentration is multiplied by instantaneous discharge at the time of sample collection and a conversion factor to yield flux in short tons (2,000 pounds) per day. The flux values are then used as the dependent variable and discharge is the independent variable, and a linear regression equation is used to describe the relation (Helsel and Hirsch, 1992). The two models gave similar results for monthly mean phosphorus flux. The second more direct method of estimating total nitrogen and total phosphorus flux was selected to estimate total station fluxes.