USGS Washington Water Science Center
Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
By Christopher A. Curran, Christopher S. Magirl, and Jeffrey J. Duda
Measured suspended-sediment concentrations correlated best with turbidity when transformed using a cube-root model (fig. 2). The fine-grained portion (<0.0625 mm) of measured concentration was better correlated with turbidity than the total concentration. Twenty-three suspended-sediment samples (appendix A) were used to generate the regression model with total concentration ranging from 7 to 11,600 mg/L and a median concentration of 1,540 mg/L (table 1). The cube-root regression was then applied to the time series record of 15-min turbidity values from September 15, 2011, to September 10, 2013. The computed suspended-sediment concentration of time-series data ranged from 54 to >5,700 mg/L and the portion of fine-grained sediment concentrations ranged from 53 to >2,770 mg/L (table 1). Uncertainty in predicted suspended-sediment concentration, reported as 90-percent prediction intervals, was significant and ranged from 0 to 441 mg/L when the turbidity was 2 FNU (total suspended-sediment concentration of 57 mg/L), from 798 to 4,320 mg/L when the turbidity was 800 FNU (total suspended-sediment concentration of 2,180 mg/L), and from 2,750 to 9,250 mg/L when the turbidity was 1,200 FNU (predicted suspended-sediment concentration of 5,530 mg/L). All suspended-sediment concentration values and uncertainties, for both total and fine-grained portion, were calculated for each available 15-min turbidity measurement and are available in appendix B
Total suspended-sediment concentrations in the lower Elwha River were tabulated into concentration bins, which were summarized by proportion of time per week (fig. 3). There were a total of 106 weeks in the record, with 100 weeks containing the full 7 days (the other 6 weeks contained less than 7 days of data due to instrument issues or calendar artifacts). Continuous data collection was interrupted during 15 weeks, with the number of 15-min intervals collected during these partial weeks averaging 649 (range 503-669) of a maximum possible 672 15-min turbidity samples per week.
From the start of the river-restoration project until Elwha Dam was completely removed in April 2012, suspended-sediment concentrations in the lower Elwha River generally remained less than 500 mg/L (fig. 3). High flows and increased available sediment from reservoir drawdown in spring of 2012 created hydraulic conditions whereby suspended-sediment concentrations periodically exceeded 500 mg/L. Suspended-sediment concentrations were generally less than 500 mg/L in the summer and early autumn of 2012, with concentrations less than 100 mg/L common from late August to early October 2012. By late October 2012, the Lake Mills reservoir had drained completely and coarse-grained sediment began passing the remaining Glines Canyon Dam structure in large volumes. High flows in November and December 2012 created suspended-sediment conditions in the lower river consistently greater than 1,000 mg/L and often exceeding 5,000 mg/L. Relatively low flow in February allowed a slight decrease in suspended-sediment concentration, but late spring rainfall combined with larger-than-normal snowmelt caused suspended-sediment concentrations to rise above 1,000 mg/L for much of the time period from March to June 2012 (fig. 3). The lower river cleared in the late summer 2013 with suspended-sediment concentrations commonly less than 100 mg/L during several weeks in August and early September 2013.