The dataset contains daily values of hydrologic and physical parameters. These daily values were calculated from continuous (15-minute) data collected from continuous water quality monitors (CWQMs) installed at five gaged sites within the Little Blue River watershed in Independence, Missouri. The data were used as input to regression models to predict Escherichia coli (E. coli) loads.
Any data estimation was performed according to procedures described in the WEFTEC 2017 conference proceedings paper "Evaluation of Modeled Bacteria Loads Along an Impaired Stream Reach Receiving Discharge from a Municipal Separate Storm Sewer System in Independence, Mo." Periods of missing daily values were due to equipment malfunction, fouling, or poor quality data. Daily values were estimated for the modeling input data if the period of missing daily values was less than 2 weeks, with one exception of a 3-week period at site T2. Most estimated periods covered 1 to 4 days. Although estimating water-quality parameters is discouraged, missing daily values were estimated for this study to maintain continuity for comparison of modeled results among sites. If a daily value was missing for a parameter used in a regression model, then the model would not be able to estimate a daily load for that day, and the estimated recreation-season load (the sum of all the estimated daily loads during the recreation season) would be incomplete for comparison among sites. Of the data values included in this data release, approximately 1.5 percent of water temperature, 2.0 percent of specific conductance, 1.4 percent of dissolved oxygen, 2.7 percent of pH, and 5.7 percent of turbidity daily values were estimated. Values were estimated during stormflows and base flows with the understanding that more uncertainty was introduced with the estimation of stormflow water-quality values compared to base flow because of the highly variable nature of water-quality parameters during stormflow events. If a site had an extended gap in the data that was unable to be estimated, that parameter at that site was considered unavailable for model calibration. This eliminated all CWQM parameters as calibration options at site T4, and it also eliminated turbidity at site T1 because approximately 1 year is missing in the middle of the record. Estimated values introduce increased uncertainty into the models, but this was considered acceptable given the low percentage of estimated values. Daily estimates were based on any partial record of 15-minute measurements for any day for which daily values had not been calculated. If the parameter being estimated was water temperature, the other sites were referenced and considered while making the estimate (taking into account differences in site conditions). Trends in the daily values before and after the missing value(s) were considered while estimating specific conductance, dissolved oxygen, and turbidity daily values for use in the regression models. When estimating turbidity, similar stormflow events at the same site were used for comparison. For some parameters, other parameters were used to estimate the missing values, such as using the established relationship between dissolved oxygen and water temperature. If sufficient data were not available on which to base the estimation, the daily values were interpolated between days with known values. Citation: Flickinger, A.; Christensen, E. Evaluation of Modeled Bacteria Loads Along an Impaired Stream Reach Receiving Discharge from a Municipal Separate Storm Sewer System in Independence, Mo. Proceedings of the WEFTEC Conference, Chicago, IL, October 1-4, 2017.