Columbia Slough Sediment Analysis
2006 Sampling
Prepared by:
City of Portland Bureau of Environmental Services January 2009
Prepared for:
Oregon Department of Environmental Quality
Page intentionally left blank
EXECUTIVE SUMMARY The City of Portland, Bureau of Environmental Services (BES) and the Oregon Department of Environmental Quality (DEQ) have been engaged in investigating and improving sediment quality in the Columbia Slough for over ten years. During this time, extensive remedial investigations, focused investigations, and feasibility studies have been conducted. BES is currently participating in the state’s Voluntary Cleanup Program to address contaminated sediments in the Slough. As part of the program, BES and DEQ signed an Intergovernmental Agreement (IGA) which describes the actions each agency will take to address sediment quality including conducting long-term monitoring. The Long-term Monitoring Plan, which is an element of the remedial approach for the Columbia Slough, includes sampling fish and sediment every 10 years to assess progress in achieving protective levels, evaluate spatial and temporal trends, and identify areas where more focused remedial efforts may be warranted. During the summer of 2006, BES collected 78 surface sediment samples for analysis of contaminants of interest (COI). The sediment samples were analyzed for polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs), metals, and pesticides. The first broad-scale sampling event was conducted in 1994. These data were comprised of sediment samples from 10 segments/reaches throughout the slough. The 2006 samples consisted of 78 sediment samples from the same reaches. In the 2006 sediment data, metals and PAHs were nearly always detected, pesticides ranged widely in percent detection, and many PCB Aroclors were never detected in any sample while others were detected infrequently. For the pesticides, several DDT compounds, chlordanes, aldrin, dieldrin and endosulfan II were some of the more frequently detected contaminants, and were detected in over half of samples. The 2006 sediment data were compared to the Columbia Slough Screening Levels. For the metals, chromium, copper and zinc showed the highest exceedence of the screening levels, with approximately 20% of the samples exceeding for each. The PAHs were typically measured at levels above the screening levels. For the pesticides, gamma chlordane (63%), 4,4’-DDE (62%), alpha chlordane (58%) and dieldrin (51%) were the analytes most frequently detected above their screening levels. Since PCBs were infrequently detected they did not have a high percentage of samples above screening levels, but for Aroclors 1248 and 1260 all detected values were above the screening level (i.e., % detection was equal to % exceed). A little less than half of the detected values for Aroclor 1254 exceeded the screening level. Spatial analysis of the 2006 data indicate a number of areas within the Slough with elevated contaminant concentrations. Whitaker Slough had high levels of DDT compounds, chlordanes, aldrin, dieldrin, endosulfan II, endosulfan sulfate and many PAHs in comparison to the rest of the Slough. The Lower Slough in the I-5/MLK reach had high levels of PAHs, PCB Aroclors 1248 and 1254, copper and zinc. The eastern end of Wapato Slough had the highest concentration of PCB Aroclor 1248, chromium, copper, lindane, heptachlor, naphthalene and nickel. For nickel, the level at this station was over an order of magnitude higher than the rest of the Slough. For lindane and heptachlor, it was the only of the 78 stations across the entire Slough where these two analytes were detected, and the detected value at this site was over two orders of magnitude higher than the detection limit. In general the pattern of sediment contamination in 2006 is consistent with the pattern of sediment contamination in 1994. The key risk drivers and contaminants of concern identified in 1994 continue to exceed screening levels in 2006. Spatial variability in contaminants was roughly similar in the two sampling events. However, there was a consistent pattern of higher contaminant concentrations in 2006 than in 1994. Many contaminants were detected at significantly higher concentrations in 2006 than in 1994. i
TABLE OF CONTENTS EXECUTIVE SUMMARY ...... i 1. INTRODUCTION...... 1 1.1. Report Organization ...... 1 2. BACKGROUND ...... 2 2.1. Site Background ...... 2 2.2. Previous Studies...... 4 2.2.1. 1994 SLRA Sampling of Entire Slough...... 4 2.2.2. 1995 Buffalo Slough Investigation ...... 5 2.2.3. 1997 Wapato Wetland Investigation ...... 5 2.2.4. 1998 Peninsula Drainage Canal ...... 5 2.2.5. 1998 Marx-Whitaker Slough Investigation ...... 6 2.3. Contaminants of Potential Concern...... 6 3. SAMPLE COLLECTION AND FIELD ACTIVITIES ...... 6 3.1. Sediment Collection Locations ...... 6 3.2. Sampling Procedures ...... 8 3.3. Sediment Collection Activities ...... 8 4. 2006 SAMPLE ANALYSES...... 9 4.1. Analytical Constituents ...... 9 4.2. Data Usability ...... 9 5. GENERAL SUMMARY OF THE 2006 SEDIMENT DATA ...... 10 5.1. Percent Detection for each analyte ...... 10 5.2. Percent Exceedence of Columbia Slough Screening Levels...... 13 6. EVALUATION OF SELECTED CONTAMINANTS...... 15 Metals...... 16 6.1. Copper...... 16 6.2. Lead...... 22 6.3. Zinc...... 26 PCBs...... 27 6.4. PCB 1248 ...... 27 6.5. PCB 1254 ...... 28 6.6. PCB 1260 ...... 29 General Observation on PCB Aroclors ...... 30 Pesticides...... 32 6.7. Alpha Chlordane...... 32 6.8. Gamma Chlordane ...... 33 6.9. Dieldrin ...... 34 6.10. Sum of DDTs ...... 37 6.11. Individual 4,4’-DDX Compounds ...... 38
ii
Semi-Volatiles ...... 43 6.12. Total PAHs ...... 43 6.13. Benzo(k)fluoranthene ...... 45 6.14. Bis(2-ethylhexyl)phthalate ...... 46 6.15. Summary Table of the 2006 Sediment Data ...... 48 Relationship between 2006 sediment data and 2005 fish tissue data ...... 50 Comparison of 2006 and 1994 data to special study areas ...... 53 Buffalo Slough...... 55 Wapato Slough...... 57 Marx-Whitaker Subbasin...... 60 7. CONCLUSIONS...... 63 8. REFERENCES...... 69
Appendices A – Map of sample locations B – Columbia Slough Screening Levels C – Data Usability Report D – 2006 Columbia Slough Sediment Data Tables
iii Columbia Slough Sediment Data Analysis
1. INTRODUCTION The Columbia Slough (hereafter “Slough”) comprises a 19-mile channel with about a dozen miles of side channels that parallels the Columbia River. Previous investigations of the Slough have identified pollutants that have impaired the quality of water, sediments, and fish tissue. The Oregon Department of Environmental Quality (DEQ) placed the Columbia Slough on the state’s 303(d) list in 1994 as water-quality limited for bacteria, phosphorus, dissolved oxygen, chlorophyll a, toxics (DDT/DDE, dieldrin, dioxins, PCBs, and lead), pH, and temperature. In response to the identification of contaminants in the Columbia Slough, the DEQ and the City of Portland (City) performed a Feasibility Study with the goal of defining a framework for addressing contaminated sediments over the long term. The remedial approach recommended based on this evaluation includes source control, sediment cleanup, and long-term monitoring and evaluation. The Long-term Monitoring Plan utilizes the collection of sediment and fish tissue samples to track the environmental quality of the Slough over time. The data collected will be used to assess the effectiveness of source control and cleanup efforts, evaluate continued need for institutional controls or other changes in management practices, and guide future efforts to improve environmental conditions. In addition, the Long-Term Monitoring Plan will be used to evaluate the potential for impacts on human and ecological receptors and to evaluate the need for the Slough fish/health advisory outreach and education effort. The second long-term fish tissue monitoring event was conducted in 2005 and is described in Columbia Slough Fish Tissue Analysis, 2005 Sampling (BES and GeoSyntec, 2007). The current sediment sampling program included all major reaches of the Columbia Slough including: Lower Slough, North Slough, Middle Slough, Upper Slough, Big Four Corners East Slough, Peninsula Drainage Canal, Buffalo Slough, and Whitaker Slough. Subsequent sediment sampling events are expected to occur on a 10-year sampling frequency. This frequency should provide sufficient time for the data to reflect natural recovery processes in sediment as source control and sediment cleanup actions are implemented. The long-term objectives of sampling sediment are three-fold: (1) assess progress in achieving protective levels; (2) evaluate spatial and temporal trends; and (3) identify areas where more focused remedial efforts may be warranted. To complete the first objective, sediment data were evaluated to determine which samples exceed Columbia Slough-specific sediment screening levels. Slough screening levels are compiled and/or developed by DEQ, updated as new information becomes available, and are used to determine if contaminants are present above levels of concern. Completing the second objective entails an assessment of key COIs from 1994 and 2006 to determine if there are spatial and temporal differences. Completing the third objective involved identifying geographic areas where COIs appeared to be consistently or significantly higher than other areas of the Slough.
1.1. Report Organization
The remainder of this Report is organized into the following sections: Section 2.0 – Background Section 3.0 – Sample Collection and Field Activities Section 4.0 – 2006 Sample Analyses Section 5.0 – General Summary of the 2006 Sediment Data
City of Portland BES - 1 - February 2008 Columbia Slough Sediment Data Analysis Section 6.0 – Evaluation of Selected Contaminants Section 7.0 - Conclusions Section 8.0 – References
2. BACKGROUND
2.1. Site Background The Columbia Slough watershed drains approximately 32,700 acres of land (Figure 2). Portland’s city limits end at approximately NE 185th Avenue on the east, but the watershed also includes Fairview Lake and Fairview Creek, and portions of Troutdale, Fairview, Gresham, Maywood Park, Wood Village, and Multnomah County within the watershed boundaries. The watershed once contained a vast system of side channels, lakes, and wetlands that comprised the floodplain of the Columbia River between the mouths of the Willamette and Sandy Rivers. High water seasonally inundated the floodplain, cutting new channels and depositing sediment. Native Americans used these waterways and the uplands for fishing, hunting, and gathering food. Over the years, the watershed and waterway have been drastically altered. The watershed once had numerous streams, ponds, and wetlands. Beginning in 1918, levees were built and wetlands were drained and filled to provide flood protection and allow for development. The waterway was channelized, and dozens of streams were diverted from natural channels to underground pipes. Today the Columbia Slough comprises a 19-mile main channel that parallels the Columbia River, as well as approximately a dozen additional miles of secondary waterways. Other remaining major surface water features include Fairview Creek, Fairview Lake, and Smith and Bybee Lakes. Floodplain development has resulted in an extensively managed surface water system that includes levees, pumps, and other water control structures in the Middle and Upper Sloughs. The levee system greatly reduced the Columbia River’s connection to its floodplain. The Slough is currently divided into three sections, based on hydraulic characteristics:
• The Upper Slough starts at the mouth of Fairview Lake on the east and flows west to the mid-dike levee at NE 142nd Avenue. It includes Big Four Corners East & West Sloughs. It receives water from Fairview Lake, Fairview Creek, Wilkes Creek, stormwater outfalls, natural springs, groundwater, and overland flow.
• The Middle Slough extends from the mid-dike levee, near NE 142nd Avenue to the Pen 2 levee, near NE 18th Avenue. It includes a substantial southern arm complex of sloughs and lakes, including Peninsula Drainage Canal, Prison Pond, Mays Lake, Johnson Lake, Whitaker Slough, Whitaker Ponds, and Buffalo Slough. The Middle Slough receives water from the Upper Slough, stormwater outfalls, natural springs, groundwater, and overland flow. Pumps are used to move water from the Upper and Middle Slough to the Columbia River or the Lower Slough.
City of Portland BES - 2 - February 2008 Columbia Slough Sediment Data Analysis
Figure 21 Columbia Slough Watershed Overview
City of Portland BES - 3 - Columbia Slough Sediment Data Analysis
• The Lower Slough starts at the Pen 2 levee, near NE 18th Avenue, and extends approximately 8.5 miles to the Willamette River. The lowlands of the Lower Slough Watershed are subject to flooding because they are not protected by levees. Water flow and levels in the Lower Slough are affected primarily by the Columbia River and Willamette River and the ocean tides, and to a lesser extent by pumping. During high tide, the Columbia and Willamette Rivers create a backwater effect that complicates flow patterns. The Columbia Slough Watershed now includes several types of land uses: residential neighborhoods, commercial and industrial development, agriculture, Portland International Airport, interstate highways, railroad corridors, and large open spaces. Much of Portland’s industrial and commercial land is located within the watershed. In addition to industrial development in the area north of Columbia Boulevard and the Rivergate area, land is preserved for industrial uses in the Columbia South Shore area between NE 82nd and NE 185th Avenues north of Sandy Boulevard. Over time, extensive alteration of the Slough’s watershed, due to industrial and residential development, has had a deleterious effect on the environmental quality of the watershed. As development occurs, the natural topography, hydrology, and vegetation are altered and impervious surfaces such as streets, parking lots, and buildings are placed over much of the land. As a result of urbanization, industrial releases, alteration of water flows, and runoff from agricultural land, the Columbia Slough has polluted water, sediments, and fish.
2.2. Previous Studies A reconnaissance study indicated elevated levels of heavy metals and some toxic organic compounds in Columbia Slough sediments. In 1993, a joint DEQ-Metro fish sampling effort found high levels of PCBs and pesticides (such as DDTs and dieldrin), in the fish tissues of five fish caught at one location near the St. Johns Landfill. This finding of toxic organic compounds in fish raised concerns about impacts to the people of immigrant communities who frequently fish from the Slough. In September 1993, the Oregon Department of Health Services issued a health advisory about eating carp from the Slough. Data from previous studies was used to evaluate spatial and temporal variations through comparison to data collected by BES in 2006. General scope and findings of each of these studies is provided below.
2.2.1. 1994 SLRA Sampling of Entire Slough Following on the 1993 fish sampling event and the issuance of the health advisory, the City, under Consent Order with DEQ, conducted more extensive fish tissue and sediment sampling in 1994, during the screening level risk assessment (SLRA) phase of the Columbia Slough Sediment Project. The SLRA, completed in 1995, involved a general “screening” of the Columbia Slough to identify areas that present the highest potential risks to wildlife and human health. The SLRA tested 300 sediment samples from sites along 31 miles of the Slough for over 140 chemicals. Fish tissue samples were also collected and tested for a variety of toxic chemicals. At each sampling site, the potential risks from sediment contaminants were assessed for different receptors: aquatic life (benthic, or bottom-dwelling organisms and fish), wildlife (blue heron and river otter), and humans. The results of these assessments were combined to
City of Portland BES - 4 - Columbia Slough Sediment Data Analysis identify the most contaminated sediment sites in the Columbia Slough. Each sediment sample was given a "hazard score," based on the overall risk it posed. Based on their hazard score, the 300 sediment samples were then divided into four priority groups: A, B, C, and D. Priority A sites had the highest hazard scores and posed the greatest potential risk relative to all the sites, while Priority D sites had the lowest hazards scores and posed the lowest potential risk. This ranking helped set priorities for action by identifying the areas of greatest concern. The Slough SLRA confirmed that banned toxic organics, such as PCBs and pesticides, are found in fish tissues at levels which may pose an unacceptable risk to human health, although the levels were not as high as those found in the earlier 1989 sampling effort. Many of the organic chemicals found in fish tissues have been banned for many years. No one type of fish showed a significantly higher level of contamination in all segments of the Slough. Carp generally had a slightly higher level of contamination, probably due to their higher lipid content. Contaminant concentrations were almost invariably higher in non-edible portions of fish, regardless of species. Heavy metals, particularly lead and chromium, were commonly found in sediments at levels that exceeded conservative screening levels based on impacts to benthic organisms and/or, via bioaccumulation, to wildlife that consume aquatic organisms.
2.2.2. 1995 Buffalo Slough Investigation Fifty sediment samples were collected from the Buffalo Slough in 1995 as an element of a focused Remedial Investigation (RI) (Parametrix 1996). The chemicals selected for analysis were based on the chemicals of potential concern identified by the screening level risk assessment. The chemicals analyzed included 14 semi-volatiles, 11 pesticides, 6 PCB Aroclors, and 8 metals or inorganics. The study found that people who consumed fish from Buffalo Slough had significant risks from PCBs and pesticides (e.g., DDTs, chlordane, dieldrin). Concentrations of several metals (e.g., lead, copper, zinc) also indicated potential toxicity to benthic organisms. However, sediment bioassays suggested negligible to low risks to benthic organisms, potentially due to low bioavailability.
2.2.3. 1997 Wapato Wetland Investigation A focused investigation of Wapato Wetlands in 1997 showed that metals contamination in this area was not widespread. Chromium, copper and manganese contamination extended from outfall 55A only approximately 6 meters into the wetland. Cobalt, molybdenum and nickel concentrations were not elevated in the vicinity of the outfall. Sediment bioassays showed that invertebrate survival and growth were not significantly affected by metals contamination in Wapato Wetland, even in the vicinity of the outfall, suggesting that the contaminants were not bioavailable to benthic invertebrates. DEQ determined that no sediment cleanup or removal is required in this portion of the Slough and provided a No Further Action (NFA) determination (August 31, 1998). BES also installed a vault-type of stormwater filtering facility to reduce contaminants in stormwater discharged to this section of the slough.
2.2.4. 1998 Peninsula Drainage Canal A survey of fish consumption was completed for Peninsula Drainage Canal and risks were calculated to fall within acceptable risk levels. DEQ provided a conditional NFA (December 17, 2002) for this reach, based on remaining contaminant concentrations (below baseline but exceeding conservative risk-based levels) that could pose a threat to benthic organisms. This segment of the Slough will continue to be monitored to ensure that natural recovery is effective in achieving further contaminant reduction.
City of Portland BES - 5 - Columbia Slough Sediment Data Analysis
2.2.5. 1998 Marx-Whitaker Slough Investigation Sediments in the Marx-Whitaker Slough were sampled to determine the nature and extent of contamination. The sediment investigation found that DDTs dieldrin, endosulfanII, endosulfan sulfate and petroleum hydrocarbons were elevated in sediment throughout the subbasin. The pesticide levels were above concern thresholds for the protection of benthic life, and their average concentration in subbasin sediment were ten-fold higher than the average levels measured in the Columbia Slough for the SLRA. Sediment toxicity tests indicated sediment was toxic to midge larvae growth and survival, and results were correlated with the concentrations of several pesticides (dieldrin, DDT, endosulfan II) and petroleum hydrocarbons. Risk analysis suggested that fish-eating birds that forage at the site may be adversely affected by DDE contamination.
2.3. Contaminants of Potential Concern Potential receptors of contaminants from sediment in the Columbia Slough include humans, via direct contact or fish ingestion; wildlife, via direct contact and ingestion of impacted benthic organisms or fish; and benthic organisms, via direct contact. Based on results of the 1995 SLRA, the only exposure route of concern for humans is ingestion of fish that have accumulated contaminants from the sediments, water column, or food source (BES 1995). Potential exposures of concern for ecological receptors include: ingestion of contaminated fish by birds or mammals; uptake by fish of contaminants in sediment, water column and food sources; and uptake of contaminants by sediment dwelling organisms. A wide range of contaminants including semi-volatiles, pesticides, metals, and PCBs were found to be present in Slough sediment at concentrations of concern for one or more of these pathways. To evaluate which contaminants in sediment present a potential unacceptable risk to human health or the environment, sediment data obtained during the SLRA was compared to Columbia Slough specific screening levels and baseline concentrations (BES 2006a). Screening levels are provided in a draft spreadsheet prepared by DEQ (DEQ 2006), provided in Appendix B. This table provides both risk-based levels and “baseline” concentrations. Baseline concentrations reflect the general level of contamination found throughout the Slough based on the 1994/95 data. Exceedances of baseline indicate there is likely a nearby source area for which some active cleanup is warranted. Ultimately, the goal for the Slough is to control sources and remediate elevated areas of sediment contamination such that the baseline concentrations in the Slough will decline to below the risk-based levels.
3. SAMPLE COLLECTION AND FIELD ACTIVITIES Field activities were carried out in general accordance with the Sampling and Analysis Plan for the Collection and Processing of Bulk Sediment Samples in the Columbia Sough (SAP) prepared by the City of Portland and dated *** 2006 (BES 2006***). Deviations from the SAP are described in this report where appropriate.
3.1. Sediment Collection Locations A systematic sample design was used to select the initial sampling locations, approximately every half mile, in the Columbia Slough and arms of the Slough. Additional locations were added to target additional outfalls or areas of concern. The statistical benefit of a randomly selected start point (which is typically done with systematic random sampling) was deemed negligible for long-term trend analysis.
City of Portland BES - 6 - Columbia Slough Sediment Data Analysis
Sampling locations were pre-selected for all major Slough reaches, totaling 78 sediment samples for the project. Some sampling locations were moved or additional ones added to target additional outfalls or areas of concern:
• Lower Slough: Seven sample points were added or moved in order to insure that there are sample locations near all thirteen former CSOs.
• St Johns Landfill: Two sample locations were moved and two were added to better mesh with sediment sampling that Metro conducted as part of the landfill remedial investigation. Avoided sampling near Union Carbide because substantial data is already available for this area as a result of the remedial investigation conducted by this facility .
• Wapato Wetlands: One sample point was added in this area of interest.
• Portland International Raceway (PIR): A sample point was added at the PIR pump station (this site is not located on the main channel of the Lower Slough).
• I-5: 2 sample points were added, and 2 others were moved to be closer to former CSO outfalls and to collect more data in this area of interest.
• Multnomah County Drainage District (MCDD) #1: the sample point at the east end of the lower slough was moved to downstream of the confluences of discharges from MCDD#1. The sampling point at the west end of the middle slough, at MCDD#1, was moved east to better represent this broad area.
• Peninsula Drainage Canal: One sample point was added in this area of interest.
• Middle Slough: A sample point was added to the main channel at the Port of Portland deicing outfall.
• Whitaker Slough: Four sample points were added in this area of interest.
• Upper Slough: One sample point was added to sample both sides of the mid-dike levee. One sample point was added west of NE148th because a water quality facility will be built near this location.
• Confluence Areas: sediment sample locations were either moved (no farther than ¼ mile) or added to sample downstream of confluences of the main slough and its side channels. A map of the sample collection locations are included in Appendix A. It should be noted that the movement of stations within the systematic sampling framework weakens the statistical strength of the sampling design. Various ways of reducing the impact of this issue on conclusions were used, as described in the Evaluation of Selected Contaminants section. For future sampling events, it is recommended that the stations within the statistically- based framework are maintained and not moved. The density of the stations within the statistically-based framework can be reduced if additional stations are needed to address areas of interest, but fidelity to a systematic, stratified-random, or other approach should be maintained.
City of Portland BES - 7 - Columbia Slough Sediment Data Analysis
3.2. Sampling Procedures All sample collection was performed between July 2006 and September 2006 as set forth in the SAP. The SAP provided collection method and sampling site selection guidelines to be implemented by Field Operations based on site-specific conditions. The majority of sediment samples were obtained using a large Ekman dredge to a depth of approximately 10 centimeters. On occasion, the smaller Ponar dredge was utilized for sites with gravelly substrate. Once a successful sample had been obtained it was then spooned either directly from the grab sampler or a decontaminated stainless steel bowl into sample containers. Prior to filling labeled sample jars, the sample depth range, number of sub-samples required, collection method, sediment description, percent volume and description of any debris removed from sample, etc., were recorded on the field data sheet (FDS). Samples were placed in a chilled cooler for transport to the Water Pollution Control Laboratory (WPCL). A small Ekman dredge was used to obtain sediment for field parameters. Redox potential (Eh) and pH were measured using the Orion Model 230A meter coupled with a silver/silver chloride probe. Field parameters were recorded on site-specific Field Data Sheets (FDS). Stainless steel sediment dredges, bowls and spoons were decontaminated using a five-step method of: soapy tap water, tap water, 100% acetone, deionized water, and ultra pure deionized water between each sampling event. Equipment was wrapped in aluminum foil for transport and dredges were put in clean heavy-gauge plastic bags.
3.3. Sediment Collection Activities Sites were pre-determined based on maps and geo-spatial analysis. Sampling locations were approximately 0.5 miles apart, with the exception of targeted priority sites. BES successfully sampled all 78 sites. While all resources were fully utilized to precisely locate and sample at the pre-determined sampling locations, on occasion there was some slight deviation from the prescribed sampling locations, as allowed by the SAP. This was due mainly to terrain limitations and local site conditions. The majority of sites were within 50 feet of the prescribed locations. Maximum variance is estimated to be 300 feet from pre-determined sampling points. Pre-determined site locations were navigated to using a Garmin global positioning system (GPS) unit coupled with aerial photos and then accessed via jet boat, a small rowboat, wading or walking, depending on terrain. Reconnaissance of, and arranging access to, sediment sampling sites was a critical task of this project. The Port of Portland, City of Portland Water Bureau, Multnomah County Drainage District and several industries along the Slough granted the City permission to access the Slough via their property. Upon arrival at the designated location, a new waypoint was created to reflect the actual sampling location. The Multnomah County Drainage District was actively dredging a section of the Middle Slough from approximately NE 55th and the Copart Bridge (roughly NE 68th) during the sampling period of July through October 2006. After discussion with project managers, a decision was made to proceed with sampling in the dredged areas with the implicit understanding that the samples might have been significantly disturbed or were, in fact, side-castings. It is not known to what degree the reach had been disturbed but it is important to note that Middle Slough Reach sites MS06 and MS12 may have been affected by dredging. At the time of sampling, Wapato Slough had likely been dry for several weeks. As there was no overlying water, samples were taken directly from the ground with either a hand trowel or stainless steel shovel.
City of Portland BES - 8 - Columbia Slough Sediment Data Analysis
4. 2006 SAMPLE ANALYSES
4.1. Analytical Constituents Previous sediment studies in the Columbia Slough identified certain PCBs, pesticides, PAHs, and metals as constituents of concern. In order to assess the low concentrations for certain organic analytes, some sediment was sent to Battelle for selected analyses, as they have the capability of performing low-level analysis on sediment samples. The complete list of constituents measured for this project and associated methods are provided below (Table 1).
Table 1: Sediment Analyses
Constituents Method PCB Aroclors General NS&T (GC/ECD) Organochlorine pesticides General NS&T (GC/ECD) PAHs + phthalates + phenol General NS&T (GC/MS) Total Metals EPA 6020 TPH (gas, diesel, oil) NWTPH-HCID, -Gx, -Dx
4.2. Data Usability BES Investigation & Monitoring Services (IMS) conducted an independent data validation to ensure the data were usable. All data were evaluated using the project Quality Assurance Project Plan (QAPP) (BES 2006b) and U.S. EPA Contract Laboratory Program National Functional Guidelines (NFGs) for Data Review (USEPA 1999, 2004, 2005) for guidance in evaluating the following:
• Sample chain of custody (COC) and receipt documentation, preparation and analytical holding times, and reporting and detection limits for chemicals of interest
• Laboratory data quality, in terms of precision, accuracy, representativeness, completeness, and comparability (PARCC) Data were reviewed by IMS and qualifiers applied, where necessary. Qualified data are still considered valid and usable, except for those analytes which may have been rejected (R qualified). No data were rejected subsequent to this review. Analytes may also have been qualified with any of the following additional qualifiers: “J” (estimated concentration), “U” (analyte non-detected), “B” (blank contamination), or “M” (potential false positive). The data usability report is included in Appendix C. Some data were qualified as estimated where for other end uses requiring higher quality data, the data might have been rejected. Specifically, recoveries and field duplicate results indicate that some organochlorine pesticide detections may be false positives, which is not uncommon with EPA Method 8081 analysis. These data were qualified with “M” or “UM” as appropriate. Also, Battelle experienced chronic phenol laboratory contamination which resulted in little of the data being usable as reported. Some of the phenol data are valid; however, spatial analysis is not possible with the 2006 phenol sediment dataset. Since data collected for this project are City of Portland BES - 9 - Columbia Slough Sediment Data Analysis used primarily for management decisions, some data were qualified without being rejected in order to avoid rejecting valid data.
5. GENERAL SUMMARY OF THE 2006 SEDIMENT DATA The 2006 sediment sampling effort analyzed concentrations of 70 contaminants and measured 15 ancillary parameters (e.g., grain size, pH) at 78 stations across the Columbia Slough. The results for each sampling location and parameter are included as a table in Appendix A. The table also lists the Columbia Slough Screening Level (DEQ 2007) where available, and color codes in red those results that exceed the screening level.
5.1. Percent Detection for each analyte Table 2 below summarizes the percent detection for each contaminant, indicates the number of non-detected, estimated and unqualified results for each contaminant, and identifies the range and average for each. A few general patterns are evident in evaluating the percent detections. Because they are naturally occurring, metals were nearly always detected, ranging from 96- 100% detection. Polycyclic Aromatic Hydrocarbons (PAHs) were also nearly always detected, and ranged from 97-100% detection. In contrast, many PCB Aroclors were never detected in any sample, or were detected infrequently. Six of the nine Aroclors were never detected (Aroclors 1016, 1221, 1232, 1242, 1262 and 1268). Aroclors 1248 (12%) and 1260 (5%) were detected infrequently. Aroclor 1254 was the most frequently detected Aroclor, and was detected in 28% of the samples. Pesticides ranged widely in percent detection, from 0-99%. Several DDT compounds, chlordanes, aldrin, dieldrin and endosulfan II were some of the more frequently detected contaminants, and all were detected in over half of samples. For the DDx compounds, the 4-4’- DDx isomer was detected more frequently than the 2,4’-DDx isomer. Pesticides that were never detected include alpha-, beta- and delta-BHC, endrin ketone and heptachlor epoxide. In general, the semi-volatile organic compounds had the highest number of results qualified as estimates1. Diethylphthalate (50 estimates out of 78 samples analyzed), di-n-butylphthalate (42), and phenol (32) had the highest proportion of qualified results, while among the pesticides gamma chlordane (26) had the highest number of results qualified as estimates. In contrast, none of the PCB Aroclors were qualified as estimates.
1 Estimate qualifiers indicate that the analyte was detected, but the reported value is an estimate due to quality assurance/quality control issues. Estimates were indicated on graphs, and used at the reported values in calculations of medians and significance tests.
City of Portland BES - 10 - Columbia Slough Sediment Data Analysis
Table 2: Percent detection, percent estimates and range of the 2006 sediment data %Estimates of samples Range Analyte % Detect analyzed (ppm) Antimony 96% 3% 0.10 - 4 Arsenic 99% 1% 1.00 - 12 Cadmium 99% 1% 0.10 - 2 Chromium 100% 1% 15 - 320 Copper 100% 0% 10 - 297 METALS Lead 100% 0% 6 - 177 Nickel 100% 1% 12 - 841 Zinc 100% 0% 46 - 576
Range Analyte % Detect %Estimates (ppb) Aroclor 1016 0% 0% 0.47 - 0.47 Aroclor 1221 0% 0% 0.27 - 4 Aroclor 1232 0% 0% 0.60 - 4 Aroclor 1242 0% 0% 0.51 - 0.51 Aroclor 1248 12% 0% 0.15 - 125
PCBs Aroclor 1254 28% 0% 0.16 - 330 Aroclor 1260 5% 0% 0.15 - 91 Aroclor 1262 0% 0% 0.45 - 0.45 Aroclor 1268 0% 0% 0.27 - 0.27
City of Portland BES - 11 - Columbia Slough Sediment Data Analysis
Table 2: (cont.) Percent detection, percent estimates and range of the 2006 sediment data Range Analyte % Detect %Estimates (ppb) 2,4'-DDD 85% 1% 0.03 - 34 2,4'-DDE 31% 3% 0.04 - 3 2,4'-DDT 12% 3% 0.05 - 28 4,4'-DDD 99% 14% 0.02 - 95 4,4'-DDE 99% 3% 0.02 - 166 4,4'-DDT 59% 17% 0.02 - 77 Alpha-BHC 0% 0% 0.04 - 0.04 Alpha-Chlordane 86% 6% 0.02 - 9 Aldrin 62% 24% 0.02 - 4 Beta-BHC 0% 0% 0.02 - 0.02 Delta-BHC 0% 0% 0.02 - 0.02 Dieldrin 69% 0% 0.02 - 67 Endosulfan I 3% 1% 0.02 - 4
PESTICIDES Endosulfan II 54% 14% 0.02 - 33 Endosulfan Sulfate 47% 4% 0.03 - 104 Endrin 15% 1% 0.02 - 1 Endrin Aldehyde 4% 1% 0.03 - 1 Endrin Ketone 0% 0% 0.02 - 0.02 Gamma-BHC (Lindane) 1% 1% 0.02 - 3 Gamma-Chlordane 87% 33% 0.02 - 19 Heptachlor 1% 0% 0.03 - 7 Heptachlor Epoxide 0% 0% 0.04 - 0.04 Methoxychlor 12% 12% 0.13 - 3
City of Portland BES - 12 - Columbia Slough Sediment Data Analysis
Table 2: (cont.) Percent detection, percent estimates and range of the 2006 sediment data Range
Analyte % Detect %Estimates (ppb) 2-Chloronaphthalene 3% 0% 0.13 - 2 Acenaphthene 100% 18% 0.13 - 336 Acenaphthylene 97% 9% 0.17 - 85 Anthracene 100% 3% 0.74 - 588 Benzo(a)anthracene 100% 1% 0.72 - 978 Benzo(a)pyrene 100% 3% 1 - 1,061 Benzo(b)fluoranthene 100% 3% 1 - 1,061 Benzo(g,h,i)perylene 100% 3% 2 - 889 Benzo(k)fluoranthene 100% 4% 1 - 1,140 Bis (2- ethylhexyl)phthalate 100% 13% 19 - 7,679 Butylbenzylphthalate 97% 10% 0.11 - 868 Chrysene 100% 1% 2 - 1,345 Dibenzo(a,h)anthracene 100% 5% 0.31 - 231 Dibenzofuran 100% 14% 0.06 - 197
SEMI-VOLATILES Diethylphthalate 71% 64% 1 - 476 Dimethylphthalate 88% 15% 0.12 - 185 Di-n-butylphthalate 94% 54% 5 - 2,118 Di-n-octylphthalate 62% 1% 0.13 - 3,371 Fluoranthene 100% 1% 1 - 3,259 Fluorene 100% 8% 0.11 - 302 Indeno(1,2,3-cd)pyrene 100% 3% 1 - 739 Naphthalene 99% 17% 0.49 - 369 Phenanthrene 100% 10% 0.76 - 3,045 Phenol 54% 41% 7 - 1,318 Pyrene 100% 1% 1 - 3,287
5.2. Percent Exceedence of Columbia Slough Screening Levels There are no regulatory standards for sediment contaminant concentrations in Oregon. However, DEQ has developed Columbia Slough Screening Levels that can be used as a first step in identifying potential problem areas in need of upland source control or sediment cleanup (DEQ 2007). The Columbia Slough Screening Levels are listed with each analyte in the table in Appendix A. The report uses the “Sediment Screening Level” from the DEQ tables. These levels incorporate “baseline” calculations, which try to account for background concentrations City of Portland BES - 13 - Columbia Slough Sediment Data Analysis across the Slough. Baseline concentrations reflect the general level of contamination found throughout the Slough based on the 1994/95 data. Exceedances of baseline indicate there is likely a nearby source area for which some active source control or cleanup is warranted. The City will be working with DEQ to recalculate these values using the 2006 data and using recent Bayesian advances in statistical approaches for estimating background concentrations (Qian and Lyons 2006). Over time as progress is made in reducing contaminant concentrations, sediment data will be compared to the “Source Control Screening Level”, which is generally based on human health and ecological risk. The percentages of samples exceeding the screening levels for each analyte are listed in Table 3. For the metals, chromium, copper and zinc showed the highest exceedence of the screening levels, with approximately 20% of the samples exceeding for each. The PAHs were typically measured at levels above the screening levels. Benzo(k)fluoranthene, (62%), fluoranthene (60%), benzo(b)fluoranthene (58%) and pyrene (56%) were most frequently detected above the screening level. Consistent with the PAH results, #6 fuel oil (87%) and motor oil (99%) had very high exceedence percentages. Since PCBs were infrequently detected they did not have a high percentage of samples above screening levels, but for Aroclors 1248 and 1260 all detected values were above the screening level (i.e., % detection was equal to % exceed; detection limits were two orders of magnitude below screening levels). A little less than half of the detected values for Aroclor 1254 exceeded the screening level. For the pesticides, gamma chlordane (63%), 4,4’-DDE (62%), alpha chlordane (58%) and dieldrin (51%) were the analytes most frequently detected above their screening levels. As with percent detection, the 4,4’-DDx isomer more frequently exceeded screening levels than the 2,4’- DDx isomer.
City of Portland BES - 14 - Columbia Slough Sediment Data Analysis
Table 3: Percent Exceedence of the Columbia Slough Screening Level Type Analyte % Exceed Type Analyte % Exceed Antimony 1% #6 Fuel Oil 87% Dx Arsenic 14% - Diesel 20%
Cadmium 6% TPH Kerosene 41%
Chromium* 21% NW Motor Oil 99% Copper* 22% Aroclor 1248 12% Metals Lead* 10% Aroclor 1254 12%
Nickel 8% PCBs Aroclor 1260 5% Zinc* 21% 2,4'-DDD 15% Acenaphthene 1% 2,4'-DDE 0% Acenaphthylene 0% 2,4'-DDT 5% Anthracene 15% 4,4'-DDD* 45% Benzo(a)anthracene 46% 4,4'-DDE* 62% Benzo(a)pyrene 50% 4,4'-DDT* 31% Benzo(b)fluoranthene 58% Alpha-BHC 0% Benzo(g,h,i)perylene 10% Alpha Chlordane* 58% Benzo(k)fluoranthene* 62% Aldrin 24%
iles Bis(2-ethylhexyl) phthalate* 32% Beta-BHC 0% t Chrysene 50% Pesticides Dieldrin 51% Dibenzo(a,h)anthracene 14% Endosulfan I 0%
Semi-Vola Dibenzofuran 0% Endosulfan II 0% Fluoranthene 60% Endrin Aldehyde 0% Fluorene 6% Gamma-BHC (Lindane) 0% Indeno(1,2,3-cd)pyrene 53% Gamma Chlordane* 63% Naphthalene 1% Heptachlor 1% Phenanthrene 49%
Phenol 65% *-analytes selected for further evaluation Pyrene 56%
6. EVALUATION OF SELECTED CONTAMINANTS In order to provide more detailed analysis of the sampling results, a number of the analytes were selected for further evaluation. These analytes are indicated by asterisk in Table 3. In general, these analytes were selected for further evaluation because they frequently exceeded screening levels, or had been identified as key risk drivers in previous analyses such as the Screening Level Risk Assessment (BES 1995). Additional analytes may be evaluated based on specific needs identified in particular Slough segments. In addition, some analytes were also City of Portland BES - 15 - Columbia Slough Sediment Data Analysis evaluated as a group since they have similar chemical properties and spatial patterns. The DDT compounds, chlordanes and PAHs were summed and evaluated together as total DDTs, total chlordanes and total PAHs.
Metals
6.1. Copper
2006 Patterns Figure 3: 2006 Copper by Section The highest value for copper, 297 ppm, was detected at Copper vs Section station WAS03, at the east end 54* of Wapato Slough. This value LOWER SLOUGH was three times higher than all NORTH SLOUGH of the other values detected WAPATO SLOUGH throughout the Slough (Figure PENINSULA DRAINAGE
3). Copper levels in Wapato n o i
t MIDDLE SLOUGH Slough decreased moving west c Se from this location. WAS02 had BUFFALO SLOUGH considerably lower copper WHITAKER SLOUGH levels (90.7 ppm) but still exceeded the screening level UPPER SLOUGH of 54 ppm, while WAS01 at the BIG FOUR CORNERS west end of the Slough was 0 50 100 150 200 250 300 just below the screening level *-C olumbia Slough Screening Lev el ppm Copper with a value of 52.7 ppm. Several other metals showed a similar pattern in Wapato Slough (see below). The Lower Slough also had high copper values in comparison to the rest of the Slough. The Lower Slough had two of the three highest copper values, the second highest median value after Wapato Slough (Table 4) and the greatest number of stations above the copper screening level, with nine of the 23 values above 54 ppm. In the Lower Slough most of the high copper values are clustered around the I-5/MLK transportation corridor (Figure 4). This area receives runoff from outfalls 59-65 and a suite of private outfalls indicated in the map. Two stations near I-205 also exceed the screening level, and three stations in close proximity in Middle and Whitaker Sloughs also were measured above screening levels. All the values measured in North Slough, Buffalo Slough, and Upper Slough were below the screening level. Comparison of copper levels indicated that the medians did not differ significantly2 amongst the different Slough sections, though the p value was marginally non-significant.
2 Kruskal-Wallis nonparametric test was used due to unequal variances and non-normality; p=0.06
City of Portland BES - 16 - Columbia Slough Sediment Data Analysis
Figure 4: Copper levels across the Columbia Slough. Circles in red indicate values that exceed the Columbia Slough Screening Level. Aerials in insets show stations in relation to City outfalls (labeled green dots) and private outfalls (yellow triangles).
City of Portland BES - 17 - Columbia Slough Sediment Data Analysis
Figure 5: Copper by year and Slough section. The pink diamonds Comparison with 19943 indicate medians for each section and year. Comparison of 2006 copper levels to 1994 copper levels indicate that Copper by Year 2006 copper levels are higher than Section Year 54 1994 Year 4 LOWER SLOUGH 1994 levels . The largest increase in 2006 1994 1994 2006 NORTH SLOUGH the median and mean between the 2006 1994 two years appeared in Lower Slough WAPATO SLOUGH 2006 and Wapato Slough (Figure 5). 1994 PENINSULA DRAINAGE However, the 1994 sampling event 2006 1994 MIDDLE SLOUGH did not sample as far upstream into 2006 1994 BUFFALO SLOUGH Wapato Slough, with the furthest 2006 upstream random station 1994 WHITAKER SLOUGH 2006 corresponding roughly to WAS02 in 1994 UPPER SLOUGH 2006. Nearly all the median values 2006 1994 BIG FOUR CORNERS by section in 2006 were higher than 2006 in 1994, with the exception of Upper 0 50 100 150 200 250 300 and Whitaker. ppm Copper
3 There are some important caveats that accompany the comparison of 2006 and 1994 data throughout this report. The first is that two points in time do not establish a trend – a series of repeated samplings over time is needed to determine whether sediment contaminants are “increasing” or “decreasing”. The second is that the two sampling events used different station placement approaches – in 1994 a stratified random approach was used to provide a probabilistically-based estimate of sediment concentrations, combined with targeted stations which were placed to assess specific point sources. The random samples from 1994 were used to obtain an unbiased estimate of contaminant concentrations. In 2006 a systematic approach to station placement was used. However, stations were also added and in some cases moved to address points of interest, which has the potential to introduce bias. The analyses were conducted both using and excluding all the added or moved stations, and the general results and conclusions were similar for both analyses. The graphs and statistics presented are based on using all the 2006 data so that there are not gaps in spatial coverage, but suggestions are provided in the recommendation section for avoiding these issues in the future. Third, the sampling depth used in this sampling event – 10 cm – is different than the sampling depth used in the 1994 sampling (2 cm). Both are commonly used sampling depths and there are good reasons for using either (10 cm is the “bio-active zone”, and 2 cm represents the most recently deposited sediments). However, this has the potential to affect comparisons of the 2006 results to the 1994 results, as the deeper sampling depth used in 2006 has the potential to sample down into or below the sediment layer sampled in 1994, depending on sedimentation rates. The sedimentation rate in the Slough is unknown, but given that the Slough is a hyper-eutrophic system with high levels of TSS, it is possible that over the 12 years between sampling events 10 cm of new sediment would be deposited, in which case the deeper sampling depth in 2006 would not confound the temporal comparison. In the recommendations section it is recommended that if temporal trends are a key objective of future sampling events, that the shallower 2cm sampling depth be used. Lastly, some portions of the Slough have been dredged over the time period, and this would clearly affect the temporal comparison if the dredging were recent enough that 10 cm of new sediment have not been deposited since the dredging occurred. However, the mainstem of the Middle and Upper Slough between NE 22nd Ave and NE 158th Ave are the only areas that have been dredged, so this would not affect the differences in the two sampling periods observed in the Lower, Wapato, Buffalo, or Whitaker sections of the Slough, where some of the largest changes were observed. 4 Ln transformed data. Two-sample T-Test. T-Value = -3.41 P-Value = 0.001
City of Portland BES - 18 - Columbia Slough Sediment Data Analysis
Table 4: Means, medians and maximums within each Slough section for key contaminants.
Lower Wapato Peninsula Buffalo Middle Whitaker Upper Big Four North Slough Slough Slough Drainage Slough Slough Slough Slough Corners 1994 2006 1994 2006 1994 2006 1994 2006 1994 2006 1994 2006 1994 2006 1994 2006 1994 2006 Mean 33.32 46.47 30.68 35.5 49.33 147.7 23.63 36.9 113.9 137.7 39.07 44.17 51.72 49.56 35.47 37.63 27.34 36.87 Chromium Median 30.45 48.2 29.5 37.7 48 77 26 35.7 63.25 97.6 41 43.6 42 39.1 36.5 37.5 25.5 37.75 Max 118 80.5 36.3 43.4 59 320 30 39.8 287 253 60 71.4 228 130 43 58.9 38 45.8
Mean 33.84 46.89 28.18 29.23 55.37 146.8 25.7 38.83 17.88 21.9 32.97 43.34 37.88 43.74 35.69 33.66 33.05 41.7 Copper Median 27.1 47.6 27 32.4 51.4 90.7 30.2 40.4 16.5 22 34.5 43.5 37.7 40.1 36.65 36.7 31.85 38.25 Max 109 99.9 35.8 34.7 67 297 36.1 40.9 25.5 23.7 53.1 85.8 60.9 81.9 44.8 42.5 58 60.3
Mean 44.42 58.88 19.13 19.7 102.7 73.03 28.08 40.33 36.75 30.37 42.39 34.15 146.8 73.12 48.17 30.56 41.58 23.57 Lead Median 28.5 47.6 21 21.7 98 77.6 32 38 36.5 26.2 42 34.9 86 65 51 31.5 30.5 22.95 Max 225 177 26.3 25.5 127 81.4 39 47.5 48 40 66 63.3 942 150 97 51.7 120 33.1
Mean 177.6 269.1 109.1 120.3 304.7 336.7 172.8 251.3 103.8 90.63 213.5 264.3 210.8 246.2 168.6 198.6 144.5 192.7 Zinc Median 128 273 114.8 125 289 315 199.5 232 103.2 88.1 212 283 222 219.5 199.5 228 111 188 Max 548 576 136 146 375 390 223 293 140 121 336 394 351 562 222 245 284 314
Mean 2.803 4.114 1.055 1.753 14.67 15.92 0.575 1.093 2.905 0.79 6.331 10.77 9.473 19.21 5.248 7.351 2.749 19.77 4,4'-DDD Median 1.95 2.83 1.035 1.32 15 18.91 0.585 0.61 1.5 0.78 5.5 9.86 3.15 9.115 4.15 6.435 1.95 11.63 Max 12 31.69 1.6 2.63 23 25.39 1 2.65 7.9 1.05 18 19.4 63 94.58 16 13.59 7.3 61.81
Mean 3.166 8.291 1.393 2.267 10 20.05 3.255 6.15 1.36 1.197 6.454 16.68 11.95 31.22 4.675 10.75 3.629 25.09 4,4'-DDE Median 1.85 7.26 1.7 2.35 12 23.4 2.75 5.53 1.375 1.23 5.8 14.31 3.2 13.84 5 11.68 3.5 15.52 Max 17 22.29 2 2.74 15 32.07 7.4 7.73 2.1 2.34 20 31.32 90 165.7 9 20.95 6.3 61.34
Mean 1.384 0.372 1.383 0.173 3.167 1.617 1.23 5.003 1.95 1.617 2.842 5.575 5.132 17.18 2.969 1.553 2.179 7.345 4,4'-DDT Median 1.3 0.02 1.55 0.02 2.2 1.87 0.585 6.27 1.85 0.02 2.1 2.7 2.55 2.695 2.45 1.285 2.4 1.205 Max 4.2 5.19 1.9 0.48 5.3 2.96 3.5 8.72 2.7 4.81 9.4 19.2 23 76.62 13 4.79 3 36.82
Mean 1.09 1.065 0.375 0.197 1.183 1.317 0.178 0.02 19.42 0.79 0.783 1.245 1.174 3.811 0.588 1.275 0.504 1.692 Alpha- Median 0.355 0.66 0.45 0.16 1.6 1.91 0.145 0.02 1.65 0.84 0.74 1.25 0.9 3.805 0.59 1.6 0.525 1.35 Chlordane Max 19 3.78 0.53 0.41 1.6 2.02 0.33 0.02 74 1.51 1.7 2.47 4.3 9.09 1.6 1.98 0.82 3.51
Mean 2.358 1.184 0.345 0.25 2.673 3.923 0.718 0.29 33.11 1.05 1.337 1.911 1.546 5.154 0.928 4.23 0.553 1.74 Gamma- Median 0.68 0.55 0.41 0.26 2.8 4.27 0.48 0.42 1.055 0.81 1.3 1.96 1.35 4.95 0.945 2.165 0.605 1.605 Chlordane Max 48 5.01 0.5 0.47 4.3 4.54 1.8 0.43 130 2.32 3.8 5.04 5.5 11.06 2.2 18.56 0.9 4.01
Mean 16.94 28.44 18.7 0.15 26 41.62 10.08 0.15 102.3 0.15 28.38 0.15 37.48 0.15 21.71 0.15 23.7 0.15 PCB 1248* Median 14.50.15210.15310.157.750.15540.15260.1535.50.15200.15240.15 Max 44 102.1 29 0.15 34 124.6 21 0.15 280 0.15 70 0.15 70 0.15 39 0.15 46 0.15
Mean 45.61 31.83 57 1.4 78 30 28.25 0.16 108.3 0.16 86.41 2.119 113.2 3.384 65.92 10.68 72.75 4.527 PCB 1254* Median 39 0.16 64.5 0.16 94 39.67 22.5 0.16 115 0.16 80 0.16 110 0.16 61 13.89 72.5 0.16 Max 120 329.5 88 3.88 100 50.17 56 0.16 140 0.16 210 15.29 210 45.29 120 22.41 140 16.29
Mean 37.53 4.108 46.33 0.15 63.67 0.15 22.85 0.15 75.5 0.15 70.01 0.15 91.67 11.56 53.08 0.15 57.38 0.15 PCB 1260* Median 32 0.15 52 0.15 76 0.15 18.5 0.15 75.5 0.15 65 0.15 87 0.15 49.5 0.15 59 0.15 Max 100 91.18 72 0.15 83 0.15 45 0.15 100 0.15 170 0.15 170 71.53 96 0.15 110 0.15
Mean 580.9 2409 437.5 513.8 856.2 2259 337.3 581.5 276.1 567.7 883.7 2146 1860 2851 645.4 3276 226.9 682.9 Total PAHs Median 256.5 1962 414.2 577.5 756.2 2168 403 528.6 286.9 401.5 709.1 1630 724.4 1387 516.8 1358 199.9 669.6 Max 3082 10550 787.8 700.7 1102 3138 449.3 808.5 399 931.4 2722 5967 12854 13287 3290 17753 424.9 1267
Bis (2- Mean 574 1167 140.8 124 404.8 174.9 43 94.74 1010 173.7 804.4 2293 857.7 2930 835.4 1590 143.4 543.6 ethylhexyl) Median 290 525.6 82 125.7 165 149.7 46 76.34 190 145.2 830 2149 455 2534 795 1484 53.5 436.2 phthalate Max 6500 4996 460 214.4 1200 288.2 60 144.2 4400 285 2500 5786 3600 7679 2100 3917 470 1233
*-Large decreases in the means and medians of PCBs are due to markedly lower detection limits in 2006.
City of Portland BES - 19 - Columbia Slough Sediment Data Analysis
Chromium
2006 Patterns As with copper, chromium was highest at the east end of Wapato Slough and decreased moving west. Buffalo Slough also had elevated chromium levels, and all three stations within Buffalo Slough exceeded the screening level value of 58 ppm (Figure 6). The highest value in Buffalo Slough (253 ppm) occurred at the east end of Buffalo Slough. This location also had a high chromium value in 1994 (287 ppm), and resulted in source investigation of the businesses and outfalls draining to this section. In the process of inspection, BES did not find any obvious sources of chromium, but did recommend a number of BMPs to businesses in the area to reduce the possibility of pollutant runoff into the Slough. The current findings have been forwarded to the Stormwater Section of the Source Control Program, which will use updated findings and source investigation approaches to identify possible chromium sources into the subbasin. The highest median values were at Buffalo (97.6), Wapato (77.0), and Lower Slough (48.2). Whitaker Slough had the third highest chromium value (130 ppm), but had a lower median and percent of exceedence than the Lower Slough. The values in Whitaker Slough that exceeded the screening level were in close proximity to the east end of Buffalo Slough (Figure 7) which may suggest common sources in that area, though this would need to be verified by field source investigations. All values in North Slough, Peninsula Drainage Canal and Big Four Corners were below the screening level. Statistical analysis indicated that the medians were not statistically different amongst the different Slough sections, though as with copper the p value was marginally non-significant5. Figure 6: Chromium by Slough section
Chromium vs Section 58 LOWER SLOUGH
NORTH SLOUGH
WAPATO SLOUGH
PENINSULA DRAINAGE n o i
t MIDDLE SLOUGH c
Se BUFFALO SLOUGH
WHITAKER SLOUGH
UPPER SLOUGH
BIG FOUR CORNERS
0 50 100 150 200 250 300 350 ppm Chromium
5 Kruskal-Wallis nonparametric test was used due to unequal variances and non-normality; p=0.06
City of Portland BES - 20 - Columbia Slough Sediment Data Analysis
Figure 7: Chromium levels across the Columbia Slough. Circles in red indicate values that exceed the Columbia Slough Screening Level. Aerials in insets show stations in relation to City outfalls (labeled green dots) and private outfalls (yellow triangles).
City of Portland BES - 21 - Columbia Slough Sediment Data Analysis
Figure 8: Chromium by year and Slough section. The pink Comparison with 1994 diamonds indicate medians for each section and year. Comparison of 2006 sediment data to 1994 sediment data Chromium by Year indicated that chromium levels SECTION Year 58 1994 Year LOWER SLOUGH in 2006 are significantly higher 2006 1994 6 1994 2006 NORTH SLOUGH than in 1994 . The largest 2006 1994 increase in the mean and WAPATO SLOUGH 2006 median occurred in the Lower 1994 PENINSULA DRAINAGE Slough and in Wapato Slough 2006 1994 MIDDLE SLOUGH (Figure 8), although in the latter 2006 1994 BUFFALO SLOUGH case this was driven by the 2006 1994 sample located at the east end WHITAKER SLOUGH 2006 of Wapato Slough, which was 1994 UPPER SLOUGH not captured by the random 2006 1994 BIG FOUR CORNERS samples in 1994. All medians 2006 except Whitaker Slough were 0 50 100 150 200 250 300 350 ppm Chromium higher in 2006 than in 1994, though Middle and Upper Slough were essentially even in the two years.
6.2. Lead Figure 9: Lead by Slough Section.
2006 Patterns Lead vs Section The Lower Slough and Whitaker 90* Slough were the only sections LOWER SLOUGH which exceeded the screening NORTH SLOUGH level for lead (Figure 9). WAPATO SLOUGH
However, the highest mean and PENINSULA DRAINAGE median lead values were at MIDDLE SLOUGH Wapato and Whitaker Slough. A Kruskal-Wallis test indicated that BUFFALO SLOUGH the differences amongst the WHITAKER SLOUGH
Slough sections were highly UPPER SLOUGH significant7. BIG FOUR CORNERS
0 50 100 150 200 *-Columbia Slough Screening Level ppm Lead
6 Ln-transformed data had equal variances, but no transformation produced normality. Year differences were significant with both parametric (T-Value = -3.29 P-Value = 0.001) and nonparametric(W = 9662.5, P-Value = 0.0004) tests. 7 Data were non-normal with unequal variances, and transformations did not produce normality or equal variances. Kruskal-Wallis: H = 26.46 DF = 8 P = 0.001
City of Portland BES - 22 - Columbia Slough Sediment Data Analysis
The lead values that exceeded the screening level were largely located around the I-5/MLK transportation corridor, with the exception of one exceedence in the vicinity of the St. Johns Landfill and outfalls 53E and 54 (LS06; Figure 11), which had the highest lead value detected in 2006 (177 ppm). The high values above the screening level in Whitaker Slough were near I-205 and just upstream and downstream of Prison Pond. Lead was one of the few metals that did not exceed the screening level at Wapato Slough.
Comparison with 1994 Comparisons of the 2006 lead data to the 1994 lead data indicate that there are no significant differences between the two years8. Several 2006 medians and means are equal to or slightly lower than the 1994 medians. In addition, the maxima in many sections decreased considerably, particularly in Whitaker Slough, where some of the highest values in 1994 were no longer evident in 2006 (Figure 10). The highest value in Whitaker Slough in 1994 was in the vicinity of NuWay Oil, which has since undergone cleanup.
Figure 10: Lead by year and Slough section.
Lead by Year SECTION Year 90 1994 Year LOWER SLOUGH 2006 1994 1994 2006 NORTH SLOUGH 2006 1994 WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006 0 100 200 300 400 500 600 700 800 900 ppm Lead
8 Values were ln-transformed and met assumptions for normality and equal variances. T-Value = 0.20 P-Value = 0.844.
City of Portland BES - 23 - Columbia Slough Sediment Data Analysis
Figure 11: Lead levels across the Columbia Slough.
City of Portland BES - 24 - Columbia Slough Sediment Data Analysis
Figure 12: Zinc levels across the Slough.
City of Portland BES - 25 - Columbia Slough Sediment Data Analysis
6.3. Zinc Zinc vs Section 314* 2006 Pattern LOWER SLOUGH Two of the three highest zinc NORTH SLOUGH values in the Slough were detected at LS20 (576 ppm) WAPATO SLOUGH and LS17 (540 ppm), which PENINSULA DRAINAGE were in close proximity in the MIDDLE SLOUGH I-5/MLK transportation BUFFALO SLOUGH corridor (Figure 12). Whitaker Slough only had one value WHITAKER SLOUGH above the screening level, but UPPER SLOUGH it was the second highest BIG FOUR CORNERS value in the Slough, and was 0 100 200 300 400 500 600 in close proximity to locations *-C olumbia Slough Screening Lev el ppm Zinc in the Middle Slough that Figure 14: Zinc by Slough section. exceeded the screening level (Figure 14). Approximately one-third of the values in both the Lower and Middle Sloughs exceeded the screening level. All values in North Slough, Peninsula Drainage Canal, Buffalo Slough and Upper Slough were below the screening level. ANOVA indicated weakly significant differences amongst the sections9. Figure 13: Zinc by year and Slough section.
Comparison with 1994 Zinc by Year Section Year 314* Zinc levels were significantly 1994 Year LOWER SLOUGH higher in 2006 than in 199410. 2006 1994 1994 2006 NORTH SLOUGH There was a large increase in 2006 1994 WAPATO SLOUGH the median in the Lower 2006 1994 Slough from 128 to 273 ppm, PENINSULA DRAINAGE 2006 and an increase in the Middle 1994 MIDDLE SLOUGH Slough median from 212 to 2006 1994 BUFFALO SLOUGH 283 ppm (Figure 13). Most 2006 1994 WHITAKER SLOUGH 2006 section medians were 2006 greater than 1994 with the 1994 UPPER SLOUGH 2006 exception of Buffalo Slough, BIG FOUR CORNERS 1994 which was slightly lower, and 2006 Whitaker Slough, which was 0 100 200 300 400 500 600 *-Columbia Slough Screening Level ppm Zinc
9 Square root-transformed data were normal, and the ANOVA significance test probability (F=2.47 p=0.020) was slightly smaller than the Bartlett’s test probability for equal variances (16.9 p=0.031). A non-parametric Kruskal-Wallis was also significant: H = 18.35 DF = 8 P = 0.019 10 Square root transformed data were normal for 2006 and slightly non-normal for 1994, with equal variances. T- Value = -3.58 P-Value = 0.000
City of Portland BES - 26 - Columbia Slough Sediment Data Analysis essentially equal between the two years.
PCBs Analyses were conducted for a total of nine PCB Aroclors. Six of the nine Aroclors were never detected. The only Aroclors detected were 1248, 1254 and 1260. The results for these Aroclors are presented below.
6.4. PCB 1248
Figure 15: PCB Aroclor 1248 by Slough section. Non-detects are 2006 Pattern shown at the detection limit. All detected values of Aroclor 1248 were in the Aroclor 1248 by Slough Section
Lower Slough, with the Section 10* Values exception of the station in LOWER SLOUGH Detected the eastern end of Wapato NonDetect NORTH SLOUGH Wetland, where the highest detected value in the WAPATO SLOUGH Slough was observed PENINSULA DRAINAGE
(Figure 15). The data MIDDLE SLOUGH appeared highly BUFFALO SLOUGH dichotomous: the Aroclor was either detected without WHITAKER SLOUGH qualification or were UPPER SLOUGH undetected, and there was BIG FOUR CORNERS a large difference between 0 20 40 60 80 100 120 140 ppb Aroclor 1248 the detection limit and the *-Columbia Slough Screening Level detected values. This suggests that the detection limits for Aroclor 1248 are adequate to capture Aroclor 1248 at the levels at which it is typically present in the Slough. This is in contrast to the patterns in the 1994 data, as described below. Because of the high proportion of nondetects for this analyte, no statistics were conducted to evaluate differences amongst sections. However, the Lower Slough is clearly the area in which Aroclor 1248 is most frequently detected, and values ranged from six to ten times the screening level. The highest value in the Slough (124.6 ppb) was observed at WAS03, the station at the east end of Wapato Slough (Figure 19).
Comparison with 1994 In general, the detection limits improved greatly over the previous sampling effort in 1994. Detection limits were much lower and consistent, whereas 1994 detection limits varied greatly from sample to sample. In 1994 detection limits ranged from 3.3 – 280 ppb, with an average detection limit of 26.8 ppb; in 2006 the detection limit was 0.15 ppb for all samples. In 1994 only one value was detected in the random stations in the entire Slough, and that value was well below the levels of many of the detection limits that were achieved at other stations across the Slough. With such a dataset it is difficult to tell whether a nondetect is due to very low values of the analyte, or to analytical interferences that prevent detection. As is evident from Figure 16,
City of Portland BES - 27 - Columbia Slough Sediment Data Analysis the advances in analytical technology that have occurred from 1994 to 2006 have greatly improved the ability to evaluate patterns of PCB contamination in sediments. The high and highly variable detection limits, and very low level of detection of Aroclors in the 1994 data made comparisons between the two sampling periods of limited value. A comparison of Aroclor 1248 is provided for illustrative purposes, but not for the other Aroclors. One qualitative difference in comparing the two years of data is that the high detected values in 2006 in the Lower Slough and in Wapato Slough are well above the 1994 nondetected values in those sections, suggesting that the 2006 data had higher levels that were not observed in 1994.
Figure 16: PCB 1248 by year and Slough section. Non-detects are shown at the detection limit.
PCB 1248 by Section, Year SECTION YEAR 10* 1994 YEAR Detect? LOWER SLOUGH 2006 1994 Detect 1994 1994 Nondetect NORTH SLOUGH 2006 2006 Detect 1994 2006 Nondetect WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006 1994 JOHNSON LAKE 2006 0 50 100 150 200 250 300
*-Columbia Slough Screening Level Aroclor (ppb)
6.5. PCB 1254 Aroclor 1254 was more frequently and broadly detected than the other two detected Aroclors. Peninsula Drainage Canal and Buffalo Slough were the only sections where Aroclor 1254 was never detected (Figure 17). The largest number and highest detections were in the Lower Slough, which had four of the highest values detected. The highest value (329.5 ppb) was detected at LS22, close to outfall 65 just below the MCDD pump station. Roughly 1.5 miles downstream are the three next highest values, in close proximity to I-5 and Outfalls, 61, 61A, 62 and 62A (Figure 19).
City of Portland BES - 28 - Columbia Slough Sediment Data Analysis
Aroclor 1254 by Slough Section Section 24* Values LOWER SLOUGH Detected NonDetect NORTH SLOUGH
WAPATO SLOUGH
PENINSULA DRAINAGE
MIDDLE SLOUGH
BUFFALO SLOUGH
WHITAKER SLOUGH
UPPER SLOUGH
BIG FOUR CORNERS
0 50 100 150 200 250 300 350 *-Columbia Slough Screening Level ppb Aroclor 1254
Figure 17: PCB 1254 by Slough section. No-detects are shown at the detection limit.
6.6. PCB 1260 Aroclor 1260, like Aroclor 1248, was infrequently detected and only detected within two sections – the Lower Slough and Whitaker Slough (Figure 18). The highest value at nine times the screening level (91.2 ppb) was detected at LS06, the same station where the highest lead value was found, in the vicinity of the St. Johns Landfill and outfalls 53E and 54 (Figure 19). Whitaker Slough had the other three detections, which ranged from four to seven times the screening level. Two were located just upstream of I-205 and one just below the outlet of Johnson Lake.
Aroclor 1260 by Slough Section Section 10* Values LOWER SLOUGH Detected NonDetect NORTH SLOUGH
WAPATO SLOUGH
PENINSULA DRAINAGE
MIDDLE SLOUGH
BUFFALO SLOUGH
WHITAKER SLOUGH
UPPER SLOUGH
BIG FOUR CORNERS
0 10 20 30 40 50 60 70 80 90 ppb Aroclor 1260 *-Columbia Slough Screening Lev el
Figure 18: PCB 1260 Slough section.
City of Portland BES - 29 - Columbia Slough Sediment Data Analysis
General Observation on PCB Aroclors One noticeable pattern in the PCB Aroclor data was that for any one station only one Aroclor was ever detected, if any at all were detected. None of the Aroclors were detected at the same station. This is likely due to analytical limitations and the nature of Aroclors. Aroclors are commercial mixtures of congeners, and once an Aroclor is released into the environment the various congeners undergo differential rates of weathering and breakdown, and so the mixture becomes less and less like the original Aroclor over time. Identifying Aroclor patterns in the analytical output thus becomes very difficult, and it may be difficult to “see” a less concentrated Aroclor behind the pattern of the Aroclor in higher concentration.
City of Portland BES - 30 - Columbia Slough Sediment Data Analysis
Figure 19: PCB Aroclors across the Slough.
City of Portland BES - 31 - Columbia Slough Sediment Data Analysis
Pesticides
6.7. Alpha Chlordane
2006 Pattern Figure 20: Alpha chlordane by Slough section. The highest values of alpha Alpha Chlordane by Slough Section chlordane were measured at Section 1* Whitaker Slough, which had seven Values LOWER SLOUGH of the highest values measured in Detected Estimated NORTH SLOUGH the Columbia Slough (Figure 20), a NonDetect mean and median value twice the WAPATO SLOUGH value of all other sections, and only PENINSULA DRAINAGE one value below the screening level MIDDLE SLOUGH of 1 ppb11. ANOVA indicated highly significant differences amongst BUFFALO SLOUGH 12 sections , with Whitaker Slough WHITAKER SLOUGH being significantly higher than UPPER SLOUGH Buffalo Slough, North Slough and BIG FOUR CORNERS Peninsula Drainage Canal. The highest values in Whitaker Slough 0 1 2 3 4 5 6 7 8 9 *-Columbia Slough Screening Level ppb A-CHLORDANE were all measured above, below and within the Prison Pond area (Figure 26). The lowest values were in the lower 2 miles of the Lower Slough, North Slough and the Peninsula Drainage Canal. The North Slough and Peninsula Drainage Canal were the only sections with all values below the screening level of 1 ppb.
Comparison with 1994 As with most of the organochlorine compounds, detection limits in 2006 were much lower and less variable than in 1994 (Figure 21). Buffalo Slough had an extremely high detection limit of 74 ppb on one sample in 1994 due to analytical interference, but even excluding this sample the detection limits in 1994 ranged as high as two orders of magnitude above the detection limits attained in 2006.
11 The chlordane screening level is for CAS #12789-03-6, or “technical” (i.e., total) chlordane. Technical chlordane is typically a blend of 12 to 20 chlordane and chlordane-like compounds. Though there are more than 140 possible compounds in technical chlordane, alpha and gamma-chlordane typically compose approximately 60-85% of technical chlordane (ATSDR 1994). For the purposes of evaluating alpha- and gamma-chlordane individually, each was compared to the 1 ppb screening level for total chlordane. 12 Ln-transformed data were still non-normal, but with equal variances. Differences amongst sections were highly significant with both nonparametric (H = 30.26 P = 0.000) and parametric (F=6.14 p=0.000) tests. Non-detects were set at the detection limit for these analyses.
City of Portland BES - 32 - Columbia Slough Sediment Data Analysis
The alpha chlordane data were highly non-normal, and Alpha Chlordane by Section, Year standard transformations did Section Year 1* 1994 Year Detect? not produce normality. A LOWER SLOUGH 2006 1994 Detected nonparametric Mann-Whitney 1994 1994 NonDetect NORTH SLOUGH 2006 2006 Detected U test was used to compare 1994 2006 NonDetect WAPATO SLOUGH differences between the two 2006 1994 PENINSULA DRAINAGE years. Although these results 2006 1994 should be interpreted with MIDDLE SLOUGH 2006 caution because of the highly 1994 Buffalo Slough 1994 BUFFALO SLOUGH ND @ 74 2006 skewed data, they indicated 1994 WHITAKER SLOUGH with high significance that the 2006 1994 UPPER SLOUGH alpha chlordane values were 2006 1994 higher in 2006 than in 1994. BIG FOUR CORNERS 2006 This result is notable since a 0 5 10 15 20 highly conservative assumption A-Chlordane (ppb) was used in setting nondetect values equal to the detection Figure 21: Alpha chlordane by Slough section and year. limit, and nondetects were much more common and higher in 1994. The higher levels in 2006 are driven by much larger median values in Whitaker Slough, Upper Slough, Big Four Corners and Lower Slough (Figure 21).
6.8. Gamma Chlordane
2006 Pattern Figure 22: Gamma chlordane by Slough section. Gamma chlordane had a much higher proportion of results Gamma Chlordane by Slough Section qualified as estimates (Figure Section 1* Values LOWER SLOUGH 22). Upper Slough had the Detected Estimated NORTH SLOUGH highest value (18.6 ppb) at a NonDetect station (US07) two stations below WAPATO SLOUGH Fairview Lake (Figure 26) PENINSULA DRAINAGE although this was qualified as an estimate. Whitaker Slough had MIDDLE SLOUGH five of the six and nine of the 13 BUFFALO SLOUGH highest values. Similar to alpha WHITAKER SLOUGH chlordane, the North Slough and UPPER SLOUGH Peninsula Drainage Canal were the only sections with all values BIG FOUR CORNERS below the screening level of 1 0 5 10 15 20 ppb G-CHLORDANE ppb. All values in Whitaker *-Columbia Slough Screening Level Slough and Wapato Slough exceeded this value, and all but one of the values in Upper Slough and Big Four Corners. Gamma chlordane had the same problems meeting statistical assumptions that were evident for alpha chlordane, so that tests of significance should be weighed with caution. Both parametric and nonparametric tests indicated highly significant differences amongst sections, with Whitaker Slough being significantly higher than North Slough, Peninsula Drainage Canal and Lower Slough.
City of Portland BES - 33 - Columbia Slough Sediment Data Analysis
Comparison with 1994 The 1994 data had two very high detected values; 130 ppb in Figure 23:Gamma chlordane by year and Slough section. Buffalo Slough and 48 ppb in Lower Slough. Interestingly, the G-Chlordane by Section, Year parametric t-test showed no SECTION Year 1* 1994 '94 Lower Slough LOWER SLOUGH significant differences between 2006 48 ppm 1994 the two years, whereas the WAPATO SLOUGH 2006 nonparametric Mann-Whitney U 1994 MIDDLE SLOUGH 2006 test indicated that gamma 1994 NORTH SLOUGH chlordane in 2006 was 2006 1994 PENINSULA DRAINAGE significantly higher than in 1994. 2006 1994 '94 Buffalo Slough This is likely due to the fact that BUFFALO SLOUGH 2006 130 ppm nonparametric statistics would 1994 WHITAKER SLOUGH 2006 Year Detect? decrease the influence of the two 1994 1994 Detected UPPER SLOUGH extreme values in 1994. The 2006 1994 NonDetect 1994 2006 Detected BIG FOUR CORNERS violation of statistical 2006 2006 NonDetect assumptions means that the 0 5 10 15 20 tests of significance should be G-Chlordane (ppb) viewed with caution. There were large increases in the median in Whitaker Slough, Upper Slough and Big Four Corners, although many of these data were qualified as estimates (Figure 23).
6.9. Dieldrin
2006 Pattern The highest values of dieldrin were Figure 24: Dieldrin by Slough section. measured at Whitaker Slough, which had seven of the eight Dieldrin by Slough Section highest values measured in the Section 1* Values Columbia Slough (Figure 24) LOWER SLOUGH Detected and a median value five times Nondetect NORTH SLOUGH the screening level of 1 ppb. Middle Slough (3.4 ppb) and WAPATO SLOUGH Upper Slough (2.0 ppb) also PENINSULA DRAINAGE had median values above the MIDDLE SLOUGH screening level, whereas BUFFALO SLOUGH dieldrin was not detected at any stations in North Slough, WHITAKER SLOUGH Peninsula Drainage Canal, and UPPER SLOUGH
Buffalo Slough. The data did BIG FOUR CORNERS not meet statistical 0 10 20 30 40 50 60 70 ppb Dieldrin *-Columbia Slough Screening Level
City of Portland BES - 34 - Columbia Slough Sediment Data Analysis assumptions, but non-parametric analysis suggested highly significant differences with section13
Comparison with 1994 As with most of the organochlorine compounds, detection limits in 2006 were much lower and less variable than in 1994, ranging from one to two orders of magnitude lower in 2006 as compared to 1994. Large increases in the median were evident at Whitaker Slough (1994 median=1.4; 2006=5.2 ppb) and Middle Slough (1994=1.7; 2006=3.4 ppb) (Figure 25). In the sections where no dieldrin was detected - North Slough, Peninsula Drainage Canal, and Buffalo Slough – there were large decreases in the median due to the large decreases in detection limits. There were no significant differences between the two years, possibly due to the confounding effects of decreasing detection limits and increasing levels in the sections where dieldrin was found in higher concentrations.
Figure 25: Dieldrin by year and Slough section.
Dieldrin by Section, Year SECTION Year 1* 1994 Year Detect? LOWER SLOUGH 2006 1994 Detect 1994 Nondetect 1994 WAPATO SLOUGH 2006 Detect 2006 2006 Nondetect 1994 MIDDLE SLOUGH 2006 1994 NORTH SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006 0 10 20 30 40 50 60 70 *-C olumbia Slough Screening Lev el Dieldrin (ppb)
13 Data were non-normal with unequal variances, even with data transformations. Kruskal-Wallis indicated highly significant differences with section: H = 25.50 DF = 8 P = 0.001
City of Portland BES - 35 - Columbia Slough Sediment Data Analysis
Figure 26: Alpha- and gamma-chlordane across the Slough.
City of Portland BES - 36 - Columbia Slough Sediment Data Analysis
6.10. Sum of DDTs A calculation of the sum of all DDT compounds – 2,4’-DDT, 4,4’-DDT, 2,4’-DDE, 4,4’-DDE, 2,4’- DDD and 4,4’-DDD was used to provide a cumulative picture of the variation in DDT compounds across the Columbia Slough in 2006. No analyses of 2,4’-DDx compounds were conducted in 1994, so only the 4,4’-DDx compounds were used in the Total DDTs sum comparing the two years. For these analyses, non-detected samples were set at the detection limit.
2006 Pattern Whitaker Slough had the Figure 27: Total DDTs by Slough section. highest levels of DDT compounds (Figure 27). The Total DDTs vs Section two highest values were adjacent to each other in LOWER SLOUGH Prison Pond (Figure 31). It is NORTH SLOUGH not known whether these high WAPATO SLOUGH values are due to localized PENINSULA DRAINAGE
runoff, or whether it is the n o i
t MIDDLE SLOUGH result of settling and c Se accumulation from upstream BUFFALO SLOUGH areas in Whitaker Slough, since the levels throughout the WHITAKER SLOUGH upstream reach are elevated in UPPER SLOUGH comparison to much of the rest BIG FOUR CORNERS of the Slough, and Prison Pond 0 100 200 300 400 ends in a culvert that may act ppb Total DDTs as a settling basin. In general, the area west of where Whitaker Slough branches off from Middle Slough to Big Four Corners was an area of elevated DDx concentrations, perhaps a legacy of past agricultural uses in the area. The differences amongst sections were highly significant14.
Comparison with 1994 The sum of DDTs (using only the 4,4’-DDx compounds) was significantly higher in 2006 than in 199415. This was driven by the fact that all the means and medians except in Buffalo Slough and North Slough were higher in 2006 than in 1994, with the largest changes in Whitaker Slough and Big Four Corners (Figure 28). This pattern is notable given the fact that the very conservative assumption of setting undetected samples at the detection limit was used for this analysis, and detection limits were an order of magnitude or more higher in 1994 than they were
14 Ln-transformed data were normal and had equal variances. ANOVA: F=4.91 p=0.000. 15 Ln-transformed data were slightly non-normal but with equal variances. Both parametric (T-Value = -3.47 P-Value = 0.001) and nonparametric (W = 9464.0 p= 0.0000) were highly significant.
City of Portland BES - 37 - Columbia Slough Sediment Data Analysis in 2006.
Figure 28: Total DDTs by year and Slough section.
Sum DDTs by Year Section Year 1994 Year LOWER SLOUGH 2006 1994 1994 2006 NORTH SLOUGH 2006 1994 WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 BIG FOUR CORNERS 1994 2006 0 50 100 150 200 250 300 350 6.11. Individual 4,4’-DDXp pCob Tompoundstal DDTs
2006 Pattern DDT – DDT had a high proportion of non-detect and estimated results, so statistics were not performed on the data. The three highest values were in Whitaker Slough; two were the adjacent samples (WS09, WS10) in Prison Pond, and then one sample (WS13) above 122nd Ave (Figure 31). The high value at Big Four Corners is at the end of a stream segment with few outfalls nearby. One possible source is erosion and overland runoff from historical and current farming activities in that area. The North Slough was the only section with all values below the screening level of 2.5 ppb. The Lower Slough had only three detected values out of 23 samples, with only one exceeding the screening level (Figure 29). DDE – Most of the 2006 DDE data were detected and unqualified. The two highest values were in Whitaker Slough in Prison Pond, and the two stations at the end of Big Four Corners were the next highest values (Figure 31). All but one of the values at Big Four Corners and Wapato Slough were above the screening level of 7 ppb, as well as all but two of the values in Upper Slough. North Slough and Buffalo Slough were the only sections with all values below the screening level. The medians and means of all sections except North Slough, Peninsula Drainage Canal and Buffalo Slough exceeded the screening level (Figure 29). ANOVA indicated significant differences amongst the reaches, with Buffalo Slough being significantly lower than Lower, Middle and Wapato Slough16. DDD – The patterns in the 2006 DDD data were very similar to the DDE data. Whitaker Slough had the two highest observed values – at Prison Pond – and Big Four Corners had the next
16 Ln-transformed data were slightly non-normal, with equal variances. Both parametric and nonparametric tests were highly significant. ANOVA F= 4.51, p=0.000. Kruskal-Wallis H = 23.21 P = 0.003
City of Portland BES - 38 - Columbia Slough Sediment Data Analysis highest value at BFC01 at the end of the waterbody (Figure 31). North Slough and Buffalo Slough were the only sections with all values below the screening level. The medians and means of all sections except North Slough, Peninsula Drainage Canal and Buffalo Slough exceeded the screening level (Figure 29). ANOVA indicated significant differences amongst sections17, with ANOVA indicated significant differences amongst the reaches, with Buffalo Slough being significantly lower than Lower, Middle and Wapato Slough, and Peninsula Drainage Canal being significantly lower than these three sections, as well as Upper Slough and Big Four Corners.
Comparison with 1994 DDT – No statistics were performed because of the high proportion of non-detects and estimates in both years. Lower Slough, North Slough, Wapato Slough, Buffalo Slough and Upper Slough had slightly lower means and medians in 2006, at least partly as a result of much lower detection limits (Figure 30). Big Four Corners, Whitaker Slough and Peninsula Drainage Canal had some high detections that were not apparent in the 1994 data, and as a result the means were higher in these sections in 2006 than in 1994, although the increase in the median was much smaller in Whitaker, and the median decreased slightly in Big Four Corners. DDE – DDE results were significantly higher in 2006 than in 199418. The medians and means increased in all sections except Buffalo Slough, with the largest increases in Whitaker Slough, Big Four Corners and Wapato Slough (Figure 30; Table 4). DDD - DDE results were significantly higher in 2006 than in 199419. The largest increases in the median and mean were in Whitaker Slough, Big Four Corners, Upper Slough and Middle Slough (Figure 30).
17 Ln-Transformed data were slightly non-normal with equal variances. Both parametric and nonparametric tests were highly significant. ANOVA F=6.60 P=0.000. Kruskal-Wallis H = 35.26 DF = 8 P = 0.000 18 Ln-transformed data had equal variances but were non-normal. Difference was highly significant with both parametric (T-Value = -5.14 P-Value = 0.000) and nonparametric (W = 8861.5, p=0.0000) tests. 19 Ln-transformed data had equal variances but were non-normal. Difference was highly significant with both parametric (T-Value = -2.54 P-Value = 0.012) and nonparametric (W = 10114.0, p= 0.0018) tests.
City of Portland BES - 39 - Columbia Slough Sediment Data Analysis
Figure 29: DDX compounds by Slough section.
4,4'-DDT by Slough Section
Section 2.5* Values LOWER SLOUGH Detected Estimated NORTH SLOUGH NonDetect
WAPATO SLOUGH
PENINSULA DRAINAGE
MIDDLE SLOUGH
BUFFALO SLOUGH
WHITAKER SLOUGH
UPPER SLOUGH
BIG FOUR CORNERS
0 10 20 30 40 50 60 70 80 ppb DDT *-C olumbia Slough Screening Lev el
4,4'-DDE by Slough Section
Section 7* Values LOWER SLOUGH Detected Estimated NORTH SLOUGH NonDetect
WAPATO SLOUGH
PENINSULA DRAINAGE
MIDDLE SLOUGH
BUFFALO SLOUGH
WHITAKER SLOUGH
UPPER SLOUGH
BIG FOUR CORNERS
0 20 40 60 80 100 120 140 160 180 ppb DDE *-C olumbia Slough Screening Level
4,4'-DDD by Slough Section
Section 6.1* Values LOWER SLOUGH Detected Estimated NORTH SLOUGH NonDetect
WAPATO SLOUGH
PENINSULA DRAINAGE
MIDDLE SLOUGH
BUFFALO SLOUGH
WHITAKER SLOUGH
UPPER SLOUGH
BIG FOUR CORNERS
0 20 40 60 80 100 ppb DDD *-C olumbia Slough Screening Lev el
City of Portland BES - 40 - Columbia Slough Sediment Data Analysis
Figure 30: DDX compounds by year and Slough section.
4,4'-DDT by Se ction, Year SECTION Year 7* 1994 Year Detect? LOWER SLOUGH 2006 1994 Detect 1994 Nondetect 1994 NORTH SLOUGH 2006 Detect 2006 2006 Nondetect 1994 WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006 0 10 20 30 40 50 60 70 80 *-Columbia Slough Screening Level 4,4'-DDT (ppb)
4,4'-DDE by Section, Year
Section Year 7* 1994 Year Detect? LOWER SLOUGH 2006 1994 Detect 1994 Nondetect 1994 NORTH SLOUGH 2006 Detect 2006 2006 Nondetect 1994 WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006 0 20 40 60 80 100 120 140 160 180 *-Columbia Slough Screening Level 4,4'-DDE (ppb)
4,4'-DDD by Section, Year
Section Year 6.1* 1994 Year Detect? LOWER SLOUGH 2006 1994 Detect 1994 Nondetect 1994 NORTH SLOUGH 2006 Detect 2006 2006 Nondetect 1994 WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006 0 20 40 60 80 100 *-Columbia Slough Screening Level 4,4'-DDD (ppb)
City of Portland BES - 41 - Columbia Slough Sediment Data Analysis
Figure 31: Total DDTs across the Slough.
City of Portland BES - 42 - Columbia Slough Sediment Data Analysis
Semi-Volatiles
6.12. Total PAHs
2006 Pattern Total PAHs were calculated as the sum of the following PAH compounds: naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benzo(a)anthracene, chrysene, benzo(a)pyrene, indeno(1,2,3-cd)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene, benzo(b)fluoranthene, benzo(j)fluoranthene, benzo(k)fluoranthene. To calculate these sums, nondetected samples (which were rare for PAHs) were set at the detection limit. The highest value for total PAHs in the Columbia Slough was observed in the Upper Slough at US07 (Figure 32; Figure 33). This station is outside Portland so no information is available on nearby outfalls or drainage areas. It is two stations downstream of Fairview Lake, Figure 32: Total PAHs by Slough section. but the station just below the lake has dramatically lower Total PAHs vs Section
PAH levels so the lake is LS20SED LOWER SLOUGH unlikely to be the source of the contamination. NORTH SLOUGH
Other high values were WAPATO SLOUGH detected near both the I-205 PENINSULA DRAINAGE n o i and I-5 transportation t MIDDLE SLOUGH c
corridors (Figure 38), which Se would be expected since BUFFALO SLOUGH WS08SED PAHs are strongly associated WHITAKER SLOUGH W S07SED US07SED with petroleum products. UPPER SLOUGH
Concentrations appeared to BIG FOUR CORNERS be significantly different among sections20. 0 5000 10000 15000 20000 ppb Total PAHs
Comparison with 1994 Total PAHs were significantly higher in 2006 than in 199421. The largest increases in the median and mean occurred in Lower Slough, Wapato Slough and Upper Slough. All the 2006 values in Wapato Slough are higher than the maximum value in Wapato Slough in 1994 Figure 34.
20 Parametric ANOVA did not detect any significant differences amongst sections (F=1.36, p=0.228). Non parametric Kruskal-Wallis did indicate significant differences ( H = 19.54 DF = 8 P = 0.012), perhaps due to the non-normality of the data affecting the power of the parametric statistical test. 21Ln-Transformed data were slightly non-normal but with equal variances. T-Value = -6.04 P-Value = 0.000
City of Portland BES - 43 - Columbia Slough Sediment Data Analysis
Figure 33: Total PAHs across the Slough.
City of Portland BES - 44 - Columbia Slough Sediment Data Analysis
Total PAHs by Year
SECTION Year 1994 Year LOWER SLOUGH 2006 1994 1994 2006 NORTH SLOUGH 2006 1994 WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006 0 5000 10000 15000 20000 ppb Total PAHs
Figure 34: Total PAHs by year and Slough section.
6.13. Benzo(k)fluoranthene There was a high degree of correlation amongst the PAH compounds. Patterns in benzo(k)fluoranthene are presented below since this had the greatest number of samples exceeding the screening level, but the patterns are representative of most other PAHs because of strong correlations in their concentrations.
2006 Pattern The highest values of benzo(k)fluoranthene occurred in Benzo(k)fluoranthene by Slough Section Whitaker Slough (WS08) near I- Section 87* Detected? LOWER SLOUGH 205 and in the Upper Slough Detected Estimated (US07) two stations below NORTH SLOUGH Fairview Lake. High values were WAPATO SLOUGH also detected in the Lower Slough in the vicinity of I-5. All PENINSULA DRAINAGE values in North Slough, Buffalo MIDDLE SLOUGH Slough and Peninsula Drainage BUFFALO SLOUGH
Canal were below the screening WHITAKER SLOUGH level of 87 ppb. The Lower UPPER SLOUGH Slough was the only section where both the median and the BIG FOUR CORNERS mean exceeded the screening 0 200 400 600 800 1000 1200 ppb Benzo(k)fluoranthene level (Figure 35). No significant differences amongst sections Figure 35: Benzo(k)fluoranthene by Slough section. were detected in statistical tests.
City of Portland BES - 45 - Columbia Slough Sediment Data Analysis
Comparison with 1994 Benzo(k)fluoranthene was significantly higher in 2006 than in 199422. Large increases in the median were evident in most sections except the North Slough, Peninsula Drainage Canal, and Buffalo Slough, which showed smaller increases. While none of the medians were above the screening level in 1994, six of the nine medians exceeded the screening level in 2006.
Benzo(k)fluoranthene by Year Section Year 87 1994 LOWER SLOUGH Year 2006 1994 NORTH SLOUGH 1994 2006 2006 1994 WAPATO SLOUGH 2006 1994 PENINSULA DRAINAGE 2006 1994 MIDDLE SLOUGH 2006 1994 BUFFALO SLOUGH 2006 1994 WHITAKER SLOUGH 2006 1994 UPPER SLOUGH 2006 1994 BIG FOUR CORNERS 2006
JOHNSON LAKE 1994 2006 0 200 400 600 800 1000 1200 ppb Benzo(k)fluoranthene
Figure 36: Benzo(k)fluoranthene by year and Slough section.
6.14. Bis(2-ethylhexyl)phthalate
2006 Pattern Middle Slough and Whitaker Slough had the highest maximum and median concentrations of bis(2-ethylhexyl)phthalate, and were the only sections in which the median exceeded the screening level. North, Wapato, Peninsula Drainage and Buffalo all had low values well below the screening level.
22 Ln-transformed data were normal with equal variances. T-Value = -6.33 P-Value = 0.000
City of Portland BES - 46 - Columbia Slough Sediment Data Analysis
Bis(2-ethylhexyl)phthalate by Slough Section Section 1816* LOWER SLOUGH Values Detected NORTH SLOUGH Estimated
WAPATO SLOUGH
PENINSULA DRAINAGE
MIDDLE SLOUGH
BUFFALO SLOUGH
WHITAKER SLOUGH
UPPER SLOUGH
BIG FOUR CORNERS
0 1000 2000 3000 4000 5000 6000 7000 8000
*-C olumbia Slough Screening Level ppb Bis(2-ethylhexyl)phthalate
Figure 37: Bis(2-ethylhexyl)phthalate across the Slough.
Comparison with 1994 The medians showed large increases at Middle and Whitaker Slough. All medians in 1994 were below the screening level, and 1994 concentrations overall were significantly lower than 2006 concentrations23.
Bis(2-ethylhexyl)phthalate by Year Section Year 1816 LOWER SLOUGH 1994 Year 2006 1994 NORTH SLOUGH 1994 2006 2006 WAPATO SLOUGH 1994 2006 PENINSULA DRAINAGE 1994 2006 MIDDLE SLOUGH 1994 2006 BUFFALO SLOUGH 1994 2006 WHITAKER SLOUGH 1994 2006 UPPER SLOUGH 1994 2006 BIG FOUR CORNERS 1994 2006 0 1000 2000 3000 4000 5000 6000 7000 8000 ppb Bis(2-ethylhexyl)phthalate
Figure 38: Bis(2-ethylhexyl)phthalate by year and Slough section.
23 W = 19705.0; p= 0.0001
City of Portland BES - 47 - Columbia Slough Sediment Data Analysis
6.15. Summary Table of the 2006 Sediment Data
The findings from the previous analyses are summarized in Table 5
Table 5: Summary Table of the 2006 Sediment Data Metals PCBs Copper Chromium Lead Zinc 1248 1254 1260
Wapato Wapato Lower Lower Wapato Lower Lower Highest Value (WAS03 - (WAS03 - (LS06 - (LS20 - (WAS03 - (LS22 - (LS06 - 297.0) 320.0) 177.0) 576.0) 124.6) 329.5) 91.2)
Wapato - Buffalo - Wapato Wapato Wapato - Wapato - Highest Median All - 0.15 90.7 97.6 - 77.6 - 315.0 0.2 39.7 Buffalo - North - Buffalo - Numerous Numerous Lowest Median PDC - 35.7 All - 0.15 22.0 21.7 88.1 - 0.15 - 0.16 Screening Level 54 58 90 314 10 24 10 Differences amongst No No Yes Yes NA NA NA Sections? Differences between Years? Yes Yes No Yes NA NA NA
City of Portland BES - 48 - Columbia Slough Sediment Data Analysis
Table 5 (cont): Summary Table of the 2006 Sediment Data Pesticides Semi-Volatiles Alpha Gamma Total Benzo(k) Dieldrin DDT DDE DDD Chlordane Chlordane PAHs Fluoranthene
Whitaker Upper Whitaker Whitaker Whitaker Whitaker Upper Whitaker Highest Value (WS11 - (US07 - (WS14 - (WS09 - (WS10 - - (WS10 - (US07 - (WS08 - 9.1) 18.6) 66.8) 76.6) 165.6) 94.6) 17753) 1140.3)
Whitaker - Whitaker - Whitaker PDC - Wapato - Wapato - Wapato Whitaker - Highest Median 3.8 4.95 - 5.2 6.3 23.4 18.9 - 2168 195.9 Several - Several - Buffalo - PDC - Buffalo - Lowest Median North - 0.16 North - 0.26 Buffalo - 39.6 0.02 0.02 1.23 0.61 402 Screening Level 1 1 1 2.5 7 6.1 NA 87 Differences amongst Yes Yes Yes NA Yes Yes Yes No Sections? Differences between Years? Yes Yes No NA Yes Yes Yes Yes
City of Portland BES - 49 - Columbia Slough Sediment Data Analysis
Relationship between 2006 sediment data and 2005 fish tissue data Several approaches were taken in an attempt to identify correlations between the 2006 sediment data and the 2005 fish tissue data. There were two key challenges in accomplishing this effort. The first was the mobile nature of fish, which means that over the course of their life they move around and accumulate contaminants from a number of locations, and since the distance over which they roam and the length of time they stay in any one location is unknown, it is difficult to know which sediment stations should be used, or how much each station should be weighted to capture residence time at that location. The second was that the spatial coincidence of fish sampling locations and sediment stations varied; in some situations there were one or more stations very close to the fish sampling locations while others were further away. The approaches used to correlate sediment contaminant concentrations with fish tissue contaminant concentrations were to regress fish tissue concentrations against 1) the closest sediment station, and 2) all sediment stations within a mile of the fish sampling location (under the assumption that carp have a home range of roughly a mile). A final approach was to compare fish and sediment concentrations by Slough sections. Generally these approaches revealed very weak or statistically non-significant relationships between sediment and fish tissue contaminant concentrations. Examples of the relationships between sediment contaminant concentrations and fish tissue concentrations are provided, using zinc and 4,4-DDE, since these are contaminants that were generally detected in both fish and sediment with a limited number of qualifiers, and were contaminants of concern because they exceeded sediment screening levels and fish tissue Acceptable Tissue Levels. The first approach was to use the sediment station nearest where the net was set to collect fish tissue samples. Tissue concentrations for all fish collected in the same net were averaged, since they were not statistically independent, and because there was only one sediment station against which to regress the fish levels. Using this approach, there was no significant relationship between sediment zinc and fish tissue zinc:
Source DF SS MS F P24 Regression 1 93.6 93.6 0.36 0.558 Residual Error 15 3919.1 261.3 Total 16 4012.6
S = 16.1639 R-Sq = 2.3% R-Sq(adj) = 0.0%
24 The abbreviations are from a standard ANOVA table. DF=degrees of freedom; SS=sum of squares; MS=mean square; F=F statistic; P=probability of obtaining an F value as large or larger than the listed F statistic.
City of Portland BES - 50 - Columbia Slough Sediment Data Analysis
Fish vs. Sediment Zinc using Nearest Station 120
110 ) c n i 100 Z m p p
( 90 sh i F f 80 e o
ag 70 er v A 60
50 0 100 200 300 400 500 600 Sediment (ppm Zinc)
Figure 39: 2006 Fish tissue zinc levels plotted against zinc levels in the sediment station nearest to the fish tissue collection point.
The results for 4,4’-DDE were also non-significant using the nearest sediment station:
Source DF SS MS F P Regression 1 2227 2227 1.25 0.281 Residual Error 15 26658 1777 Total 16 28884
S = 42.1565 R-Sq = 7.7% R-Sq(adj) = 1.6%
Figure 40: 2006 Fish tissue DDE levels plotted against DDE levels in the sediment station nearest to the fish tissue collection point.
Sediment vs. Fish DDE using Nearest Station 180
160
) 140 b p
p 120 ( E D 100 D sh i 80 e F 60 erag v
A 40
20
0 0 2 4 6 8 10 12 14 16 18 Sediment DDE (ppb)
City of Portland BES - 51 - Columbia Slough Sediment Data Analysis
The relationship using sediment stations within one mile of where fish were captures was also non-significant: Source DF SS MS F P Regression 1 120.7 120.7 0.48 0.500 Residual Error 15 3796.2 253.1 Total 16 3916.9 S = 15.9085 R-Sq = 3.1% R-Sq(adj) = 0.0%
Figure 41: 2006 Fish tissue zinc levels plotted against zinc levels in the sediment stations within one mile of the fish tissue collection point.
Fish vs Sediment Zinc Using Stations w/in 1 mile 120
110 )
c 100 n i Z 90 ppm ( h s i 80 F n a e 70 M
60
50 100 200 300 400 500 Mean Sediment (ppm Zinc)
The relationship with 4,4’-DDE was similarly non-significant. Aggregating the analysis up to the section level did not produce a significant relationship. For example, the 4,4’-DDE relationship at the section level was:
Source DF SS MS F P Regression 1 1121 1121 0.59 0.473 Residual Error 6 1458 1910 Total 7 12579
S = 43.7001 R-Sq = 8.9% R-Sq(adj) = 0.0%
City of Portland BES - 52 - Columbia Slough Sediment Data Analysis
Scatterplot of Fish vs Sed
140
120
100 h
s 80 Fi
60
40
20
0 5 10 15 20 25 30 35 Sed
Figure 42: 2006 Fish tissue zinc levels plotted against zinc levels in the sediment stations within each Slough section.
Part of the problem in identifying a significant relationship between sediment and fish tissue is the fact that some of the fish that had to be grouped together because they were collected in the same location had highly variable tissue concentrations among the fish. For example, in the plot below, fish collected at sampling location I (Middle Slough) and F (mouth of Whitaker Slough) had a wide range of tissue concentration values (Figure 43).
Fish Tissue Contaminant Variability w/in a Sampling Location 200
150 ) b p p ( E D
D 100 e u ss i T sh i
F 50
0 A B C D E F G H I J K L M N O P Q Sampling Group
Figure 43: Variability in tissue DDE levels among fish collected at the same location.
Comparison of 2006 and 1994 data to special study areas In response to the findings of the 1994 SLRA, the city initiated more focused special studies of some key priority areas, including Buffalo Slough, Wapato Slough, and the Mark-Whitaker subbasin (Figure 44). The results from these special studies are compared the the Slough-wide studies in 1994 and 2006 below. City of Portland BES - 53 - Columbia Slough Sediment Data Analysis
Figure 44: Locations of special studies initiated in response to the findings of the 1994 Screening Level Risk Assessment.
City of Portland BES - 54 - Columbia Slough Sediment Data Analysis
Buffalo Slough In 1995, BES conducted a focused study of contamination in Buffalo Slough based on the results of the 1994 SLRA. The results of some of the key contaminants of concern are shown below in comparison to results from Buffalo Slough obtained during the 2006 and 1994 Slough- wide sampling events. For chromium, the results from 1994 and 2006 are very consistent with the more extensive results from the 1995 focused study. The results from all years show the consistent pattern of lower contaminant concentrations in subbasins A, B and C, and much higher concentrations in subbasin D (Figure 45). Chromium levels in subbasin A are at or below the screening level value of 58 ppm, levels in subbasin B and C are one to two times the screening level, and levels in subbasin D generally range from three to five times the screening level. The levels in subbasin D appear to exhibit an increasing trend from west to east. Interestingly, there are two exceptions to this pattern in subbasin D: two of the values at the eastern end of subbasin D – which are adjacent to the only City stormwater outfall (Outfall #73) and one of the three private outfalls in subbasin D (CS-345), are among the lowest chromium values in all of Buffalo Slough. These results would tend to suggest that these two outfalls are not the source of chromium, and that the other two private outfalls (CS-453 & CS-454) or a more general source such as non- point or overland runoff are possible sources of the chromium in subbasin D. It is also important to consider, as noted on page 20, that stations nearby at the mouth of Whitaker Slough also had high chromium values, so air deposition or a regional source are also a possibility. The results suggest that there is a significant source of chromium in subbasin D, and that elevated levels of chromium were present in 1994 and continue to be present in 2006.
City of Portland BES - 55 - Columbia Slough Sediment Data Analysis
Figure 45: Comparison of the 1995 Buffalo Slough study chromium data to the 1994 and 2006 results.
Alpha-chlordane – another key risk driver in Buffalo Slough – exhibits a different pattern amongst the subbasins. The frequency of detection of alpha-chlordane is highest in subbasins A & B, with none of the samples in subbasin C and only one of the 23 samples in subbasin D resulting in detected values (Figure 46). The patterns of contamination are confounded by highly variable detection limits in 1994 and 1995. For example, the detected values in subbasins A & B are lower than the detection limits for many of the samples in subbasins C & D. However, the pattern of less frequent detections and possibly lower values in the two eastern subbasins is somewhat corroborated by the 2006 results, which showed detections in subbasins A (0.84 ppm) and B (1.51 ppm) and a nondetection in subbasin D with a detection limit of 0.02 ppm. Gamma-chlordane – another key risk driver in Buffalo Slough – was detected in only one of the 50 samples taken in 1995. For this reason the results for this risk driver were not plotted, since the results from 1995 provide little insight beyond the 1994 and 2006 results plotted in Figure 23 and Figure 26.
City of Portland BES - 56 - Columbia Slough Sediment Data Analysis
Figure 46: Comparison of the 1995 Buffalo Slough study alpha chlordane data to the 1994 and 2006 results.
Wapato Slough BES conducted a sediment characterization of Wapato Wetland in 1997 in response to potential ecological risks identified in the 1994 SLRA associated with elevated metals near outfall 55A which discharges to the wetland. The 1997 focused study showed that metals were generally high in the vicinity of the outfall, but decreased with distance away from the outfall. The 1994 and 2006 sampling locations mostly fall outside the area of the 1997 focused study, but the two stations that do fall within the 1997 study area are consistent with the spatial patterns observed in the 1997 study for chromium (Figure 47), copper (Figure 48) and nickel (Figure 49), three of the contaminants of concern.
City of Portland BES - 57 - Columbia Slough Sediment Data Analysis
Figure 47: Comparison of the 1975 Wapato Slough study chromium data to the 1994 and 2006 results.
Based on the results of the 1997 focused study, BES installed a pollution reduction facility at Outfall 55A to address this source of pollutants to the ecologically valuable wetlands. Since the contamination from 55A was limited to the immediate vicinity of the outfall, and the 2006 samples did not occur within this limited area, the recent samples do not provide information on the effectiveness of the stormwater treatment. However, the 1994 and 2006 sampling efforts extended further east into the wetland, showed a consistent pattern of increasing metals concentrations at the eastern end of Wapato Slough, and identified what is likely another area of potential concern in Wapato Slough. Chromium, copper and nickel – three contaminants of concern at Wapato Slough – showed similar spatial patterns, although with some important variations. Chromium exhibited a peak in the immediate vicinity of Outfall 55A well above the Columbia Slough Screening Level, with lower levels generally below the screening level moving away from the outfall. This pattern persisted until approximately 2000 ft upstream of the outfall, at which point chromium levels increase dramatically and actually reach their highest point at the sampling point furthest upstream (east) in the wetland (Figure 47). Copper shows a similar pattern, although the value at the furthest east end of the wetland is lower than the peak reached in the vicinity of the outfall (Figure 48). Nickel exhibits another variation on this pattern. The peak in the vicinity is not as pronounced as for the other two metals, but the peak at the eastern end of the wetland is (Figure 49). This 2006 station at the eastern end is the highest value of nickel observed in the entire Columbia Slough by far (841 ppm), and is over an order of magnitude higher than the next highest value observed at the station just downstream (66.3 ppm). City of Portland BES - 58 - Columbia Slough Sediment Data Analysis
Figure 48: Comparison of the 1997 Wapato Slough study copper data to the 1994 and 2006 results.
While the contaminant sources near Outfall 55A were evaluated by the 1997 focused study and addressed through stormwater treatment, the contamination at the eastern end has not been. The only known city or private outfall at this end of the wetland is city Outfall 56A, which occurs directly adjacent to the 1994 station approximately halfway between the two elevated 2006 stations at the eastern end of the wetland. Since concentrations of the key metals at the easternmost end are over twice the levels immediately adjacent to the outfall, it is either unlikely that the outfall is the only source of metals to the wetland, or there are complex patterns of dispersal and sedimentation that transport the contaminants in an easterly direction and deposit them at the eastern end. Other potential sources include overland runoff from adjacent facilities or illicit discharges.
City of Portland BES - 59 - Columbia Slough Sediment Data Analysis
Figure 49: Comparison of the 1997 Wapato Slough study nickel data to the 1994 and 2006 results.
Marx-Whitaker Subbasin In 1998, BES conducted a focused study of sediment contamination in the Marx-Whitaker subbasin, based on the results of a historical sediment sample evaluated in the SLRA. The historical sample had elevated levels of metals, but did not measure organic contaminants. The focused study found that metals, SVOCs and PCBs did not pose significant ecological risks, but that pesticides (including DDTs, dieldrin, endosulfan II and endosulfan sulfate) and PAHs were elevated in sediment throughout the basin, and posed potential risks to benthic organisms and wildlife. Spatial patterns in the 1998 data show a consistent pattern amongst the contaminants. Levels are generally lower in the 4 upstream stations (T1-T4), and higher in the 4 downstream stations (T5-T8) such that there is little or no overlap in the range in concentrations in these two halves of the subbasin (Figure 50 – Figure 52). Sediment pollutant concentrations show a slight declining trend from T1 (furthest upstream) to T4 in the middle of the subbasin, with T3 and T4 generally showing the lowest concentrations measured in the subbasin in 1998. There is a dramatic increase in going from T4 to T5, with concentrations increasing 5-10-fold between these two stations. Concentrations are generally lower in the downstream stations below T5 (T6-T8) for dieldrin and endosulfan sulfate, while for DDE T8 had a concentration just below the peak observed at T5.
City of Portland BES - 60 - Columbia Slough Sediment Data Analysis
Figure 50: Comparison of the 1998 Whitaker Slough study DDE data to the 1994 and 2006 results.
The only outfalls to discharge to the subbasin between T4 and T5 are two private outfalls: CS215, which discharges from the south bank, and CS-408, which discharges from the north bank. It is not known whether either of these two outfalls are responsible for the marked increases in pesticide concentrations that occurs between the stations upstream and downstream of these stations. Comparison with the 2006 data shows varying patterns. 4,4’-DDE appeared to be markedly lower in 2006, as all three values were below the lowest value observed in 1998 (Figure 50). Dieldrin and endosulfan sulfate do not appear to be lower in 2006. Dieldrin had a similar range between the two years, and the highest 2006 value is just above the highest 1998 value (Figure 51). In contrast, the highest value of endosulfan sulfate in 1998 was 2.5 times the highest value observed in 2006, though the range of all stations other than the peak at T5 was similar to the range observed in 2006 (Figure 52).
City of Portland BES - 61 - Columbia Slough Sediment Data Analysis
Figure 51: Comparison of the 1998 Whitaker Slough study dieldrin data to the 1994 and 2006 results.
City of Portland BES - 62 - Columbia Slough Sediment Data Analysis
Figure 52: Comparison of the 1998 Whitaker Slough study endosulfan sulfate data to the 1994 and 2006 results.
7. CONCLUSIONS Patterns in the 2006 data are consistent with 1994 – many of the same key contaminants and areas of concern as identified in 1994 In general the pattern of sediment contamination in 2006 is consistent with the pattern of sediment contamination in 1994. The key risk drivers and contaminants of concern identified in 1994 continue to exceed screening levels in 2006. Spatial variability in contaminants was roughly similar in the two sampling events. Whitaker Slough and Lower Slough tended to have higher concentrations in both sampling years, and Peninsula Drainage Canal and North Slough tended to have low concentrations in both years.
Improved detection limits have greatly enhanced the ability to capture patterns in the PBT (persistent, bioaccumulative and toxic) contaminants There have been clear and dramatic improvements in the detection limits for PBT contaminants such as PCBs and pesticides. In some cases detection limits have decreased by two or more
City of Portland BES - 63 - Columbia Slough Sediment Data Analysis orders of magnitude from 1994 to 2006, and this has greatly improved the ability to identify current patterns of contamination, and will improve the ability to detect trends in the future.
The 2006 data identified several “hot spots”, or areas with elevated concentrations for several contaminants. These include:
Whitaker Slough The area from the eastern extent of Whitaker Slough through Prison Pond to I-205 had the highest concentrations of a number of contaminants and elevated concentrations of nearly all the contaminants of concern. The highest DDX-compounds, chlordanes, aldrin, dieldrin, and many PAHs were observed in this area. This area also had the five highest concentrations of endosulfan II and endosulfan sulfate. Focused source investigations should be conducted in this area because of the complex assortment of sources, which include I-205 and erosion and runoff from historical and current farms in the eastern portion of the section. Prison Pond has high concentrations of several contaminants, particularly DDTs, but whether these are due to runoff from local outfalls or transport from the upstream reaches high in DDTs and settling and concentration in the waters of Prison Pond impounded by the culvert is unclear.
Lower Slough in the I-5/MLK reach PAHs, PCB Aroclors 1248 and 1254, copper and zinc were all high in this dense transportation corridor. The PAHs, copper and zinc are likely related to highway and street runoff. This area would likely benefit from stormwater treatment of the runoff from the outfalls draining busiest transportation corridors along this section.
The eastern end of Wapato Slough The station at the eastern end of Wapato Slough – WAS03 – had the highest concentration of a number of contaminants. It had the highest concentration of PCB Aroclor 1248, chromium, copper, lindane, heptachlor, naphthalene and nickel. For nickel, the level at this station was over an order of magnitude higher than the rest of the Slough. For lindane and heptachlor, it was the only of the 78 stations across the entire Slough where these two analytes were detected, and the detected value at this site was over two orders of magnitude higher than the detection limit. It had the second highest dibenzofuran concentration, and was in the top ten highest concentrations for fluorene, acenaphthene, phenanthrene and antimony.
The Upper Slough station US07, two stations downstream of Fairview Lake This station was unusual in that it had the highest values in the Slough for several PAHs, Total PAHs and gamma chlordane, yet the stations around this one had very low levels of these contaminants, suggesting a highly localized source, and that contaminant transport is limited in this reach. There is a station between this station and Fairview Lake that does not exhibit similarly high values, suggesting that Fairview Lake is not likely the source of these contaminants. The City of Portland does not have outfall information for this area since it is outside the city, and so cannot provide insight on potential sources. DEQ might consider investigating outfalls in the vicinity of US07, particularly any that drain I-84 because of the very high PAH signal at this station.
To illustrate the hot spots, each station was given a percentile rank for each contaminant – indicating which percentile the station fell into for each contaminant. These were averaged over
City of Portland BES - 64 - Columbia Slough Sediment Data Analysis all contaminants, so that the resulting percentile is the average percentile a given station ranked for all contaminants. The resulting map shows many of the hot spots indicated in the previous discussion (Figure 53). The Lower Slough in the vicinity of I-5/MLK scores highly, with stations LS17, 19, 20 & 21 in the upper 20%. Whitaker Slough near I-205 also had high average percentiles, with stations WS7, 8, 10 & 11 scoring in the top 20%. One finding evident from this analysis that was not evident from the contaminant-specific analyses is that the nearby stations along the Middle Slough – MS7, 8, 9, 10 &11 – also scored highly. The Upper Slough station US07 is also in the top 20% of ranks across contaminants. Source investigations, or re-evaluation of existing source investigations in the context of these findings where source investigations have been conducted, would help to identify effective actions for reducing loadings of key contaminants of concern into the Slough.
The 2006 data also identified areas with consistently low concentrations of contaminants The percentile rank map also shows areas that tended to have low concentrations of contaminants. Peninsula Drainage Canal and North Slough had consistently low concentrations of nearly all the contaminants (Figure 53). The only exceptions to this were that Peninsula Drainage Canal had the highest cadmium concentration in the Slough and the fifth highest antimony concentration, and North Slough had the highest phenol concentration in the Slough. Other than these exceptions the two sections had almost uniformly low concentrations of all the key contaminants of concern. Buffalo Slough also tended to rank low for all contaminants with the exception of the chromium pattern described earlier, and had the fifth highest phenol concentration. The stations at the mouth of the Columbia Slough also had uniformly low concentrations of all contaminants without exception. This is potentially due to the fact that this reach is an erosional environment where sediment contaminants are unlikely to accumulate. These areas could be used as possible “reference site” areas for evaluating Slough contamination in the future. These are areas that are likely exposed to the general region-wide levels of ambient contamination – and so could help to identify “background” levels of contamination – but do not appear to be subject to high levels of local pollutant loads. They may help to characterize the concept of “how clean is clean enough?”; in other words, they characterize contaminant concentrations that would be expected in areas with limited pollutant sources, but below which it would be difficult to attain because of regional pollutant sources, atmospheric deposition and other ambient contaminant sources.
City of Portland BES - 65 - Columbia Slough Sediment Data Analysis
Figure 53: Average of the percentile rank for each station across all contaminants.
City of Portland BES - 66 - Columbia Slough Sediment Data Analysis
Sediment contaminants appear to be higher in 2006 than in 1994 Although the spatial patterns and key contaminants of concern are generally similar between the two sampling events, there was a consistent pattern of higher contaminant concentrations in 2006 than in 1994. It is important to emphasize that there are a number of confounding factors between the two sampling events that could influence these patterns, including different sampling depths, station placement, and dredging of Slough sediments25. The most important of these is that the 2006 sampling event used a 10 cm sampling depth, and this has the potential to sample down into earlier sediments which might tend to have higher contaminant levels. However, 12 years elapsed between the 1994 and 2006 sampling events. Although sedimentation rates in the Slough are unknown, it is a hyper-eutrophic system with significant stormwater inputs of suspended solids, and it is quite possible that a majority if not the entirety of the 10 cm was deposited in the 12 years between the two sampling events. Keeping these caveats in mind, the notable pattern in the comparison of 2006 and 1994 sediment contaminants is the consistency of higher levels observed in 2006. Means and medians were consistently higher in 2006 across analytes and sections, and were generally highly significant under both parametric and nonparametric tests. In addition, the detection limits in 1994 were considerably higher than in 2006, and these analyses used the highly conservative approach of setting the value at the detection limit for contaminants which were non-detected in a sample. While two points in time do not make a trend, the pattern is clear and broad enough to suggest that Slough-wide contamination levels are higher in 2006 than in 1994. The reasons for these potentially higher levels are unknown. For metals and PAHs, it is likely that higher numbers of people, traffic, industries, and development have increased the loading of these contaminants into the Slough. For the organochlorine pesticides, however, the higher levels are an unexpected result. Production and use of all of these compounds has been banned for two or more decades in the U.S. While they are highly persistent compounds that break down very slowly, their use has been banned for enough time that decreases in their environmental concentrations should begin to become evident, and it would seem unlikely that higher concentrations should be observed over time. One possible hypothesis is that these contaminants were so broadly used for so long that amounts have accumulated in soils and other reservoirs, and continue to be introduced into the Slough through erosion and runoff. It is also possible that high rates of development in the upper watershed have increased soil disturbance, erosion and washoff, and increased the rate at which legacy pesticides have been introduced into the Slough, in spite of the fact that the total amount of legacy pesticides have remained constantly or decreased some due to breakdown. This pattern of higher contaminant levels observed in recent sediment sampling is inconsistent with a USGS study in Johnson Creek, which found that DDT had decreased by an order of magnitude in surface waters since 1989-90 (Tanner and Lee 2004). It is also inconsistent with the recent analyses of contaminants in fish tissues, which found decreases in Slough fish tissues for many of these organochlorine pesticides (GeoSyntec and BES 2007). It may be that the sediment represents a repository that accumulates these compounds, but the
25 See footnote 3 on page18 for a description of the important caveats in the comparison between the 2006 and 1994 data.
City of Portland BES - 67 - Columbia Slough Sediment Data Analysis sediment-bound contaminants are not very bioavailable and so appear to be decreasing in surface waters and fish tissues. Further evaluation and source investigation into potential sources and fate and transport of banned organochlorine pesticides and PCBs may help to explain these results. DEQ is investigating the role that dispersion or “spreading out” of contaminated sites might have in elevating the median values across the Slough sections. A preliminary evaluation suggested that in 1994, stations within 200 feet of an outfall tended to have higher average metals concentrations than stations greater than 200 feet, while in 2006 stations within 200 feet of an outfall tended to have lower average metals concentrations than stations further than 200 feet.
Future sampling efforts should investigate approaches for improving the ability to link sediment and fish tissue concentrations The 2005 fish tissue and 2006 sediment data collection efforts revealed very little correlation in their contaminant levels. This may reflect an ecological reality, in that fish are mobile and although a fish may be captured in the same location as another, it may have had very different movement patterns and therefore a different history of exposure to and uptake of sediment contaminants. However, there may be ways to improve the sampling design to improve investigations into relationships between sediment and fish tissue concentrations. These include co-locating fish and sediment collection stations, collecting fish and sediment samples in the same year, and trying to use locations where movement range is limited (e.g., Peninsula Drainage Canal, sections with culverts) to reduce the impact of mobility on tissue variability. The City will also continue analyses using ancillary parameters (lipid content, age) as covariates to explain some of the variability and further examine the possibility of a relationship, but the initial analyses suggest the relationship is non-existent, highly complex, or explained by factors beyond those included in the initial analyses.
Future sampling efforts should be designed to maximize the statistical rigor and comparability of sampling approaches, and enhance the ability to detect trends There are three recommendations that would increase the strength of future sampling efforts to detect changes in sediment concentrations over time. 1) Maintain adherence to a probabilistically-based sampling approach, such as a systematic, stratified random, or generalized random tessellation stratified sampling design. It is acceptable to add stations to such a sampling network to evaluate areas of interest, but every effort should be made to avoid moving stations within the probabilistic network to evaluate such areas, as it weakens the statistical strength of the design. If limited resources preclude using additional stations to address areas of interest, it is preferable to reduce the density of the stations within the network while adhering to its probabilistic design principles (e.g., place stations every ¾ of a mile instead of every ½ mile), and use the extra stations made available for placement in areas of interest. 2) Because a key objective of the sampling effort is to evaluate changes over time, future sampling efforts should sample the top 2 cm of sediment. Commonly used sediment sampling depths are the top 2 or 10 cm. The choice of sampling depth is dependent on the study objectives. The top 10 cm represents the “bio-active zone” in which a majority of benthic organisms move, feed and reside, and it is often used to evaluate risks to benthic organisms. The top 2 cm represents the most recently deposited sediments, and is often used to evaluate changes in sediment contamination over time, since it is likely that these sediments have been recently deposited. Both objectives are relevant to the Columbia Slough sediment sampling effort. However, sampling of the top 10 cm complicates the ability to evaluate changes in sediment contaminants over time, since sediment deposition rates in the Slough are unknown, and it is possible that a 10 cm sampling depth may sample down into a sediment layer that was City of Portland BES - 68 - Columbia Slough Sediment Data Analysis present from the previous sampling period. Although the Slough is hyper-eutrophic and has high inputs of stormwater solids and thus may have high sedimentation rates, sampling the top 2 cm would greatly reduce the ambiguity about whether the sampled sediments are recent or were present during the previous sampling event. 3) Evaluate the periodicity of future sampling events. One of the key objectives stated in the Long Term Monitoring Plan (LTMP) is to track changes in sediment contamination over time. Tracking trends is a complicated effort that requires repeated sampling events over time to separate out actual temporal changes from random variability in the data. The number of sampling events and the time period of sampling required to adequately evaluate trends is highly dependent on the system and constituents being sampled, but as a very rough guide for constituents that can be sampled annually (i.e., that don’t have strong seasonal variation) it generally takes 5-10 sampling events to begin to detect a trend. The LTMP calls for sampling sediment contaminants every ten years. However, this could mean that changes over time or not evident until 50-100 years have passed. BES should work with the BES statistician to evaluate the best sampling design for detecting changes over time. If the evaluation suggests more frequent sampling is recommended and the level of resource required poses concerns, options such as a reduced number of stations or constituents analyzed could be explored to increase the ability to detect trends without unduly increasing costs.
8. REFERENCES ATSDR 1994. Toxicological Profile for Chlordane. Agency for Toxic Substances and Disease Registry, http://www.atsdr.cdc.gov/toxprofiles/tp31.html. ATSDR 2005. Toxicological Profile for Arsenic, Draft for Public Comment. Agency for Toxic Substances and Disease Registry, http://www.atsdr.cdc.gov/toxprofiles/tp2.html. BES 1995. Columbia Slough Sediment Project, Screening Level Risk Assessment Report, City of Portland, Bureau of Environmental Services. 1995a. February 1995. BES 1997. Columbia Slough Sediment Remedial Investigation/Feasibility Study, Endangerment Assessment Report for Buffalo Slough. Final Report, City of Portland, Bureau of Environmental Services. January 1997. DEQ 2006. Draft Columbia Slough Screening Levels. http://www.deq.state.or.us/lq/cu/nwr/ColumbiaSlough/ GeoSyntec and BES. 2007. Columbia Slough Fish Tissue Analysis - 2005 Sampling. Final Report, City of Portland, Bureau of Environmental Services. http://www.portlandonline.com/shared/cfm/image.cfm?id=175911 Parametrix 1994. Columbia Slough Sediment Remedial Investigation/Feasibility Studies. Final Work Plan. Prepared for the City of Portland Bureau of Environmental Services. June 1994. Qian, S. and R. Lyons. 2006. Characterization of Background Concentrations of Contaminants Using a Mixture of Normal Distributions. Envir. Sci. Technol. 40: 6021-6025.
Tanner, DQ and KK Lee. 2004. Organochlorine Pesticides in the Johnson Creek Basin, Oregon, 1988-2002. Scientific Investigations Report. United States Geological Survey.
USEPA 1999. USEPA Contract Laboratory Program National Functional Guidelines for Organic Data Review. EPA-540-R-99-008 (OSWER 9240.1-05A-P). Office of Emergency and Remedial Response (OERR). October 1999.USEPA 2004. USEPA Contract Laboratory
City of Portland BES - 69 - Columbia Slough Sediment Data Analysis
Program National Functional Guidelines for Inorganic Data Review. EPA-540-R-04-004 (OSWER 9240.1-45). Office of Superfund Remediation and Technology Innovation (OSTRI). USEPA 2005. USEPA Contract Laboratory Program National Functional Guidelines for Superfund Organic Methods Data Review (Draft Final). EPA-540-R-04-009 (OSWER 9240.1-46). Office of Superfund Remediation and Technology Innovation (OSTRI). USEPA 2006. PCB ID - Congener-Specific PCB (Aroclor) Composition Data. Office of Prevention, Pesticides, and Toxic Substances (OPPTS) http://www.epa.gov/toxteam/pcbid/aroclor_comp_frame.htm
City of Portland BES - 70 - Columbia Slough Sediment Data Analysis
APPENDIX A
Sample Location Map
71 Columbia Slough Sediment Data Analysis
72 Columbia Slough Sediment Data Analysis
APPENDIX B
Columbia Slough Screening Levels
73 Columbia Slough Sediment Data Analysis
Columbia Slough Sediment Screening Levels - (2/15/06) Source Control Screening Level Sediment Screening Level (mg/kg-dry weight) (mg/kg) Concentration Basis Concentration Basis Metals Antimony 3 toxicity 3 toxicity Arsenic 7.9 background 8.4 baseline Cadmium 0.5 background 1.9 baseline Chromium 30 background 58 baseline Copper 12 background 54 baseline Lead 2 background 90 baseline Manganese 1100 toxicity 1100 toxicity Mercury (total, elemental) 0.2 toxicity/bioaccum 0.2 toxicity/bioaccum Selenium 0.1 bioaccum unknown Silver 4.5 toxicity 4.500 toxicity Nickel 20 background 34 baseline Zinc 53 background 314 baseline Polychlorinated Biphenyls Aroclor 1254 0.010 MRL 0.024 baseline Aroclor 1260 0.01 MRL 0.01 MRL Aroclor 1248 0.010 MRL 0.01 MRL Pesticides beta-BHC 0.001 MRL unknown gamma-BHC 0.0009 toxicity 0.00900 toxicity DDD 0.001 MRL 0.00610 baseline DDE 0.0010 MRL 0.00700 baseline DDT 0.001 MRL 0.00250 baseline Endosulfan 0.350 bioaccum 0.35035 bioaccum Endrin Aldehyde 0.003 toxicity 0.0030 toxicity Aldrin 0.0010 MRL 0.001000 MRL Chlordane 0.0010 MRL 0.00100 MRL Dieldrin 0.001 MRL 0.001000 MRL Heptachlor 0.00 MRL 0.001000 MRL
74 Columbia Slough Sediment Data Analysis
Semivolatile Organics 2-Methylnaphthalene 0.02 toxicity 0.02 toxicity Acenaphthene 0.29 toxicity 0 toxicity Acenaphthylene 0.16 toxicity 0.16 toxicity Anthracene 0.057 toxicity 0 toxicity Benzo(a)anthracene 0.032 toxicity 0.072 baseline Benzo(a)pyrene 0.032 toxicity 0.0900 baseline Benzo(b)fluoranthene NA 0.104 baseline Benzo(g,h,i)perylene 0.3 toxicity 0.3 toxicity Benzo(k)fluoranthene 0.03 toxicity 0.09 baseline Bis(2-ethylhexyl)phthalate 0.75 toxicity 1.816 baseline Chrysene 0.057 toxicity 0.1290 baseline Dibenzo(a,h)anthracene 0.06 toxicity 0.060 toxicity Dibenzofuran 5.1 toxicity 5 toxicity Fluoranthene 0.111 toxicity 0.1440 baseline Fluorene 0.077 toxicity 0.0770 toxicity Indeno(1,2,3-cd)pyrene 0.017 toxicity 0.074 baseline Naphthalene 0.176 toxicity 0.176 toxicity Phenanthrene 0.042 toxicity 0.08 baseline Phenol 0.048 toxicity 0.048 toxicity Pyrene 0.053 toxicity 0.1960 baseline Pentachlorophenol 0.100 bioaccum 0.10 bioaccum Organotins Tributyltin (pore water) 0.00006 toxicity 0.00006 toxicity Tributyltin (sediment) 0.190 bioaccum 0.19 bioaccum Petroleum Petroleum hydrocarbons 80 toxicity 80 toxicity Dioxins (2,3,7,8 TCDD TECs) 0.000009 toxicity 0.000009 toxicity
75 Columbia Slough Sediment Data Analysis
APPENDIX C
Data Usability Report
76 Columbia Slough Sediment Data Analysis
CITY OF PORTLAND BUREAU OF ENVIRONMENTAL SERVICES COLUMBIA SLOUGH LONG-TERM MONITORING PROGRAM
2006 SEDIMENT DATA DATA USABILITY REPORT
1.0 INTRODUCTION The City of Portland Bureau of Environmental Services (BES)/Oregon Department of Environmental Quality (DEQ) Long-Term Monitoring Plan for the Columbia Slough (BES 2007) involves collecting fish and sediment samples every 10 years, in order to assess status and trends. During the summer of 2006, BES personnel collected 78 discrete sediment samples and an additional 14 QC samples for analysis of contaminants of interest (COIs) and general analytical parameters. BES Investigation & Monitoring Services (IMS) conducted an independent data usability assessment to ensure the data are usable.
1.1 Abbreviations
Below is a list of abbreviations used in this report:
ARI Analytical Resources, Inc. BBP butylbenzylphthalate BDO Battelle Laboratories Duxbury Operations BEHP bis(2-ethylhexyl)phthalate BES City of Portland Bureau of Environmental Services CCV Continuing Calibration Verification CLP EPA Contract Laboratory Program COC Chain of Custody COI Contaminant of Interest CRQL Contract Required Quantitation Limit DDD 1,1-dichloro-2,2-bis(p-chlorophenyl)ethane DDE 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene DDT 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane DEP diethylphthalate DEQ Oregon Department of Environmental Quality DFR Daily Field Report DMP dimethylphthalate DBP di-n-butylphthalate
77 Columbia Slough Sediment Data Analysis
DnOP di-n-octylphthalate ECD electron capture detector FDS Field Data Sheet FO City of Portland BES Field Operations ICAL Initial Calibration ICC Initial Calibration Check IMS BES Investigations and Monitoring Services LCS Laboratory Control Spike MDL Method Detection Limit MRL Method Reporting Limit MS Matrix Spike MSD Matrix Spike Duplicate NFG EPA National Functional Guidelines NS&T NOAA Status & Trends PAHs Polyaromatic Hydrocarbons PCBs Polychlorinated Biphenyls PD Percent Difference QAPP Quality Assurance Project Plan QC Quality Control RPD Relative Percent Difference RSD Relative Standard Deviation SRM Standard Reference Material TOC Total Organic Carbon TPH-Dx Total Petroleum Hydrocarbons—Diesel Extended TPH-Gx Total Petroleum Hydrocarbons—Gasoline Extended TPH-HCID Total Petroleum Hydrocarbons—Hydrocarbon Identification WPCL City of Portland BES Water Pollution Control Laboratory
78 Columbia Slough Sediment Data Analysis
1.2 Sample Analyses
Analyses and laboratories utilized for this project are summarized in the following table:
Laboratory Analyses ARI Grain Size, TOC (one batch) Battelle (BDO) PAHs, phthalates, phenol, pesticides, PCB Aroclors BES WPCL Metals, TPH-HCID, TPH-Dx, total solids Field Measurements Eh, pH TestAmerica TOC, TPH-Gx
2.0 QAPP COMPLIANCE All data were evaluated using the project Quality Assurance Project Plan (QAPP) (BES 2006-- Draft) and U.S. EPA Contract Laboratory Program National Functional Guidelines (NFGs) for Data Review (EPA 1999, 2004, 2005) for guidance in evaluating the following: • Field practices, field quality control (QC) samples, daily activity logs, and sample collection logs; • Sample chain of custody (COC) and receipt documentation, preparation and analytical holding times, and reporting and detection limits for chemicals of interest; and • Laboratory data quality, in terms of precision, accuracy, representativeness, completeness, and comparability (PARCC).
Sample handling and laboratory practices are discussed in Section 3.0. Field practices are discussed below. There were no deviations from the QAPP except as noted.
Daily Activity Logs Daily activity logs consist of daily field reports (DFRs) and sample field data sheets (FDSs). DFRs and FDSs are used to record general and sample-specific information regarding site conditions, time of collecting, visual and physical sample characteristics, sampling difficulties, and any information relating to potential contaminant sources. These were reviewed by both the Field Operations project lead and by IMS for completeness and consistency. No problems were encountered except for minor issues with sample tracking. Two duplicate samples were mis-identified on the sample tracking sheet used to insure all samples were collected. One duplicate sample location identified was not a valid location and both entries were corrected using original sample information recorded on FDSs. Corrections were properly initialed by the individual making the correction.
Field QC Samples Field QC samples are used to assess sample collection procedures, environmental conditions during sample collection and shipment, and the adequacy of equipment decontamination. They are also used to estimate field precision and accuracy.
79 Columbia Slough Sediment Data Analysis
The field decon blanks were analyzed for most of the same parameters as sediment samples (excluding TPH, TOC, grain size, etc.). Results of field blank analyses are discussed individually below under the appropriate sections. The results of the field duplicate samples are used to assess overall sample precision. The relative percent difference (RPD) is used to estimate the combination of field (sampling) and laboratory (analytical) precision. For a few samples where PAH or pesticide RPDs exceeded acceptance criteria, either the parent sample or the field duplicate was also used by BDO for laboratory duplicate analysis, thus three values were reported for a single sample location. Where three values were available for a single sample location, relative standard deviation (RSD) was calculated to assess sample precision. Precision goals are presented in the QAPP. Field duplicate results are discussed below under the sections for individual analyses.
3.0 LABORATORY DATA QUALITY REVIEW General Ninety-two samples consisting of 78 sample locations, eight field duplicates, five field decontamination blanks, and one equipment blank were received at WPCL between July 25 and September 26, 2006. Laboratory sample receipt forms indicate that, except for one sample jar from location LS03, all sample containers arrived intact, and all container labels matched the COC documentation. Sediment from the broken sample jar was transferred to a new sample jar and subsequently utilized for grain size analysis. Samples for analysis by TestAmerica or ARI were transported by courier from WPCL to TestAmerica under COC conditions on the same or next business day that samples were received. Samples analyzed by BDO were shipped via FedEx overnight except as described below in Sections 3.2 and 3.3. All samples were analyzed using methods specified in the QAPP. BDO conducted the PCB Aroclor, organochlorine pesticide, and PAH analyses using NOAA Status and Trends (NS&T) analytical methods, rather than the EPA methods recommended in the QAPP. NS&T methods are standard analytical methods optimized for the analysis of fish tissues and marine sediments, and are comparable to standard EPA methods; therefore, data comparability was judged to not be affected. Each sample is assigned a unique sample id/tracking number by BES Field Operations (FO) that differs from the sample location id. Laboratory data reported by BDO is reported both by analytical batch number and the FO unique id. Laboratory QC issues are discussed below using batch numbers since some QC issues were only associated with specific batches. Field sample collection was not systematic for most areas of the Slough, therefore, most analytical batches include samples from various reaches of the Slough. Table 1 (attached) includes a list of analytical batches with the corresponding sample locations and FO sample ids. Various data were qualified by the labs as estimated using several qualifiers to illustrate various laboratory QC failures. Following review of laboratory reports, case narratives, and field QC data, these qualifiers were carried through and replaced with qualifiers presented below. Additional qualifiers were added, where necessary. Qualified data are still considered valid and usable, though some data should be used with caution. Qualifiers used for 2006 sediment sampling are listed below: J Estimated concentration JB Estimated due to blank contamination JH Estimated concentration, possible/probable high bias due to QC failure
80 Columbia Slough Sediment Data Analysis
JL Estimated concentration, possible/probable low bias due to QC failure M Estimated concentration, sample and duplicate are “out of control”. Potential false positive due to gross RPD failure, one duplicate result at or below the MDL, and QC failure. Data should be used with caution U Not detected above MDL UB Analyte detected in method blank. Sample result is concentration reported by laboratory but less than reported method blank concentration UJ Estimated below MDL or MRL, reporting limit may be inaccurate or imprecise UM Not detected above MDL; however, sample and duplicate “out of control”. Duplicate value is greater than 10x MDL for the non-detect sample result
Some QC issues identified would typically result in some data collected for this project being qualified with “R” for rejected if the data were intended for use in support of litigation or enforcement actions. However, data use here is primarily for guiding management decisions, and data were not rejected since most significant issues could be attributed to non- homogeneous samples and data could still be considered an estimate for a particular sample location. A detailed summary of all data review results by analysis group follows.
3.1 Metals Holding Time The holding time criterion for the analysis of preserved sediment samples (6 months) included in the QAPP was not exceeded. Blanks Method blanks were analyzed at the appropriate frequency specified in the QAPP. Copper was detected in one laboratory method blank at 2.73 mg/kg and nickel was detected at 2.04 and 2.63 mg/kg in two laboratory method blanks. These values were slightly above method reporting limits and sample values were all greater than 5 times the method blank concentrations; therefore, no qualifiers were assigned (EPA 2005). No other metals were detected in laboratory method blanks. Copper, nickel, and zinc were detected in the equipment blank at concentrations of 0.40, 0.38, and 5.35 ug/l, respectively. Metals were detected in four out of five field decontamination blanks and included copper (only one detect at 0.32 ug/l) and zinc (7.60; 1.85; 4.41; and 1.10 ug/l). Since metals sediment data is reported in mg/kg, metals detected in field and equipment blanks likely do not affect sediment data, therefore, no qualifiers were assigned. No other metals were detected in the equipment or field blanks. Matrix Spike/Matrix Spike Duplicates (MS/MSD) Matrix spike samples were analyzed at the appropriate frequency specified in the QAPP. Two of fourteen antimony MS recoveries (73.8 and 73.8%) and one zinc recovery (128%) were outside the 75-125% recovery range specified in the QAPP. All other criteria for these metals (i.e., SRM, MSD RPDs, CCV) were within QAPP-specified control limits; therefore, analytical accuracy was judged to be acceptable and no further action was taken. None of the MS/MSD RPDs exceeded the 20% limit specified in the QAPP.
81 Columbia Slough Sediment Data Analysis
Laboratory Control Samples (LCS) and/or Standard Reference Materials (SRMs) Commercially-obtained SRM samples were also analyzed at the appropriate frequency specified in the QAPP. One of the fifteen SRM recoveries for arsenic (79.6%) and one for nickel (129.5%) were outside the 80-120% recovery range specified in the QAPP. All other criteria for these metals (i.e., MS/MSD, CCV) were within QAPP-specified control limits; therefore, analytical accuracy was judged to be acceptable and no further action was taken. Duplicates Field duplicate samples were analyzed at the appropriate frequency specified in the QAPP. Several of the metals results for field duplicates did not meet the 20% RPD specific in the QAPP. A 35% RPD is sometimes used for metals duplicate soil samples (EPA 2004); however, most metals RPDs for this project were well within acceptance criteria. Results and actions taken are summarized below:
METALS FIELD DUPLICATE PRECISION FAILURES Sample Parameter RPD Comment LS10 Antimony 25% Both values less than 5x MRL, no action taken WS04 Antimony 31.1% Duplicate value slightly greater than 5x MRL, both values qualified with “J” WS10 Cadmium 20.9% Both values qualified with “J” US07 Antimony 60.5% Numerous RPD failures also for pesticides and PAH results indicating non- Arsenic 35.7% homogenous sample matrix despite sample homogenization in the field. Field Chromium 28.3% crew noted metal shards in sediment, and sample was more sandy/gravelly than Nickel 25.9% most slough sediment samples. Values qualified with “J”.
3.2 Pesticides and PCBs Sample Preservation and Holding Time All samples were preserved properly as specified in the QAPP except for 23 samples shipped to Battelle in Duxbury, Massachusetts. Fourteen sediment samples were received by Battelle at 7.4oC and an additional eight sediment samples and one field decontamination blank were received at 20.5oC (68.9oF) four days after being relinquished by WPCL due to a shipping delay. All sample containers were received intact by Battelle and did not appear to have leaked or been damaged. Other than during shipping as noted above, samples were properly preserved, either refrigerated or frozen. Due to the persistent nature of the target analytes, this was not anticipated to impact the quality of the data and no action was taken. Due to the shipping delay, one field decontamination blank was extracted one day out of hold time (7 days for extraction, 40 days for analysis). No analytes were detected in this sample, the equipment blank, or any of the other field decontamination blanks, therefore, all results for this sample were qualified with a “UJ” (EPA 2005). Sediment samples were received by Battelle and archived (i.e., frozen) until all samples collected under for this project had been received. All samples were analyzed within 365 days of collection (EPA 1995, 2003) and within 40 days of extraction. Blanks Method blanks were analyzed at the appropriate frequency specified in the QAPP. 4,4’-DDE was detected at 0.13 ug/kg in the method blank for Battelle batch 06-0308. All 18 sample
82 Columbia Slough Sediment Data Analysis results were greater than 5x the concentration detected in the blank, therefore, none of the sample results were qualified (EPA 2005). No analytes were detected above method detection limits in the equipment or field decontamination blanks. Surrogates All surrogate recoveries were within the control limits specified in the QAPP except for the following: • MSD surrogate recovery was reported at 297% for PCB152 in batch 06-0274 which was outside the laboratory control limits of 40-120%. The analyst noted a gas fluctuation which likely impacted the surrogate recovery for that sample. MS/MSD results and the RPD for Aroclor 1254 were within control limits, therefore, the sample data was not believed to be affected and no action was taken. • Several recoveries were outside the control limits (40-120%) specified by the lab but were within the control limits included in the QAPP (30-150%). No action was taken.
LCS LCS samples were analyzed at the appropriate QAPP-specified frequency. All recoveries were within BDO-specified control limits (40-120%) except for the following:
PESTICIDE LCS RECOVERY FAILURES Batch Analyte Recovery Comments 06-0263, heptachlor 134% (Water sample) Heptachlor was not detected in any samples in this batch 06-0276 129% and no action was taken. 06-0274 heptachlor 137% Exceeded laboratory control limits (40-120%) but was within control limits specified in the QAPP (30-150%--EPA 2005). No action taken. 06-0308 endrin 124% Endrin, heptachlor, alpha-BHC, beta-BHC, and gamma-BHC were not heptachlor detected in any of the samples in this batch and no action was taken 160% alpha-BHC 123% beta-BHC 128% gamma-BHC 138% methoxychlor 178% Methoxychlor recoveries were also high in the MS/MSD samples for this batch, thus, all methoxychlor detects were qualified with “JH” for estimated with possible high bias. Non-detects were not qualified. 4,4’-DDT 154% Battelle reported that 4,4’-DDT was over-responding in the bracketing CCV and values were qualified by Battelle with “Q” as the data may be biased high. These 4,4’-DDT values were subsequently qualified with “JH” as described below under Calibrations.
Matrix Spike/Matrix Spike Duplicates (MS/MSD) MS/MSD samples were analyzed at the appropriate QAPP-specified frequency. MS/MSD recoveries and RPDs were within the control limits for accuracy (% recovery varies by analyte) and precision (also varies by analyte) specified in the QAPP except as summarized in the table below.
83 Columbia Slough Sediment Data Analysis
Note that laboratory data typically is not qualified based only on MS/MSD results, MS/MSD failures must be evaluated in conjunction with other laboratory QC data. Calibrations Calibrations were not specified in the QAPP; however, initial calibration (ICAL), initial calibration checks (ICC), and continuing calibration verification (CCV) were conducted at the appropriate frequency (EPA 2005). The following exceedences were reported by Battelle: • Heptachlor was high in two ICCs (batches 06-0274 and 06-0308); however, heptachlor was not detected in any samples included in these batches and no qualifiers were assigned. • Numerous CCV exceedences were reported for 2,4’-DDT, 4,4’-DDT, 4,4’-DDD in various analytical batches. Data that were bracketed by one or more failing CCVs were qualified by the lab as estimated. As reported by Battelle, some data may have either a high or low bias and are qualified with “J” or “UJ”, as appropriate. For specific batches, 4,4’-DDT over- responded and these data were qualified with “JH”. • For batches 06-0352 and 06-0367, Battelle reported that CCVs failed calibration criteria for aldrin and gamma-chlordane and that all results for these batches should be considered biased high. Data above the MDL for these batches have subsequently been qualified with “JH”.
PESTICIDE MS/MSD RECOVERY & PRECISION FAILURES Batch Analyte Recovery/ Comments Precision 06-0274 heptachlor MS/MSD Heptachlor was not detected in any samples in this batch and no action was taken. 195%/198 % 4,4’-DDD MSD 129% Exceeded laboratory control limits (40-120%) but was within control limits for 4,4’-DDT specified in the QAPP (23-134%--EPA 2005). Also, sample 4,4’-DDE MSD 124% spiked at less than 5x sample concentration, thus, no action taken. 06-0308 4,4’-DDT MS/MSD Battelle also reported that CCVs for this batch exceeded acceptance criteria and qualified selected data with “Q”. Battelle reported that 186%/164 CCVs for this batch were mostly over-responding and that 4,4’-DDT % sample data for this batch may be biased high. 4,4’-DDT and methoxychlor data that were bracketed by one or more failing CCVs methoxychlor MS/MSD were qualified by Battelle with “Q” and were subsequently qualified 193%/168 here with a “JH” to represent potential high bias. % 06-0367 2,4’-DDD RPDs 4,4’-DDD and 4,4’-DDE also exceeded RPDs, however, spike amounts 38.4% were less than 5x the sample concentrations, therefore, no action was 4,4’-DDD taken. Battelle attributed the QC failure to possible partial loss of ECD 52.0% 4,4’-DDE extract for the MSD sample since surrogate recoveries were in different ranges (47 - 74% and 68 – 102%) for the MS/MSD samples 55.1% 4,4’-DDT and PAH surrogate recoveries were more consistent for the two 40.0% samples. All other criteria were met (surrogate recoveries, LCS, aldrin replicates) except for several CCV results. No qualifiers were 31.0% assigned except for those analytes that were bracketed by one or a-chlordane more CCV failures as described below under “Calibrations”. 36.4% g-chlordane 35.4% dieldrin 38.2% heptachlor 31.8%
84 Columbia Slough Sediment Data Analysis
Duplicates Field duplicates were collected at the QAPP-specified frequency of one duplicate per ten samples. Where RPDs were outside control limits, other data from the corresponding analytical batch was evaluated to determine if issues appeared to affect the entire analytical batch. RPDs for analytes using other analytical methods were also reviewed to assess whether or not RPD failures simply represented non-homogeneous sample matrices. Where significant control limit failures occurred, such as where RPDs exceeded 120%, CCVs over-responded, and one sample was below the MDL and the other was greater than 10x the MDL, concentrations above the MDL were typically qualified with “M” for estimated, potential false positive. Field duplicate RPDs for all PCBs and pesticides were within control limits (30%) except as follows:
PESTICIDES FIELD DUPLICATE PRECISION FAILURES Sample Analyte RPD Comment LS04* 2,4’-DDD 125% 2,4’-DDD values for the sample and field duplicate were <0.03 ug/kg and 0.13 ug/kg. These values were qualified as “UJ” and “J” since the detected concentration was less than 5x the MDL and no other QC issues were associated with 2,4’-DDD for this batch. 4,4’-DDD 91.9% 4,4’-DDD and 4,4’-DDE RPDs were greater than 30% but less than 120%, thus these sample values were qualified with “J”. 4,4’-DDE 43.5% 4,4’-DDT 181% 4,4’-DDT and methoxychlor both were less than the MDL but detected in the duplicate at greater than 10x the MDL. Both these analytes had high methoxychlor 170% surrogate recoveries and over-responding CCVs for this batch, thus detects in the duplicate for these analytes may be false positives and/or biased high. Detects for these analytes in sample LS04 and its duplicate appeared to be inconsistent when compared to Lower Slough area data and ratios of these analytes and similar analytes (total DDT and chlordane-like compounds) in other samples. Detects were qualified as “M” for estimated, potential false positive, and non-detects were subsequently qualified as “UM”.
LS17 aldrin 40.0% A laboratory replicate for this location was also analyzed, therefore, for these analytes, the relative standard deviation (RSD) was calculated. RSDs were g-chlordane 39.1% less than 30% except for 4,4’-DDD (37.8%) which was subsequently qualified 4,4’-DDD 38.1% with “JL” for the low sample value, “J” for the middle value, and “JH” for the high sample value. Also see discussion below for laboratory replicate samples. * Some of these data had already been qualified due to laboratory QC failures
For all other field duplicates, data with RPDs between 50 and 120% were qualified with “J” if not already qualified. RPDs ranged from 160-200% for all results where the analyte was detected in one sample but the analyte was below the MDL in the duplicate. As noted in the table above, these data were all qualified with “M” for detects and “UM” for non-detects. None of the field duplicate precision failures appeared to represent systemic problems that would require additional qualifiers beyond those applied through review of other laboratory QC as detailed previously in this section. In addition to MSDs, Battelle reported results of laboratory replicate (i.e., duplicate) samples which were analyzed with each analytical batch. RPDs for all replicate samples were within control limits except as following:
85 Columbia Slough Sediment Data Analysis
• For batch 06-0274, a replicate of sample LS17 was analyzed and 2,4’-DDD (44.6%), 4,4’-DDD (37.9%), and dieldrin (56%) exceeded RPD control limits. A field duplicate for this location was also analyzed and RPDs for two additional analytes were outside control limits, therefore, for analytes where either the replicate or field duplicate RPDs exceeded control limits, the relative standard deviation (RSD) was calculated. RSDs were less than 30% except for 4,4’-DDD (37.8%) which was subsequently qualified with “JL” for the low sample value, “J” for the middle value, and “JH” for the high sample value. • For batch 06-0352, a replicate of sample BFC04 was analyzed and 4,4’-DDT (61.7%) exceeded control limits. Since the field duplicate for this batch also exceeded control limits for 4,4’-DDT (“M” qualified), this data was qualified with “J”.
3.3 PAHs, Phthalates, and Phenol Sample Preservation and Holding Time All samples were preserved properly as specified in the QAPP except for 23 samples shipped to Battelle in Duxbury, Massachusetts. Fourteen sediment samples were received by Battelle at 7.4oC and an additional eight sediment samples and one field decontamination blank were received at 20.5oC (68.9oF) four days after being relinquished by WPCL due to a shipping delay. All sample containers were received intact by Battelle and did not appear to have leaked or been damaged. Other than during shipping, samples were properly preserved, either refrigerated or frozen. The fourteen samples received at 7.4oC were not anticipated to be affected since the temperature exceedence was slight and of short duration. The samples received at 20.5oC may been impacted though most analytes would not likely be affected due to their persistent nature and low vapor pressure. However, results for some of the more volatile analytes (i.e., vapor pressure greater than 0.01 mm/Hg) in these samples may be biased low. Subsequently, most data were judged not to have been affected, however, data above the MDL for 2-chloronaphthalene, acenaphthene, naphthalene, phenanthrene, and phenol were qualified with “JL” if not already qualified due to other QC failures such as blank contamination. Non-detect 2-chloronaphthalene data were qualified with “UJ” (and not rejected) as this analyte was rarely detected and only seen in other samples when elevated concentrations of other PAHs were detected. Samples affected were LS01 – LS05, LS04 field duplicate, NS01, and NS02. Due to the shipping delay, one field decontamination blank was extracted one day out of hold time (7 days for extraction, 40 days for analysis). Some analytes were detected though results were qualified due to method blank contamination as described below. All samples were analyzed within 365 days of collection (the holding time for frozen samples) and within 40 days of extraction. Blanks Method blanks were analyzed at the appropriate frequency specified in the QAPP. No analytes were detected in method blanks other than the following:
86 Columbia Slough Sediment Data Analysis
SEDIMENT SEMI-VOLATILE METHOD BLANK DETECTIONS Batch MRL > Concentration > MDL Concentration > MRL 06-0274 chrysene, naphthalene, pyrene, BEHP, DEP, phenol DBP 06-0307 phenanthrene, fluoranthene, BEHP, DEP, DBP, phenol pyrene, naphthalene, chrysene, BBP 06-0308 fluoranthene, naphthalene, BEHP, DEP, DBP, phenol pyrene 06-0352 naphthalene, DEP, DBP BEHP, phenol 06-0352* naphthalene, DEP, DBP BEHP, BBP, phenol 06-0367 fluoranthene, naphthalene, DBP, phenol pyrene, DEP, BEHP BBP = butyl benzyl phthalate BEHP = Bis(2-ethylhexyl)phthalate DBP = Di-n-butylphthalate DEP = Diethylphthalate * = blank is for re-extract of two samples from 06-0352 batch
EPA CLP guidance (EPA 2005) requires that all analytes be below the CRQL in a method or procedural blank except for BEHP which may be present at up to 5x the CRQL. EPA (1999) also allows the five other phthalates (BBP, DnBP, DEP, DMP, and DnOP) typically reported to be present in blanks at up to 5x the CRQL. Corresponding sample results for phthalates and phenol were qualified as follows (EPA 2005): • Non-detected results were not qualified; • Concentrations greater than the MDL but less than the blank concentration were qualified with “UB” at the reported sample concentration (usually reported sample concentrations are qualified with “UB” and replaced with the value reported for the method blank. Sample concentrations were not changed since, as discussed below, laboratory contamination appeared to be at random concentrations rather than a consistent, uniform source of contamination). • Concentrations less than 5x the blank concentration were qualified with “JB”. • Concentrations greater than 5x the blank concentration were not qualified.
In general, PAHs and BEHP were commonly detected in samples at concentrations much higher than concentrations in method blanks and few data were qualified. However, much of the remaining phthalate and phenol data is qualified due to blank contamination except for elevated concentrations reported for certain samples. Numerous PAHs and phthalates were detected in field and equipment blanks; however, these analytes were also typically detected in laboratory method blanks. Analytes detected in water sample field and lab blanks were generally the same as those detected in method blanks for sediment samples. PAHs detected were all less than 0.01 ug/l and most less than 0.001 ug/l, therefore, PAHs detected in blanks were not anticipated to impact sample analytical data. Phthalates and phenol were detected at higher concentrations in field and method blanks at concentrations ranging from 0.001 ug/l to 1 ug/l. A pattern was not observed as method blank concentrations were sometimes higher than field blank concentrations and vice versa. Battelle conducted additional glassware cleaning and performed additional extractions and analyses;
87 Columbia Slough Sediment Data Analysis however, they were unable to isolate or completely correct the lab contamination issues. Due to the apparently random concentrations detected for individual analytes throughout the project, particularly phenol and phthalates, it must be assumed that some unqualified data values may be a result of blank contamination; however, except for phenol, elevated sample concentrations are likely representative of sediment conditions. Due to the somewhat random nature of laboratory blank contamination for water samples, it is unclear if analytes detected in field blanks were introduced as part of sample collection or in the laboratory. Of all the analytes detected, DnOP (0.274 ug/l) in the first field blank collected appeared to have been introduced through sample collection or sampling equipment as there were no correspondingly elevated DnOP concentrations in any of the laboratory QC data. The next two field blanks collected yielded DnOP results of 0.0172 and 0.00453 ug/l. Sediment sample data collected between the collection of the first two field blanks did not appear to have been compromised since there were many non-detect results and a rough correlation could be observed in many samples between DnOP and other phthalate data. Since much of the lower- concentration sediment data for the same analytes had already been qualified due to sediment method blank contamination, no further action was taken.
Surrogates All surrogate recoveries were within the control limits specified in the QAPP (varies by analyte) except for the following: • Phenol-d6 recoveries for samples BS01, LS01, LS02, LS17, NS01, WS10DUP, WS11, WS14 and the MS for batch 06-0307 were low and Battelle indicated that these values may be biased low. These phenol sample values had already been qualified due to blank contamination. • Napthalene-d8 recoveries were below acceptance criteria in batch 06-0307 in the method blank and the matrix spike. The naphthalene loss was attributed by the lab to slightly elevated temperature during the blow-down process and recoveries in the associated samples were all within acceptance criteria, thus, no action was taken. • Benzo(a)pyrene-d12 recovery was below acceptance limits for sample WAS03. All other recoveries for this sample (except for phenol) were acceptable, therefore, the associated target compounds (benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, indeno(1,2,2-cd)pyrene, dibenzo(a,h)anthracene, and benzo(g,h,i)perylene) for this sample were qualified with “JL” (EPA 2005). • Several PAH recoveries for method blanks and MSs were outside the control limits (40- 120%) but recoveries in associated QC and field samples were within control limits and no action was taken.
LCS LCS samples were analyzed at the appropriate QAPP-specified frequency. Additionally, all LCS recoveries were within the control limits for accuracy (40-120%) specified by the lab except for DnOP (33%) in the LCS for batch 06-0367. All other QC criteria were within control limits for this analyte for this batch and the recovery was within control limits specified in the QAPP (30-150%), therefore, no action was taken. Matrix Spike/Matrix Spike Duplicates (MS/MSD) MS/MSD samples were analyzed at the appropriate QAPP-specified frequency (BES 2006). MS/MSD recoveries and RPDs were within the control limits for accuracy (% recovery varies by analyte) and precision (also varies by analyte) specified in the QAPP except as follows:
88 Columbia Slough Sediment Data Analysis
• MS/MSD recoveries and RPDs for some analytes in batch 06-0274, 06-0307, 06-0308 were outside the control limits (50-120% recovery, 30% RPD) specified by the laboratory. However, the samples were spiked at less than 5x the concentrations in the sample and accuracy was demonstrated through the LCS, therefore, no qualifiers were assigned. Also, many of the associated sample values had already been qualified due to blank contamination. • For batch 06-0307 MS recoveries were low for 2-chloronaphthalene (37%), naphthalene (1%), and phenol (8%) and RPDs were outside control limits for 2-chloronaphthalene (65.5%), naphthalene (197%), DEP (31.7%), and phenol (150%). For batch 06-0308, recoveries were low for DEP (39%) and phenol (39%) and the RPD for DMP (31.7%) was outside acceptance criteria. The following actions were subsequently taken: a) All 2-chloronaphthalene samples for batch 06-0307 were below the MDL and were qualified with “UJ” since this compound was detected in only a few samples at low concentrations for this project (EPA 2005). b) Naphthalene data were not qualified as described above under “Surrogates”. c) All DEP and phenol data for these two batches had already been qualified due to blank contamination. d) For DMP in batch 06-0308, the RPD was only slightly outside control limits and accuracy was demonstrated through LCS and field duplicate results. No action was taken.
Duplicates Field duplicates were collected at the QAPP-specified frequency of one duplicate per ten samples. Where RPDs were outside control limits, other data from the corresponding analytical batch was evaluated to determine if issues appeared to affect the entire batch. RPDs for analytes using other analytical methods were also reviewed to assess whether or not RPD failures simply represented non-homogeneous sample matrices. Where significant control limit failures occurred, such as where RPDs exceeded 120%, CCVs over-responded, and/or one sample was below the MDL and the other was greater than 10x the MDL, concentrations above the MDL were typically qualified with “M” for estimated, potential false positive. Field duplicate RPDs for all PAHs and phthalates were within control limits (50% as specified in the QAPP) except as follows: • For sample LS17 RPD failures, refer to the discussion below of laboratory replicate (i.e., duplicate) sample results. • Numerous RPD failures occurred for sample US07 and its’ duplicate indicating non- homogenous sample matrix despite sample homogenization in the field. Field crew noted metal shards in sediment, and sample was more gravelly (10.5% and 5.9% gravel for the sample and duplicate) than typical slough sediment samples. Values were qualified with “J” for detects and “UJ” for non-detects (one analyte). For DMP, the duplicate result (954 ug/kg) was an order of magnitude greater than both the parent sample result (RPD = 172%) and all other sample results for this project. Since this sample was clearly non-homogeneous, the duplicate DMP value was qualified with “JH”. • A few gross RPD failures occurred where an analyte was not detected in one sample above the MDL but detected in the duplicate or parent sample at greater than 10x the MDL. Detects were qualified as “M” for estimated, potential false positive, and non- detects were subsequently qualified as “UM”, except if the data had already been qualified due to blank contamination. • Some data were already qualified due to blank contamination. No additional action was taken for these data.
89 Columbia Slough Sediment Data Analysis
• For all other field duplicates, data that was not within control limits was qualified with “J” if not already qualified.
In addition to MSDs, Battelle reported results of laboratory replicate (i.e., duplicate) samples which were analyzed with each analytical batch. RPDs for all replicate samples were within control limits except as following: • For batch 06-0274, a replicate of sample LS17 was analyzed and nine analytes exceeded the laboratory’s RPD control limits (30%). A field duplicate for this location was also analyzed, therefore, for these analytes, the relative standard deviation (RSD) was calculated. RSDs were less than 50% (specified in the QAPP) except for acenaphthene, DEP, and DMP. In addition, replicate sample analysis resulted in BBP and phenol exceeding RSD control limits (where the RPD for the sample and field duplicate did not exceed control limits). For RSDs greater than 50%, if not already qualified, sample values were qualified with “JL” for the low sample value, “J” for the middle value, and “JH” for the high sample value. • For batch 06-0352, a replicate of sample BFC04 was analyzed and BBP (167.9%) exceeded control limits. Since sample locations BFC04 and LS17 had gross RPD/RSD exceedences for BBP, precision failures were reviewed with all other QC data for this compound to determine if there were systemic laboratory issues with BBP reporting. No significant issues were identified and the few method blank detections for BBP were only slightly above MDLs, therefore, precision failures were judged to be likely due to non- homogeneous matrices. BBP uses are such that BBP could be present as a component of individual sediment particles such as plastics, paint, or other waste material. Since the BFC04 sample BBP value was significantly lower than the replicate value, it was qualified with “JL” as it may represent a low estimate for BBP at that sample location. • Several RPDs exceeded control limits; however, analytical results were less than 5x the MDLs or the analytes were detected in method blank and corresponding sample data had already been qualified, thus no action was taken.
None of the duplicate precision failures appeared to represent systemic problems that would require additional qualifiers beyond those applied through review of other laboratory QC as detailed previously in this section. Calibrations Calibrations were not specified in the QAPP; however, initial calibration (ICAL), initial calibration checks (ICC), and continuing calibration verification (CCV) were conducted at the appropriate frequency (EPA 2005). Several CCV PD exceedences were reported due to under-recoveries for DEHP (one exceedence) and DnOP (two exceedences). The execeedences were slightly over the control limit (25%) specified by the lab but were less than the control limit specified by EPA (40%) for “semi-volatile target compounds exhibiting poor response” (EPA 2005), therefore, no action was taken. Two slight exceedences were reported for phenol; however, associated sample data had already been qualified due to blank contamination.
3.4 Total Petroleum Hydrocarbons Sample Preservation and Holding Time All samples were preserved properly as specified in the QAPP. All samples were extracted within 14 days of collection and analyzed within 40 days of extraction.
90 Columbia Slough Sediment Data Analysis
Blanks Method blanks were analyzed at the appropriate frequency specified in the QAPP. No analytes were detected in method blanks. Surrogates All surrogate recoveries were within the control limits (50-150%) specified in the QAPP. LCS LCS samples were analyzed at the appropriate QAPP-specified frequency and all recoveries were within the control limits for accuracy (80-120%) specified in the QAPP. Calibrations All calibration check standards were within control limits (85-115%). Duplicates Field duplicates were collected at the QAPP-specified frequency of one duplicate per ten samples. TPH-HCID results for the eight parent samples and duplicates were consistent. None of the TPH-Dx field duplicate RPDs exceeded the control limits (50%) specified in the QAPP. TPH-Dx laboratory replicate (duplicates and triplicates) samples were analyzed and reported by WPCL. RPDs for four of the ten laboratory duplicates were outside control limits (50%) specified in the QAPP. Two of the RPD exceedences were for values less than 5x the reporting limit and all RSDs were within control limits (50%), therefore, no action was taken.
3.5 Total Organic Carbon Holding Time The holding time criteria specified in the QAPP for the extraction and analysis of sediment samples (28 days) was not exceeded for any samples. Blanks Method blanks were analyzed at the appropriate frequency specified in the QAPP. No analytes were detected above method detection limits in the equipment. LCS LCS samples were analyzed at the appropriate QAPP-specified frequency, and all LCS recoveries were within the control limits for accuracy (90-110%) One sample was submitted to a separate laboratory (ARI) and the LCS recovery was 86.9% which is outside the control limits specified by TA. However, this is still within the control limits specified in the QAPP (80-120%). Duplicates Field duplicates were collected at the QAPP-specified frequency of one duplicate per ten samples. RPDs for TOC exceeded the 20% limit specified in the QAPP for four of the eight field duplicate samples. This was attributed to non-homogenous sample matrices and data were qualified with “J” except for sample LS04 which was qualified with “UJ”.
4.0 DATA USABILITY Appropriate methods were used for all analyses, ensuring good comparability with other data. Analytical accuracy and precision were determined to be generally acceptable, with noted exceptions. Some phthalate and phenol data were changed to non-detects (qualified with “UB”)
91 Columbia Slough Sediment Data Analysis based on potential effects from blank contamination. Qualifiers were assigned based on other analytical QC results that exceeded project data quality criteria. No data were rejected based on this review; therefore, the project completeness goal of 95% was met. All data reported, except for phenol and some low-level phthalate data, should be considered valid as qualified, representative of the samples collected, and acceptable for further use.
All phenol data collected for this project should be considered suspect due to blank contamination. The highest phenol concentration measured for this project was 1,318 ug/kg. Phenol was detected in method blanks as high as 550 ug/kg; however, the phenol-d6 surrogate recovery for this blank was 8%. Therefore, if this method blank were surrogate corrected, it would indicate that phenol method blank contamination may have been present at concentrations higher than any reported sample concentration. Subsequently, all phenol data must be considered suspect, even if unqualified.
5.0 REFERENCES
City of Portland Bureau of Environmental Services (BES). 2006. Draft Quality Assurance Project Plan for the Sediment Sampling in the Columbia Slough. Prepared for Oregon DEQ. August 2006.
City of Portland Bureau of Environmental Services (BES). 2007. Long-Term Monitoring Plan for the Columbia Slough. Prepared for Oregon Department of Environmental Quality (DEQ). March 2007.
EPA 1995. QA/QC Guidance for Sampling and Analysis of Sediments, Water, and Tissues for Dredged Material Evaluations. EPA 823-B-95-001.
EPA 1999. USEPA Contract Laboratory Program National Functional Guidelines for Organic Data Review. EPA-540-R-99-008 (OSWER 9240.1-05A-P). Office of Emergency and Remedial Response (OERR).
EPA 2002. Guidance on Environmental Data Verification and Data Validation. EPA-240-R-02- 004 (EPA QA/G-8). Office of Environmental Information. November 2002.
EPA 2003. Analytical Methods (Table 8-2. Summary of Recommended Procedures for Sample Collection, Preservation, and Storage). U.S. Environmental Protection Agency Water Science. http://www.epa.gov/waterscience/itm/ITM/table8-2.htm.
EPA. 2004. USEPA Contract Laboratory Program National Functional Guidelines for Inorganic Data Review. EPA-540-R-04-004 (OSWER 9240.1-45). Office of Superfund Remediation and Technology Innovation (OSTRI).
EPA. 2005. USEPA Contract Laboratory Program National Functional Guidelines for Superfund Organic Methods Data Review (Draft Final). EPA-540-R-04-009 (OSWER 9240.1-46). Office of Superfund Remediation and Technology Innovation (OSTRI).
92 Columbia Slough Sediment Data Analysis
SEDIMENT ANALYTICAL BATCH CROSS REFERENCE 06-0274 06-0307 06-0308 06-0352 06-0367 Location Sample ID Location Sample ID Location Sample ID Location Sample ID Location Sample ID LS06 FO 060863 BS02 FO 060901 BS01 FO 060992 BFC01 FO 061207 BFC02 FO 061157 LS07 FO 060862 BS03 FO 060900 MS01 FO 060990 BFC03 FO 061208 BFC06 FO 061145 LS08 FO 060861 LS01 FO 060885 MS02 FO 060991 BFC04 FO 061209 MS14 FO 061133 LS09 FO 060860 LS02 FO 060884 MS03 FO 060993 BFC05 FO 061215 MS15 FO 061134 LS10 FO 060859 LS03 FO 060883 MS10 FO 060956 MS06 FO 061230 MS15DUP FO 061135 LS10DUP FO 060864 LS04 FO 060878 MS11 FO 060957 NS03 FO 061229 US01 FO 061186 LS11 FO 060858 LS04DUP FO 060879 MS12 FO 060979 WAS01 FO 061257 US02 FO 061187 LS12 FO 060856 LS05 FO 060877 MS13 FO 060980 WAS02 FO 061258 US03 FO 061185 LS13 FO 060855 MS04 FO 060910 PDC01 FO 060955 WAS03 FO 061240 US04 FO 061146 LS14 FO 060854 MS05 FO 060911 PDC02 FO 060954 WS05 FO 061226 US05 FO 061147 LS15 FO 060853 MS07 FO 060940 PDC03 FO 060953 WS05DUP FO 061228 US06 FO 061148 LS16 FO 060852 MS08 FO 060941 WS09 FO 060975 WS06 FO 061239 US07 FO 061167 LS17 FO 060851 MS09 FO 060935 WS10 FO 060976 WS07 FO 061216 US07DUP FO 061168 LS17DUP FO 060857 NS01 FO 060882 WS10DUP FO 060978 WS08 FO 061227 US08 FO 061166 LS18 FO 060850 NS02 FO 060881 WS11 FO 060977 LS19 FO 060849 WS01 FO 060931 WS12 FO 061042 LS20 FO 060848 WS02 FO 060932 WS13 FO 061111 LS21 FO 060847 WS03 FO 060933 WS14 FO 061112 LS22 FO 060846 WS04 FO 060934 LS23 FO 060845 WS04DUP FO 060936
93 Columbia Slough Sediment Data Analysis
APPENDIX D
2006 Columbia Slough Data Tables
94 Columbia Slough Sediment Data Analysis
95 Columbia Slough Sediment Data Analysis
96 Columbia Slough Sediment Data Analysis
97 Columbia Slough Sediment Data Analysis
98 Columbia Slough Sediment Data Analysis
99