TURBIDITY AND SEDIMENT MONITORING
HY 2002 through HY 2004
Sonoma Creek Watershed, California
Sonoma Ecology Center
205 First Street West
Sonoma, California 95476
707.996.0712
www.sonomaecologycenter.org
ATTACHMENTS
TABLES
Table 1. Severity Index
Table 2. Monitoring Locations
Table 3. First and Second Flush Storms, Tributary Values
Table 4. Turbidity/SSC Relationships for Station A
Table 5. First Flush Storms, Station A
Table 6. SSC Exposures, Sixteen Storms, Station A, HY 2002
Table 7. SSC Exposures, Nine Storms, Station A, HY 2003
Table 8. SSC Exposures, Twelve Storms, Station A, HY 2004
Table 9. Severity Indices, Sixteen Storms, Station A, HY 2002
Table 10. Severity Indices, Nine Storms, Station A, HY 2003
Table 11. Severity Indices, Twelve Storms, Station A, HY 2004
PLATES
Plate 1. Location Map
Plate 2. Grab Sampling Photographs
Plate 3. Measuring Stream Height
Plate 4. Automated Data Collection Station A and Depth-Integrated Sampler
Plate 5. Grab Turbidity (NTU) versus Suspended Sediment Concentration (mg/L)
Plate 6. SSC/Turbidity (Power Curve), Station A
1.0 Introduction
The Sonoma Ecology Center (SEC) is testing the hypothesis that elevated levels of turbidity and suspended sediment concentration (SSC) in stream water pose adverse impacts to salmonids of concern in the Sonoma Creek watershed (chinook salmon [Oncorhynchus tshawytscha] and steelhead trout [Oncorhynchus mykiss]). Turbidity and SSC are related: water containing SSC up to approximately 50 milligrams/liter (mg/L) shows a 1:1 ratio with turbidity measured in nephelometric turbidity units (NTUs); above 50 mg/L, the relationship between SSC and NTU is no longer 1:1 and becomes variable according to stream location (Fitzgerald, 2004). Water containing SSC greater than 27 mg/L has been defined as “turbid,” has been characterized as “not drinkable,” has been observed to effect a 50 percent drop in the catch of fish, and leads to a 10 percent drop in fish production (Anderson, 1975).
Research has correlated elevated turbidity and SSC to observed effects on sample salmonid populations, such as reduced feeding and growth rates, avoidance of turbid waters, or death. From their analysis of the published results of 80 separate studies, Newcombe and Jensen (1996) developed a severity index for ranking and analyzing the effects of excess SSC on salmonids (Table 1). The various ranks in the index have been examined as thresholds for salmonid behavior. California Regional Water Quality Control Board (RWQCB) staff have suggested that a severity index rank of 4 or greater poses significant harm to salmonids and the cold water fishery by effecting (1) salmonid mortality, (2) reductions of growth in the smolt size, (3) overall long-term reduction in growth (Fitzgerald, 2004). Trush (2001) proposed that long-term reduction in feeding rates and success (equated with a Newcombe and Jensen severity ranking of 8) results in smaller salmonids with higher mortality rates.
More information about the Newcombe and Jensen severity index is in Section 4.0, Discussion.
2.0 Methods
To characterize existing levels of turbidity and SSC in Sonoma Creek and selected tributaries, SEC implemented a program of grab sampling, depth-integrated sampling, and automated data collection in HY 2002 through HY 2004.
Our study design is explained in the Quality Assurance Project Plan (QAPP) prepared with technical assistance from California Water Resources Control Board. The QAPP is appended to the Final Report of Volunteer Monitoring of Suspended Sediment Concentration and Turbidity prepared for California Regional Water Control Board, San Francisco Bay Region (Lawton et al., 2002). The report and QAPP are available for downloading from the SEC website, www.sonomaecologycenter.org, or by calling SEC at 707.996.0712. Methods for each aspect of the study are summarized below.
Grab Sampling
SEC staff established turbidity and SSC grab-sampling stations at 21 monitoring locations in the Sonoma Creek watershed (Plate 1 and Table 2). Sixteen of the 21 locations were sampled regularly during wet storms; all 21 were sampled only during major storm events and when staffing allowed (Table 2). 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 sample bottle with stream water (Plate 2). 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]).
When taking grab samples at monitoring locations throughout the watershed, SEC staff used a staff plate or wire-weight gauge to measure stream height (Plate 3). Once stream height was established, the cross-sectional area of stream flow was estimated from previously established measurements of channel cross section at each site (Orme et al., 2004). Staff clocked orange-peel floats on the stream surface and averaged velocity of the floats to complete the calculation Discharge= Velocity x Area.
Automated Data Collection
The automated data collection system at Station A (STA) at Sonoma Valley Watershed Station, Eldridge, California (Plates 1, 4), measures stream depth, turbidity, and rainfall. Readings were taken at 10-minute intervals in HY 2002 and at 15-minute intervals in HY 2003 and HY 2004. Data was downloaded weekly from STA and uploaded into Excel spreadsheets on computers at the Sonoma Valley Watershed Station.
Depth-Integrated Sampling
A depth-integrated (DI) sampler was lowered through the water column to the stream bottom to collect samples at 5-foot intervals across the stream cross section at STB (Plate 4). Samples were collected only during and directly following wet storms.
Assessing Biological Significance
To arrive at a ranking in the severity index (Table 1), we analyzed automated turbidity data from STA along with correlated SSC values using methods developed by Newcombe and Jensen (1996). First we calculated a suspended sediment dose index for STA based on hours of aquatic exposure to SSC:
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 dose index 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 and Jensen severity index (1996) based on analyses of salmonid species in the Russian River watershed (chinook salmon, steelhead trout, and coho salmon [Oncorhynchus kisutch]). Only coho salmon were studied sufficiently to make a strong correlation between changes in environment and biological response; for coho salmon, a dose index of 4.55 correlates to a severity index of 4. However, studies for all three species show that, as the SSC dose index increases, so do the symptoms observed and ranked in the severity index. For all species 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 correlates to a severity index of 4.17.
3.0 Results
Datasets from the study include (1) three seasons (HY 2002 through HY 2004) of grab sampling data (discharge, turbidity, and suspended sediment concentration [SSC]); (2) three seasons of DI samples (turbidity, SSC) collected at STB (Sonoma Creek at Harney Road, Plate 1); and (3) three seasons of automated weather data (stream depth, turbidity, and rainfall) collected at STA.
Grab Sampling
Results of grab sampling for turbidity and SSC are shown on Plate 5. The number of grab samples taken at each location ranges from 11 samples collected on Graham Creek at the Emery Restoration Site to 150 samples collected at STA. Our confidence in the data (shown by R-squared values) range from 0.6111 at Nathanson Creek at Nathanson Creek Park to 0.9234 at Asbury Creek at Arnold Drive. SSC:turbidity ratios estimated from the trendline equations vary with location. Lower SSC:turbidity ratios were observed at Carriger Creek at Arnold Drive (27 mg/L:52 NTU) and Calabazas Creek at Dunbar Road (27 mg/L:49 NTU). Higher SSC:turbidity ratios were observed at Sonoma Creek above the Yulupa Creek confluence (27mg/L:23 NTU) and Mill Creek at Redwood Street (27mg/L:29 NTU). These latter correlations more closely approximate the previously cited 1:1 rule of thumb for SSC < 50 mg/L and turbidity < 50 NTU (Fitzgerald, 2004).
Grab sampling during first- and second-flush storms showed peak SSC values exceeding those of storms later in the season. Example first- and second-flush values are in Table 3. Maximum SSC values on tributaries exceeded 4,000 mg/L on Asbury Creek and 1,100 mg/L on Mill Creek, with values at other locations ranging from approximately 50 to 500 mg/L (Table 3).
Depth-Integrated Sampling
Results of DI sampling for turbidity and SSC at STB are shown on Plate 5. Forty-five samples collected at peak-flow times indicated a SSC:turbidity ratio at STB (27 mg/L:37 NTU) similar to that of STA directly upstream (27 mg/L:35 NTU).
Automated Data Collection
Results of automated data collection at STA are shown on Plate 6. Summaries from sixteen wet storms in HY 2002, nine wet storms in HY 2003, and twelve wet storms in HY 2004 indicate SSC (in grab samples from STA) related to turbidity (in automated readings at STA) in a natural log relationship that fits a power curve (Plate 6):
SSC (mg/L) = 0.0116*(turbidity in NTU)1.9902
Example SSC/turbidity relationships derived from the power curve equation are listed in Table 4.
Turbidity was measured in first flush storms at STA at the automated station’s maximum reading of 2,000 NTU, indicating peak SSC values on the mainstem on the order of 1,000 mg/L (Table 5). However, SSC values in grab samples were measured at STA in first flush storms as high as 4,043.4 mg/L (February 16, 2004). A new probe installed at STA with the capacity to measure to 4,000 NTU will help capture higher-end readings.
Turbidity/SSC levels dropped at STA after storms such that 90 percent of maximum values were generally reached within 0.25 hour (15 minutes), 50 percent in less than 3 hours, and 0 percent (total clearing) in less than 30 hours (Lawton et al., 2002). Total clearing occurred in an average of 15.5 hours in HY 2002, 7.1 hours in HY 2003, and 29.7 hours in HY 2004.
Assessing Biological Significance
Turbidity/SSC exposures for aquatic organisms at STA are shown in Tables 6 through 8. Values are sorted in bins according to thresholds used by Newcombe and Jensen (1996). Duration of exposure at each automated turbidity value incorporates duration in lower-value bins; therefore, duration times are cumulative as turbidity/SSC values increase.
The longest turbidity/SSC exposures observed at STA were 305.83 hours for HY 2002, in the season’s longest and wettest storm (HY 02 Storm 11, December 30, 2001, to January 6, 2002; Table 6); 166 hours for HY 2003, in the first major, longest, and wettest storm of the season (HY 03 Storm 1, December 16 to 23, 2002; Table 7); and 75 hours for HY 2004 (HY 04 Storm 9, February 16 to 19, 2004; Table 8). The maximum severity index observed at STA was 8.94 for HY 2002 (HY 02 Storm 11; Table 9); 9.04 for HY 2003 (HY 03 Storm 1; Table 10); and 8.20 for HY 2004 (HY 04 Storm 9; Table 11). Severity indices were above 4 in most storms and occasionally above 8 (Tables 9 through 11).
4.0 Discussion
As stated previously, Trush (2001) proposed that long-term reduction in feeding rates and success (equated with a Newcombe and Jensen severity ranking of 8) directly results in smaller salmonids with higher mortality rates. Therefore, salmonids exposed to situations with ranking 8 or higher on the Newcombe and Jensen severity index are subject to major physiological stress. Other observers (Ligon, 2004) have noted that the severity index rankings are most significant when resulting in major effects (e.g., at a Newcombe and Jensen severity index of 10 or above, resulting in direct mortality).
Extended times of elevated SSC exposures that result in severity indices of 4 and above also pose concern for salmonids. Newcombe and Jensen observed that a severity index of 4 results in short-term reduction in feeding rates and/or feeding success (Newcombe and Jensen, 1996). The RWQCB has suggested 4 as a potential severity index target for the protection of salmonids in the cold-water fishery (especially with regard to coho salmon; Fitzgerald, 2004).
Winter storms in Sonoma Creek in HY 2002, 2003, and 2004 were observed to sustain SSC exposures that cause minor to major physiological stress in salmonids, moderate habitat degradation, possible impaired homing, and poor condition. Severity indices for STA on mainstem Sonoma Creek have not been observed to exceed 9. Therefore SSC alone probably does not account for reduced growth rate, delayed hatching, reduced fish density, severe habitat degradation, and direct mortality. SSC exposures do, however, appear to be contributing intermittent salmonid stressors, potentially significant when severity indices are highest, i.e., in first and second flush events (long, early [usually December] storms). Between storms, SSC exposures on the mainstem generally drop within 30 hours to less significant levels.
5.0 Recommendations
Future water-quality sampling should focus on better understanding human impacts on sediment delivery to waterways. We specifically recommend that future turbidity and SSC monitoring should include the following:
· Increased sample collection in major tributaries to Sonoma Creek to estimate severity indices in these contributing waterways
· Increased sample collection at STB (e.g., DI sampling on at least falling limb of four to five storms across STB cross section) to correlate to grab and automated samples at Eldridge sampling locations
· Continued use of the automated station to monitor all previously measured parameters at STA
· Analysis of data from an additional turbidity probe installed in November 2004 at STA to back up the existing probe
· Correlation of sample results to a RUSLE-style model of sediment production for Sonoma Valley that includes land-use map coverage from the SEC GIS database
· Correlation of SSC results at each location to stream discharge.