Suspended sediment concentration and turbidity during storms.
Version November 2014, completed for EPA grant SFEP E2100-2, funded by the EPA San Francisco Bay Water Quality Improvement Fund.
Summary of methods:
Former SEC staffer Rebecca Lawton, who initiated SEC’s sediment monitoring program and ran it for 10 years, trained SEC staff and volunteers from Glen Ellen Rotary in the office and at Station A on December 4, 2013 to assure compliance with the suspended sediment and turbidity QAPP. The relevant QAPP isQuality Assurance Project Plan for Volunteering Monitoring of Suspended Sediment Concentration and Turbidity, Sonoma Creek Watershed, Sonoma County, CA.Staff and volunteers collected water sample bottles (for suspended sediment concentration measurement in the lab) and turbidity measurements around the clock at five stations along Sonoma Creek during the storm of February 7-10, 2014. Turbidity measurements are obtained in the field using a turbidometer, with no samples brought back from the field. Suspended sediment concentration requires lab processing. Staff and volunteers processed the over 100 sample bottles in SEC’s lab. Processing entails weighing the water in the sample bottle, filtering the sediment from the water, drying the sediment, and weighing the sediment. The weights of the water and the sediment enable calculation of the weight of sediment per volume of water. Staff and an intern conducted data analysis and produced summary graphs comparing 2014 findings with past findings.The results are not significantly different from past findings: acute suspended sediment levels during storms are physiologically harmful to steelhead. Data, methods, findings, and graphs are available on the Sonoma Valley Knowledge Base.
Previous Work:
Collecting, processing, and analyzing suspended sediment concentration and turbidity data for this project continues Sonoma Ecology Center’s monitoring of these parameters begun in 2002, overseen for 10 years by Rebecca Lawton. Past findings informed foundational Sonoma Ecology Center documents: the Sediment Source Analysis (Lawton, editor, 2006) and the Limiting Factors Analysis for Steelhead (Micheli, editor, 2006). Both documents informed the analysis conducted by the San Francisco Bay Regional Water Quality Control Board for the sediment Total Maximum Daily Load (TMDL) for Sonoma Creek.
Grab Sampling:
“SEC staff established turbidity and SSC grab-sampling stations at 6 monitoring locations in the Sonoma Creek watershed. Grab sampling, primarily done during and directly following rainfall, consisted of the simultaneous filling of one 15-milliliter (mL) HACH cell and one 500-mL SSC (Suspended sediment concentration) sample bottle with stream water. Turbidity cells were analyzed in the field using a HACH 2100P turbidimeter. SSC sample bottles were delivered under chain-of-custody protocols to the M.U.D. Laboratory at the Sonoma Valley Watershed Station, Eldridge, California. Methods used for analyzing the SSC grab samples derive in part from the Redwood Sciences Laboratory Standard Operating Procedures for SSC Determination and from Standard Methods (2540B-Total Solids Dried at 103 to 105 degrees Celsius [C]).”
Field and lab data were inputted word for word into Excel. Data was organized by station and time and was cross checked for input and math errors. 10% of all values were checked for accuracy. For samples that were split, the sediment weights were merged. For each final sediment weight (g) run a conversion to SSC using the following formula as found in the 2011 QAPP.
Using this data, linear regressions were run for each station and for all of the stations together.
Automated Data Collection:
The automated data collection system at Station A (STA) at Sonoma Valley Watershed Station,
Eldridge, California, measures stream depth, turbidity, and rainfall. Readings were taken at 15-minute intervals. Data was downloaded monthly from STA and uploaded into Excel spreadsheets on computers at the Sonoma Valley Watershed Station.
There are 2 turbidity sensors at STA, each with a high and low sensitivity reading. These report data in millivolts, which can be converted to NTUs based on the conversion factors given by the manufacturer. Since there are three different conversion factors listed for different readings, an overall formula was derived and a conversion factor was made for each millivolt reading.
NTU conversion= 7E-05(mV) + 0.1111
NTU = NTU conversion * mV
The maximum value that the current sensor can read is 2,000 NTU, therefore the next step was to throw out any data greater than that. The data was then smoothed – using the grab samples as the standard, each of the 4 sensor readings were examined to see which was the most accurate. If they were all close, or more than one of them were close, an average was taken. If none of them were accurate, an average was taken of the previous and next time stamps. The low sensitivity reading is nominally 1/4 the value of the high sensitivity readings. As the water becomes cloudier, the high sensitivity reading will saturate and flat-line. At that point we switched over to using the low-sensitivity channel (Allen, 2014). ). Data smoothing was completed from the time of the first rainfall to when NTU is less than 49 based on the previous protocols established (Lawton & Flores, 2006).
The sample time of each grab sample at STA was matched to the nearest automated sample, accounting for daylight savings time. Once matched, we created a graph of SSC (in grab samples from STA) against turbidity (in automated readings at STA) in a natural log relationship that fits a power curve.
Assessing Biological Significance:
As in past analyses, we assessed the meaning and significance of the stream condition data by correlating it directly to impacts on salmonids, as empirically derived by Newcombe and Jensen (1996) and communicated using their severity index.
To arrive at a ranking in the severity index, we analyzed automated turbidity data from STA along with correlated SSC values using methods developed by Newcombe and Jensen (1996). To start, data from the USGS Kenwood gauge was used to determine when discharge approximated 100 cubic feet per second and to identify potential first flush storms for past years. Data to match the first flush was then found in the STA records. Duration of each storm was calculated from first rainfall (using the rain gauge at STA) to when turbidity approximated 49 NTU.
Each previous year’s storm was smoothed in the process described above. SSC was calculated for each automated NTU value using the relationship derived in the power curve. The SSC values were sorted by value and then placed in bins set by Newcombe and Jensen using the frequency function of excel. Exposure to each SSC bin range was calculated.
From hours exposed and SSC, a suspended sediment dose index for STA was calculated:
Suspended Sediment Dose Index = natural log (SSC x Hours Exposed)
(Examples: Using this equation, exposure to 3.13 mg/L SSC for 24 hours results in a doseindex of 4. Similarly, exposure to 75.19 mg/L SSC for just 1 hour results in a dose index of 4[Fitzgerald, 2004].)
Next we correlated the calculated SSC dose indices from Sonoma Creek to the Newcombe andJensen severity index (1996) based on analyses of salmonid species in the Russian Riverwatershed (Chinook salmon, steelhead trout, and coho salmon [Oncorhynchuskisutch]). Onlycoho salmon were studied sufficiently to make a strong correlation between changes inenvironment and biological response; for coho salmon, a dose index of 4.55 correlates to aseverity index of 4. However, studies for all three species show that, as the SSC dose indexincreases, so does the severity of the symptoms observed and ranked in the index. For allspecies studied, the severity index ranking correlates to the SSC dose index as follows:
Severity Index Ranking = 0.7491(SSC dose index) + 0.7625
Using this relationship, an SSC dose index of 4.55 in the Newcombe and Jensen analysis correlates to a severity index of 4.17.Duration of exposure at each automated turbidity value incorporates duration in lower-value bins; therefore, duration times are cumulative as turbidity/SSC values increase.
References
Allen, T., 2014. Personal communication with Alex Young of Sonoma Ecology Center. July 7.
Fitzgerald, R., 2004. Salmonid Freshwater Habitat Targets for Sediment-Related Parameters. Draft. Prepared for State Water Resources Control Board, North Coast Region. October.
Lawton, R. (editor). 2006. Sediment Source Analysis, Sonoma Creek Watershed, California. Sonoma Ecology Center, Sonoma, California.
Lawton, R., R. Hunter, and J. Menze, 2002. Final Report, Volunteer Monitoring of Suspended Sediment Concentration and Turbidity and Watershed Monitoring of Road Remediation in Annadel State Park, Sonoma Creek Watershed, Sonoma County, California. Prepared for the Sonoma Ecology Center and Regional Water Quality Control Board, San Francisco Bay Region. September.
Lawton, R., and V. Flores, 2006. Suspended Sediment Concentration. Appendix D to Limiting Factors Analysis. Prepared for the Sonoma Ecology Center and Regional Water Quality Control Board, San Francisco Bay Region. December.
Lawton, R., 2014. Personal communication with Alex Young of Sonoma Ecology Center. July 1.
Micheli, E. (editor). 2006. Sonoma Creek Watershed Limiting Factors Analysis. Sonoma Ecology Center, with Stillwater Sciences and UC Berkeley Dept. of Earth and Planetary Sciences. Sonoma Ecology Center, Sonoma, CA.
Newcombe, C.P. and J.O.T. Jensen, 1996. Channel Suspended Sediment and Fisheries: A Synthesis for Quantitative Assessment of Risk and Impact. North American Journal of Fisheries Management. 16(4):693-727.