W912DW20B0005R0001- Final SDM Kenmore 3 Dec 2019.pdf

December 3, 2019 Prepared by:

Dredged Material Management Office Seattle District, U.S. Army Corps of Engineers

MEMORANDUM FOR: RECORD

SUBJECT: DETERMINATION REGARDING THE SUITABILITY OF PROPOSED DREDGED

MATERIAL FROM THE KENMORE FEDERAL NAVIGATION CHANNEL EVALUATED UNDER

SECTION 404 OF THE CLEAN WATER ACT FOR UNCONFINED OPEN-WATER DISPOSAL AT THE

ELLIOTT BAY NON-DISPERSIVE DISPOSAL SITE.

1. Introduction. This memorandum reflects the consensus determination of the Dredged Material Management Program (DMMP) agencies (U.S. Army Corps of Engineers, Washington State Department of Ecology, Washington State Department of Natural Resources, and the Environmental Protection Agency) regarding the suitability/unsuitability of up to 34,350 cubic yards

(cy) of dredged material from the Kenmore federal navigation channel for open-water disposal at the Elliott Bay non-dispersive site.

2. Background. The authorized federal navigation channel in Kenmore, WA runs approximately 2,900 feet from deep water in Lake Washington to the Kenmore industrial area, see Figure 1. The channel has an authorized width of 100 to 120 feet wide and a depth of -15 ft Lake Washington Low Water (LWLW) datum.

The channel was last dredged in 1998. Since then sediment has accumulated in the navigation channel above the authorized depth, resulting in impacts to the ability of vessels to use the navigation channel. In 2014 a DMMP characterization was completed – the results of which were inconclusive due to the presence of dioxin at levels above the disposal site management objective of 4 pptr TEQ throughout the channel, with concentrations as high as 23 pptr TEQ. The DMMP agencies concluded in 2014 that additional testing, including a bioaccumulation evaluation for dioxins would be necessary to make a final determination on the suitability of the proposed dredged material for open-water disposal at Elliott Bay.

This memo documents the additional sampling and testing that was conducted in 2019 to fully evaluate the sediment in the Kenmore navigation channel, including bioaccumulation testing for dioxin. Table 1 includes project summary and tracking information.

Table 1. Project Summary Project ranking High Proposed dredging volume 34,350 cy Proposed dredging depth -15 ft MLLW (+ 2 ft overdepth) 1st draft SAP received 4 December 2018

Elliott Bay: 30 November 2018 Comments provided on 1st draft SAP 13 December 2018

Kenmore Navigation Channel DMMP Suitability Determination

December 3, 2019 Elliott Bay: 13 December 2018

2nd draft SAP received 18 December 2018, both SAPs Comments provided on 2nd draft SAP 8 January 2019 both SAPs Final SAPs received 18 January 2019 SAPs approved 20 January 2019 Sampling dates 21 – 30 January, 18 February, 7

March and 13 March 2019 Draft data report received 5 September 2019 Comments provided on draft report 18 September 2019 2nd draft data report received 3 October 2019 Comments provided on 2nd draft report 16 October 2019 Final data report received 31 October 2019 EIM Study ID KENMO19 Recency Expiration (high rank = 3 years) No expiration – all unsuitable based on incomplete bioassay results.

Additional biological testing could alter suitability and establish recency expiration date.

3. Project Ranking, Sampling Requirements, and Bioaccumulation Study Design. This project was ranked “high” by the DMMP agencies according to the guidelines set out in the DMMP User Manual for Kenmore, WA. In a high-ranked area the number of samples and analyses are calculated using the following guidelines (DMMP, 2014):

• Maximum volume of sediment represented by each field sample = 4,000 cubic yards

• Maximum volume of sediment represented by each analysis in the upper 4-feet of the dredging prism (surface sediment) = 4,000 cubic yards

• Maximum volume of sediment represented by each analysis in the subsurface portion of the dredging prism = 12,000 cubic yards

Based on calculated volumes plus additional contingency, 9 DMMUs were needed to characterize the total project volume of 34,350 CY. DMMUs were labeled KEN01 for DMMU1, KEN02 for DMMU 2 and so on. Table 2 shows calculated volumes for each DMMU and Figure 2 locates the DMMUs within the project area. The study design included the following elements:

• The boundaries of DMMU 2 were kept consistent with the previous testing because that had been the DMMU with highest dioxin concentrations,

• A full resampling and re-testing of the dredge prism was planned, since in addition to dioxins, there had been issues with other COCs, notably pesticides and phthalates, in the 2014 sampling event. Since the recency had expired on the 2014 sampling, new data was required.

• Based on the project volume, two bioaccumulation tests were considered sufficient to evaluate bioaccumulation impacts from the project

• The single DMMU with the highest concentration of dioxin would be analyzed for bioaccumulation, or, alternatively a composite of multiple DMMUs would be analyzed if there were more than one DMMU with dioxin concentrations elevated above the rest of the DMMUs.

December 3, 2019

• All the remaining DMMUs not included in the higher end bioaccumulation test would be composited together and analyzed for bioaccumulation as a single ‘Supercomposite’.

• Since there was no way to know which DMMU would have highest dioxin concentration, all

DMMUs were sampled so they would have enough volume to be tested for bioaccumulation individually. Therefore, 5 cores were needed from each DMMU to collect sufficient volume.

• Assumed no need to re-evaluate z-samples for anti-degradation compliance, since last time freshwater z-sample bioassays were run on all DMMUs and all passed. The DMMP agencies agreed there was no reason to suspect the quality of deep buried sediment would have changed since the 2014 testing. However, as a precautionary measure, z-samples were collected and archived.

Standard Tier 3 DMMP bioaccumulation testing utilizes a modified EPA protocol (Lee et al. 1989), which included co-testing of Macoma nasuta and Nephtys caecoides in the same test chambers.

The modified EPA protocol – with static-renewal of overlying water, sediment addition, and no supplemental feeding – was selected for use in this study.

The length of exposure is a critical parameter in the design of the bioaccumulation test. The longer the exposure, the closer tissue concentrations approach a steady-state condition. The DMMP Manual does not provide a recommended exposure time specifically for dioxins, however, based on the previous testing results, existing guidance (DMMP, 2009), and experience from other projects, a 45-day exposure time was chosen.

Target Tissue Levels (TTLs) are used to compare tissue concentrations resulting from test sediments to determine suitability for open-water disposal. The 2018 DMMP User Manual includes a list of TTLs (Table 10-6), but does not include one for dioxins. Therefore it was agreed to use a similar comparison approach as was used for the Everett Marina study (DMMP, 2017), including:

a. Comparison of test sediment tissue concentrations to a reference sediment tissue concentration by use of a one-tailed t-test with an alpha level of 0.1.

b. Determination of whether there was a 20% difference between reference and test sediment tissue concentrations.

c. Evaluation of the importance of non-detect concentrations to the total TEQ calculations

d. Comparison to tissue PQLs as developed by Ecology (Ecology, 2018).

e. Evaluation of tissue concentrations from organisms living in the vicinity of the disposal site.

In consideration of the general lack of in-situ tissue data from the Elliott Bay disposal site vicinity, a separate laboratory bioaccumulation study was conducted during the same time frame to characterize dioxin concentrations in benthic tissues. Specifically, sediment was collected from 16 areas of Elliott Bay which were similar in depth and sediment type to the disposal site but outside the influence of disposal activities at the Elliott Bay site. These locations would be impacted by diffuse sources of contaminants and be consistent, in concept, with regional background. The surface sediment samples were mixed into a single composite sample (Elliott Bay Benchmark) and tested using the same bioaccumulation protocol used to test Kenmore sediments (Sealaska, 2019d).

December 3, 2019

4. Sampling. Sampling took place 21-30 January 2019 for sampling in the Kenmore navigation channel using a vibracore sampler aboard the M/V Nancy Ann. Reference sediment collection in Carr Inlet occurred on 19 February with a van Veen sampler and on 13 March with a power grab.

Additional sediment was collected from KEN01 on 7 March using a vibracore. Five cores were collected from each of the 9 DMMUs, for a total of 45 cores, see Table 3 for coordinates and core collection information. Five cores per DMMU were necessary to collect sufficient sediment to run all chemical, bioassay and bioaccumulation testing on a single DMMU. Figure 2 shows the target and actual coring locations. The sampling and analysis plan (Sealaska, 2019a) was approved by the DMMP agencies prior to the start of sampling.

All decisions regarding changing sampling locations and acceptance of cores was coordinated real-time with the DMMO. The following changes were approved:

• After three failed attempts to obtain the minimum acceptable recovery rate of 75 percent for core C-2 in KEN01, the sampling location was moved approximately 50 ft to the southwest.

Core C-2 was subsequently collected on the first attempt at the revised location,

• After one failed attempt at location C-1 due to sand and gravel in the core, the sampling location was moved approximately 35 ft to the west. An acceptable sediment core was recovered on the first attempt at the revised position for C-1.

• After three attempts at the target location for C-17 in KEN04, the highest percent recovery obtained was 70%. DMMO allowed 70 percent recovery at C-17.

• After three failed attempts at location C-23, the sampling location was moved about 25 ft to the southeast. Cores collected at the original location suffered from poor retention of native glacial clay encountered below the dredge prism. An acceptable core was collected at the revised C-23 location on the first attempt.

DMMU compositing information is shown in Table 4. One error was discovered on the last day of sediment core processing, resulting in several deviations from the sampling and compositing plan.

Contractor staff discovered that KEN07-C34 (B) (the B designates a specific interval of the sediment core [i.e., the 16.9 to 19.9 ft interval]), had been inadvertently replaced with KEN09-C43 (B). A thin section of sediment (0.1 ft) from KEN09-C43 had been added to the dredge material (mudline to – 17 ft LWLW) composite of KEN07 and the interval from –17 to –19 ft LWLW of KEN09-C43 had been added to the z-layer composite of KEN07. When the error was identified, the archive sample originally labeled “KEN07-C34 Z” was retrieved from ARI’s freezer and relabeled to the correct sample ID, “KEN09-C43 Z”. The following highlights the samples that were archived and/or submitted for analysis.

• KEN07 composite, contains 0.1 ft from KEN09-C43 (–16.9 to –17.0 ft LWLW)

• KEN07-C34 Z archive sample collected

• KEN07-Z composite contains sediment from –17 to –19 ft LWLW interval from KEN09-C43 and was rendered unusable for testing

• KEN09-43 Z archive sample collected

• KEN09-C43 composited with other four KEN09 (C41, C42, C44, and C45) cores

• KEN09-Z composite contains sediment from four cores (C41, C42, C44, and C45) instead of five.

December 3, 2019 These deviations had minimal impact on the project. No z-samples were analyzed as part of this characterization, and the small amount of DMMU material from KEN09-C43 that was combined with KEN07 (less than one measuring cup) was considered insignificant volumetrically (0.58 percent of the total volume).

Elliott Bay Benchmark Field sampling occurred in a single day on 1 February 2019. Surface sediment samples were collected from 16 locations shown in Figure 3 using a 0.1 m2 stainless steel double van Veen grab sampler deployed from the R/V Nancy Ann owned and operated by MSS (Burley, Washington).

Sediment was mixed on 2 February 2019, at ARI to yield a single composite for testing.

5. Review of Chemical Data and Bioaccumulation Study Design Modifications. Analysis of conventionals and all standard DMMP COCs was conducted by Analytical Resources, Inc. of Tukwila, WA and Axys Environmental of Sydney, BC. The approved sampling and analysis plan (Sealaska, 2019a) was generally followed. All data was validated by an external third party at EPA Stage 2B and 10% of the dioxin data was validated at Stage 4. The final data presented in this memo are considered acceptable to be used in decision-making.

All 9 DMMUs were analyzed for the full list of DMMP COCs, including dioxins and TBT at the ARI laboratory in Tukwila, WA. All analytical results are shown in Table 5. Conventional results showed that the dredged material is a loam to silt loam, with 28-40 percent sand, 42-50 percent silt and 16- 21 percent clay. Total organic carbon (TOC) was relatively high, ranging from 2.38-5.66 % in the dredged material; TOC >3.5% is considered by Ecology to be above the typical range for freshwater sediments (Ecology, 2017). All DMMP COC’s were below SLs with the following exceptions:

• Butyl benzyl phthalate detected above the SL in KEN10 (duplicate of KEN02) and

KEN07

• Total Chlordane non-detect above the SL in all 9 DMMUs

• 4,4’-DDE non-detect above the SL in KEN05

The elevated non-detect pesticide results were re-evaluated by the laboratory and reported at the lower limit of detection per DOD policy (DOD/DOE QSM, 2018),). The LOD is defined as:

The smallest concentration of a substance that must be present in a sample in order to be detected at the detection limit with 99% confidence. At the LOD, the false negative rate (Type II error) is 1%. A LOD may be used as the lowest concentration for reliably reporting a non-detect of a specific analyte in a specific matrix with a specific method at 99% confidence.

Results in Table 5 are reported at the LOD for pesticide results via EPA Method 8081. When reporting at the LOD, only KEN02 and KEN03 still had total chlordane non-detect above the SL and KEN05 still had a non-detect exceedance of 4,4’-DDE.

USACE Navigation chose to pursue bioassay testing for the two DMMUs KEN02 and KEN07 with detected butyl benzyl phthalate exceedances, and high-resolution pesticide analysis via EPA method 1699 to lower the reporting limits to below SLs for KEN03 and KEN05. Bioassay results are discussed in Section 7 and high-resolution pesticide results are discussed below.

https://2.38-5.66

December 3, 2019 High Resolution Pesticide Analysis EPA method 1699 is a high-resolution gas chromatography/high-resolution mass spectrometry (HRGC/HRMS) isotope dilution method. All high-resolution pesticide analyses were conducted by Axys Environmental of Sydney, BC. Pesticide results from EPA Method 1699 are shown in Table 6.

The first round of results from KEN03 and KEN05 showed that 4,4’-DDE was non-detect in KEN05 below the SL and that total chlordane was detected above the SL in both DMMUs. It was not expected to find total chlordane in KEN05 because the results from EPA Method 8081 were non-detect at 1.49 U µg/kg. This prompted additional high-resolution pesticide analysis of the remaining DMMUs, with the exception of KEN02 which had already been shown to have failed the larval bioassay test at the time the second round of high-resolution pesticide analyses were initiated.

Results from the second round of high-resolution pesticide analysis are also shown in Table 6, and confirmed that total chlordane was detected above the SL in all DMMUs analyzed. By the time this data was received from the analytical laboratory the sediments had long expired the bioassay holding time of 56 days, so no further testing was pursued.

Dioxin Results Dioxin results are shown in Table 5, all dioxin results are reported as sum TEQ with non-detect concentrations included at ½ the reported value (ND=1/2). KEN01 had the highest dioxin TEQ (16.5 J pptr TEQ) and was selected by the DMMP agencies to undergo bioaccumulation testing as an individual sample. Dioxin concentrations in the remaining DMMUs ranged from 6.1 J - 12.2 J pptr TEQ. These eight DMMUs were composited together into a single sample, hereafter referred to as the Supercomposite, for bioaccumulation testing. After being created, the Supercomposite was tested and found to have a dioxin concentration of 9.95 J pptr TEQ. The Elliott Bay benchmark sample was also tested for conventionals and dioxin, results are shown in Table 7. The dioxin concentration was 10.7 J pptr TEQ.

Three bioaccumulation reference samples were collected from Carr Inlet and tested for conventionals and dioxins. The results from these samples ranged from 2.5 J – 6.05 J pptr TEQ and are shown in Table 8.

KEN01 sediment volume requirements After it was decided that KEN01 would be run individually for bioaccumulation testing, it was discovered that there wasn’t enough volume of sediment to conduct the bioaccumulation testing using the standard set-up. After discussion with the contractor and the DMMP agencies, it was decided that additional sediment from the project area was needed to conduct the appropriate testing. On 13 March 2019, an additional five cores were collected from the same five sample locations (C-1 – C-5) and the additional sediment was composited with the original KEN01 sediment. Dioxin analysis of the newly composited KEN01, called KEN01-Comp, showed that the concentration was 15.0 J pptr TEQ which was less than the original concentration of 16.5 J pptr TEQ, but still above the dioxin concentration of the Supercomposite sample (9.95 pptr TEQ).

6. Bioassay Testing Bioassay testing was carried out at the EcoAnalysts, Inc. laboratory in Port Gamble, WA. The standard suite of three DMMP tests: the amphipod mortality, larval development, and polychaete growth bioassays were conducted on two test sediments (KEN02 and KEN07) and a reference

December 3, 2019 sample. Interpretation guidelines are found in Table 12. Because test sediments were from a freshwater environment and the proposed disposal location is in marine waters, the sediment required acclimation to marine conditions prior to initiating the bioassay tests. These procedures are described in detail below.

All bioassay tests were initiated on 22 March, one day past the 56-day holding time for KEN02 and within the 56-day holding time for KEN07.

Bioassay Reference Collection A bioassay reference sample was collected from Carr Inlet on 13 March. Table 9 shows field grain size results for all nine DMMUs and all Carr Inlet reference samples collected. Table 10 provides a summary of sediment collection information and Figure 4 shows the sampling areas within Carr Inlet. Reference sample CR-23 was selected for the bioassay reference because historic data from this location indicated that it best matched the grain size range of KEN02 and KEN07 (47-50% fines) based on wet-sieve field results. Field grain size results for CR-23 was 52% fines, which was considered a good match for the bioassay tests. Laboratory grain size analysis determined that the percent fines for CR-23 was 58.7%, and for KEN02 and KEN07 percent fines were 70.5 and 59.3, respectively.

Saltwater Acclimation Procedures Non-contaminant-related effects can confound bioassay results if sediment conditions are not appropriate to test organisms. For these tests, acclimation of freshwater sediments to marine conditions was necessary to eliminate potential non-treatment effects such as increasing ammonia and/or sulfide concentrations and salinity levels outside of optimal ranges for testing conditions.

Established acclimation procedures were followed prior to the initiation of the juvenile polychaete and amphipod bioassays as prescribed in the SAP. On 5 March 2019 sediments for amphipod and polychaete bioassays (plus sacrificial chambers) were layered into test chambers per the specific test methodology, topped with filtered seawater, aerated, and kept under static conditions for nine days. Ammonia, temperature, pH and salinity were measured in overlying water daily starting on 11 March, and sulfides were measured daily starting on 13 March. Porewater measurements of these same parameters were also conducted intermittently on test sediments within sacrificial chambers (on 11, 15, 18, and 21 March). Porewater measurement results are available in Appendix K of the final Kenmore data report (Sealaska, 2019c).

Low porewater pH values were observed in all acclimated test chambers on 11 March, and there was a concern that pH values were low enough (i.e., <6) in some chambers that it would impact the organism’s health. In addition, while un-ionized ammonia (UIA) concentrations were below organism toxicity trigger values, there was concern that UIA concentrations would increase in the amphipod and polychaete testing chambers as they continued to acclimate to the more typical higher pH values (7-8) associated with marine conditions. Ammonia purging via twice daily overlying water renewals began on 15 March and continued until testing was initiated 22 March, for the amphipod and polychaete test chambers. Due to the shifting pH and resulting ammonia concentrations, daily overlying water renewals were performed throughout the amphipod and juvenile polychaete testing period.

December 3, 2019 While acclimation is generally not required for the larval development test, sediment for the larval test was acclimated in the same manner as the amphipod and juvenile polychaete tests.

Specifically, an individual chamber of acclimated sediment was decanted, and the appropriate amount of sediment was distributed to each of the larval test chambers and a surrogate water quality chamber.

Bioassay Results Toxicity test results are summarized in Table 11, and comparisons of individual bioassay test controls and reference sediments to performance standards are provided in Table 12.

Amphipod 10-Day Mortality The 10-day amphipod test using Eohaustorius estuarius was initiated on 22 March 2019. E.

estuarius is the recommended species when clay content in the test sediment is <20%. The clay content in KEN07 was 17% and in KEN02 was 21.1%, above the recommended range for E.

estuarius. All organisms survived in both the control and reference sediments, meeting the DMMP’s performance standards for a valid test (Table 12). Mean amphipod mortality was 4 percent in KEN02 and was 9 percent in KEN07 (Table 11), indicating that the high clay content in KEN02 did not negatively affect organism health. There were no hits for either KEN02 or KEN07.

All water quality parameters (dissolved oxygen, temperature, salinity, and pH) were within the acceptable limits for the duration of the test. Measurements of total ammonia and total sulfides were generally made in overlying water at the beginning and end of the test. Ammonia values were below the published threshold concentrations of 30 mg/L for total ammonia and 0.8 mg/L for UIA (DMMP 2018). All sulfide concentrations in interstitial and overlying water were below the trigger value of 0.122 mg/L (DMMP 2018). Consequently, ammonia and sulfide concentrations were not expected to cause non-treatment effects.

Neanthes 20-Day Growth and Development The 20-day juvenile infaunal growth test using Neanthes arenaceodentata was initiated on 22 March 2019. No mortality was observed in the control sediment, and mean individual growth (MIG) in the control sediment was 0.401 mg/individual/day ash free dry weight (AFDW). One replicate in the control had 100 percent mortality. The laboratory was not able to identify a reason for this and the water quality data did not indicate any issues. This replicate was statistically qualified as an outlier by performing a Grubbs’ test and was not included in statistical data summaries or comparisons. All organisms survived in the CR-23 reference sediment, and MIG for the reference treatment was 0.361 mg/individual/day AFDW. Both control and reference treatments met the DMMP’s performance standards for a valid test (Table 12). All organisms survived in both KEN02 and KEN07. MIG was 0.350 mg/individual/day AFDW for KEN02 and 0.337 mg/individual/day AFDW for KEN07.

All water quality parameters (dissolved oxygen, temperature, salinity, and pH) were within the acceptable limits for the duration of the test. Measurements of total ammonia and total sulfides were made in overlying water at the beginning and end of the test. Ammonia values were below the published threshold concentrations of 0.46 mg/L UIA (DMMP 2018). All sulfide concentrations in interstitial and overlying water were below the trigger value of 3.4 mg/L (DMMP 2018).

Consequently, ammonia and sulfide concentrations were not expected to cause non-treatment effects.

December 3, 2019

Larval Development The larval development test using Mytilus galloprovincialis was initiated on 22 March 2019. All ammonia concentrations were below threshold values and are not expected to have contributed to toxicity of the sediments.

Use of the resuspension protocol is recommended for sediments with high concentrations of fines, wood waste or other flocculent material (DMMP, 2013). There is no specific definition of “high concentration” of fines in the clarification paper, however the percent fines for KEN02 (70.5%) and KEN07 (59.3%) are considered high by the DMMP agencies. The resuspension protocol was appropriately used to terminate the test and enumerate larvae.

Results are reported as control-normalized normal survival, which is calculated by dividing the number of normal larvae from the test or reference treatment at the end of the exposure period by the number of normal larvae in the seawater control at the end of the exposure period. The combined normal survivorship in the control was 0.83 (83 percent), and the mean control-normalized normal survival was 0.91 (91 percent) in the reference treatment. Both control and reference treatments met the DMMP’s performance standards for a valid test (Table 12). Mean control-normalized survival was 21 percent in KEN02 and 57 percent in KEN07.

A number of irregularities occurred during testing, including:

• One reference replicate had a much lower number of normal larvae when compared to the other four replicates for this sample. The laboratory was not able to identify a reason for this and the water quality data do not indicate any issues. This replicate was statistically qualified as an outlier by performing a Grubbs’ test and was not included in statistical data summaries or comparisons

• Salinity was outside the recommended range of 28 ± 1 ppt in CR-23; salinity was 30 ppt on Days 1 and 2. Because the 30 ppt measurement is within the tolerance range of the test organisms, it is not expected to have had a negative effect on the test.

• The hydrogen sulfide concentration measured in overlying water for KEN02 (0.0026 mg/L) on Day 0 was above the trigger value of 0.0025 mg/L (DMMP 2015). Due to a laboratory oversight, sulfides were not measured on Day 1 and therefore hydrogen sulfide concentrations could not be calculated. By test termination on Day 2, sulfide concentrations had decreased to the level that they were not detected in KEN02.

The hydrogen sulfide concentration is a calculated value of ion dissociations based on measurements of total dissolved sulfides, temperature, salinity, and pH. The dissociation curve with respect to pH is steepest within the range of 7 to 8, meaning that small pH changes within this range have a substantial impact on the calculated hydrogen sulfide concentration. For example, an increase in pH of 0.03 units results in a calculated hydrogen sulfide concentration below the trigger value. In addition, the hydrogen sulfide purging trigger was chosen based on the NOEC/LOEC threshold concentrations between 0.0041 and 0.0068 mg/L (Weston Solutions (2006). EcoAnalysts’ bioassay report states that hydrogen sulfide “values below this range are not likely to elicit negative biological effects”.

Taken together, this information provides evidence that the slightly elevated hydrogen sulfide

December 3, 2019 concentration at the beginning of the test was not likely to have caused significant non-treatment effects.

• pH in KEN02 at test termination was 6.8; below the recommended range of 7-9 . Average pH in KEN02 over the 48 hr testing period was 7.17; average pH in KEN07 was 7.33. Larval development of M. Galloprovincialis have been shown to be sensitive to timing of exposure to acidified seawater (Kapsenberg et al., 2018). The most acidified water used in the cited study was 7.4 – higher than the average pH in both of the test treatments – and that treatment resulted in only 56-64% normal larval development, compared to >95 normal development in the constant pH 8.1 and pH starting at 7.4 and increasing to 8.1 treatments. This information suggests that lower pH may have contributed to the reductions in normal development observed in both KEN02 and KEN07 sediments.

Summary Both KEN02 and KEN07 passed the amphipod and Neanthes growth tests. Due to the number of irregularities in the larval development test, it is not possible to eliminate the influence of non-treatment effects on the test results. Therefore the results of the larval test are not considered valid for decision-making by the DMMP agencies. Without a full suite of bioassay results upon which to base a determination of the suitability of the test sediments, a decision cannot be made. Without further information to demonstrate the test sediment will not have more than a minor adverse effects to benthic toxicity, both KEN02 and KEN07 are considered unsuitable for open-water disposal. Re-testing of sediments using modified methods to either account for pH-induced impacts (low pH control) or to eliminate pH induced impacts (longer acclimation) would be allowed, but is not being pursued due to project timelines.

7. Bioaccumulation Testing Bioaccumulation testing was conducted by EcoAnalysts Laboratory of Port Gamble, WA. Two species – an adult marine clam (Macoma nasuta) and an adult polychaete (Nephtys caecoides) – were exposed to sediment from the two composited test sediments, the Carr Inlet reference sediment, and the Elliott Bay Benchmark sample for 45 days. The species were co-tested, meaning they shared aquaria during the exposure. Five replicates were run for each test sediment, reference sediment, benchmark sediment, and negative control. An additional 175 ml of sediment was added to each test chamber every seven days. There was no supplemental feeding. The SAP (Sealaska, 2019a) provides additional details regarding the protocol and test conditions.

Static renewal is the standard method for DMMP bioaccumulation tests. For this project, flow-through conditions for the entire duration of exposure were allowed for the following reasons:

• It is the standard set-up used by the biological testing laboratory

• Dioxins/furans are highly hydrophobic, with log Kow values that indicate that they tend to adsorb stongly to organic components in the sediment and are therefore very unlikely to be lost to the water.

Similar to bioassay testing, established acclimation procedures were followed prior to initiating the bioaccumulation exposure tests for KEN01-Comp and the Supercomposite. The test acclimation period began 5 March 2019 for the Supercomposite and 9 March 2019 for KEN01-Comp due to the

December 3, 2019 need to collect additional sediment from KEN01. Procedures for acclimating the freshwater sediments to marine testing conditions were performed in the same manner as the bioassay samples except that the chambers were maintained under flow-through conditions with clean seawater from North Hood Canal, Washington. Acclimation typically continues until a peak in interstitial ammonia concentrations (commonly overlying water is measured as a proxy) is observed.

Water quality measurements did not start until 11 March so it was not possible to calculate unionized ammonia prior to that date. By 21 March, it was observed that ammonia concentrations in both exposure test sediments were decreasing or stable throughout testing procedures due to the continuous flow-through of overlying water during testing.

Two notable deviations to test conditions occurred. First, the tests were initiated on 22 March – two days past the standard 56-day holding time. Second, the flow rate was just above the recommended range for one replicate (CI16-8, day 33). The flow was adjusted to the appropriate range when this was observed. The replicate was within range the previous day that it was tested (day 28), therefore the flow rate exceedance of 1mL per 30 seconds lasted for a maximum of five days (total of 14.4L extra). The normal flow rate range is 38-56 mL/ 30 sec. The DMMP agencies have determined that none of these deviations had a significant impact on the test conditions or validity of the test results.

Testing was terminated on day 45 according to standard procedures. Mean organism survival in the control, test sediments and reference sediment ranged from 90-100%, and is shown in Table 13. At test termination there was sufficient tissue available for dioxin analysis in all test and reference treatments. Tissue samples were collected per the SAP and analyzed for dioxin and lipids. The analytical tissue results for all test, reference and benchmark sediments can be found in Table 14 for Macoma and in Table 15 for Nephtys.

In order to establish pre-test tissue concentrations, a sub-sample of test organisms was isolated prior to test initiation. These pre-test organisms were allowed to depurate for 24 hours, tissues removed and rinsed in deionized water, and then frozen for chemical analysis. Pre-test dioxin results are also presented in Table 14 for Macoma and Table 15 for Nephtys.

8. Bioaccumulation Test Results and Interpretation.

All dioxin concentrations discussed in this memo were calculated with non-detects equal to one-half the detection limit, and all tissue results are on a wet weight basis. Macoma and Nephtys tissue dioxin TEQ concentrations were low for the reference sediment, the Elliott Bay Benchmark sediment, and all test sediments, with TEQs well below 1 ng/kg wet weight in all cases. See Figures 5 and 6 for graphical depictions of the tissue dioxin TEQ results. For Macoma, tissue concentrations ranged from 0.149 to 0.257 ng/kg ww TEQ for the individual Carr Inlet reference and Elliott Bay Benchmark replicates and 0.146 to 0.375 ng/kg ww TEQ for the test sediment replicates.

For Nephtys, tissue concentrations ranged from 0.0650 to 0.146 ng/kg ww TEQ for the individual reference and Elliott Bay Benchmark replicates and 0.0783 to 0.183 ng/kg ww TEQ for the test sediment replicates.

Pre-test tissue concentrations are an important indicator of the validity of the test. Pre-test tissue concentrations indicate the amount of bioaccumulation that has occurred prior to the start of the laboratory bioaccumulation tests. Dioxin tissue concentrations were analyzed in triplicate for each organism. Macoma pre-test tissue concentrations were low, ranging from 0.0913 to 0.146 ng/kg ww

December 3, 2019 TEQ, and with two of the three results non-detect. Nephtys pre-test tissue concentrations were higher, ranging from 0.109 U to 0.337 J ng/kg ww TEQ. In fact, the average pre-test tissue dioxin concentration in Nephtys was higher than any of the post-test results from KEN01 and the Supercomposite. This is due, in part, to one of the pre-test replicate results being significantly higher (0.337 J) than the other two (0.169 U – 0.109 U). Note both pre-test and post-test tissue concentrations are very low and near limits of detection.

As indicated previously, a TTL for the protection of human health and other higher trophic level species has not yet been established by the DMMP agencies for dioxin. In the absence of a TTL, the DMMP User Manual (DMMP, 2018) calls for comparison of test sediment tissue concentrations to Carr Inlet reference sediment tissue concentrations using a one-tailed t-test with an alpha level of 0.10.

Sealaska Environmental used the Gehan test to statistically compare the mean TEQ for the five replicate tissue samples from each test sediment to the mean TEQ for the five replicate tissue samples from the Carr Inlet reference for each test species. The Gehan test is a nonparametric test that handles data sets with NDs and multiple DLs, it assumes comparable shapes and variability.

ProUCL recommends using the Gehan test when multiple detection limits are present in the data set (EPA, 2015). It is recommended that a minimum of 8–10 observations be used for hypothesis testing approaches such as the Gehan test; thus, results from tests applied to smaller data sets may not be reliable.

Results of using the Gehan test to compare total TEQs in test sediments to the Carr Inlet reference and Elliott Bay benchmark are shown in Table 16. KEN01 was statistically different than reference for Macoma but not Nephtys. The Supercomposite was not statistically different than reference for either species. The same comparison made against the Elliott Bay benchmark sample results showed that none of the test sediments were statistically different for either Macoma or Nephtys.

The DMMP verified the statistical results using BioStat, which are described in Table 17.

Screenshots from the BioStat comparisons are provided in Attachment 1. In summary, KEN01 was found to be statistically different from reference for both species, but when compared to the Elliott Bay benchmark, only Macoma results were statistically different. For the Supercomposite, none of the results were statistically different from reference or benchmark for either of the test species.

DMMP agencies had agreed prior to testing that a statistically significant difference between test sediments and reference results would be a necessary but insufficient conditions for finding dredged material unsuitable for open-water disposal. Several additional lines of evidence were considered in evaluating the significance of the statistical difference, including a) percentage above reference; b) the role of non-detects; and c) comparison to practical quantitation limits (PQLs) and d) comparison to dioxin concentrations measured in field-collected polychaetes from unimpacted areas near the Elliott Bay disposal site.

a) Supercomposite tissue concentrations were less than 20% above reference for both tests species. KEN01 tissue concentrations were more than 20% above reference, 57 and 30% for Macoma and Nephtys, respectively.

December 3, 2019

b) Non-detect concentrations play a significant role in the summation of TEQs for all tissue samples analyzed for this project. 75.3% of the individual congener results for Macoma in Kenmore sediments were non-detect, and 75.9% of Nephtys results were non-detect. The Carr Inlet reference sediment had an even higher fraction of congeners that were non-detect, 80% for Macoma and 85.9% for Nephtys. For Elliott Bay benchmark tissue sample results, 69.4% of congener results were non-detect for Macoma and 78.8% of congener results were non-detect for Nephtys. With non-detects set equal to one half the detection limit, the contribution of non-detects to the TEQ for the test samples ranges from 65.7 to 93.7% for Macoma and 76.1 to 99.9% for Nephtys.

Another way to demonstrate to effect of the contribution of non-detects is by comparing dioxin concentrations calculated with ND=1/2 DL to those calculated with ND set equal to zero (see Tables 14 and 15). With TEQ sums calculated with ND=0, the Macoma tissue concentrations for the Kenmore test sediments ranged from 0.0126 to 0.0774 ng/kg ww TEQ, compared with

0.146 to 0.375 ng/kg ww TEQ (ND=1.2 DL). For Nephtys, tissue concentrations in the test sediments with ND=0 ranged from 0.0000642 to 0.0439 ng/kg ww TEQ, compared with 0.0783 to 0.183 ng/kg ww TEQ (ND=1/2 DL). Overall, test sediment tissue concentrations were about an order of magnitude smaller when non-detect concentrations are not included. Alternative methods for addressing non-detects, such as Kaplan-Meier, were considered, but the large number of non-detect concentrations (>50%) precluded the use of that or similar methods.

c) Practical quantitation limits are an important consideration when concentrations are low. The PQL is the lowest concentration of an analyte that can be reliably measured within specified limits of precision and accuracy under routine laboratory operating conditions. Concentrations above the PQL can be measured with a high degree of confidence, while concentrations below the PQL are typically considered estimates. The PQL for each dioxin congener for this project was established as the lowest method calibration standard used by ARI to calibrate its instruments (Table 18). A broad evaluation of the significance of PQLs can be achieved by comparing tissue TEQs to the sum TEF-weighted PQLs. The sum of TEF-weighted PQLs was

11.41 pptr TEQ. Table 11-1 of Ecology’s SCUM II establishes programmatic sediment and tissue PQLs for establishing PQL-based SCO and CSLs. The tissue PQL for dioxins as sum TEQ is 1 pptr ww TEQ (Ecology, 2018). As noted previously, Macoma tissue concentrations ranged from 0.146 to 0.375 pptr ww TEQ and Nephtys tissue concentrations ranged from

0.0783 to 0.164 pptr ww TEQ for the test sediment replicates. Since the Ecology PQL is less than the project TEF-weighted PQL for dioxin sum TEQs, comparison with Ecology’s PQL is appropriate. All project dioxin sum TEQ tissue concentrations were less than Ecology’s sum TEQ tissue PQL.

d) The DMMP evaluated the environmental significance of the bioaccumulation results by comparing them to monitoring data from the vicinity of the Elliott Bay disposal site. In 2007, as part of a special dioxin study, specimens from the polychaete genera Glycidera, Travisia, and Maldanidae were collected from 12 off-site stations (perimeter, transect and benchmark). Of these stations, the perimeter stations are closest to the disposal site, being located approximately 0.125 (760 ft) outside the site boundary. Given their limited mobility and the distance of the collection stations from the site, all polychaetes collected from the 12 off-site stations can be considered unimpacted by dredged material disposal and representative of

December 3, 2019 background conditions in Elliott Bay. Tissue results are shown below.

Table 19. Elliott Bay in situ dioxin tissue concentrations (ND=½DL) Organism Genus Dioxin concentrations (pptr ww TEQ, ND =½DL)

Glyciderae 0.307 – 2.51 Maldanidae 0.417 to 0.525

Travisia 0.662

In comparison, the dioxin concentrations in tissue from Nephtys exposed to sediment from the Kenmore navigation channel ranged from 0.0831 to 0.183 pptr ww TEQ. As can be seen from this comparison, the Kenmore Nephtys concentrations for both Nephtys and Macoma were less than tissue concentrations from all 12 Elliott Bay polychaete samples.

Summary The bioaccumulation testing data are summarized in Table 20 and as follows: dredged material samples from Kenmore resulted in tissue concentrations that were statistically greater than reference for KEN01 but not for the Supercomposite. Kenmore tissue concentrations resulting from bioaccumulation exposures were more than 20% greater than reference concentrations for KEN01 but not the Supercomposite, despite the absolute concentrations being quite low in all samples.

Total TEQs were largely driven by the large number of non-detect results, which contributed between 65.7- 99.9% of the total TEQ sums. Even when detected, only one congener, OCDD, had concentration above the laboratory’s congener-specific PQL, all other detections were much lower than the PQL, which increases the uncertainty about the results. In addition, OCDD has one of the lowest TEFs and therefore the contribution of this congener is swamped by the non-detect results for congeners with higher TEF’s. Finally, the Kenmore polychaete tissue results had lower dioxin concentrations than that observed in various polychaete species collected in the vicinity of the Elliott Bay disposal site (but not impacted by the site) in 2007.

These lines of evidence were carefully considered by the DMMP agencies, and taken together they demonstrate that there is not likely to be a significant bioaccumulation impact to the disposal site from either KEN01 or the Supercomposite. Therefore, with respect to bioaccumulation, all DMMUs from the Kenmore federal navigation channel are considered suitable for open-water disposal.

9. Sediment Exposed by Dredging. The sediment to be exposed by dredging must either meet the State of Washington Sediment Quality Standards (SQS) or the State’s Antidegradation standard (Ecology, 2013) as outlined by DMMP guidance (DMMP, 2008).

Z-layer testing was completed in 2014, and included analysis of freshwater bioassays due to elevated nickel throughout the project. Nickel is considered to have an elevated natural background in the area (RSET 2015). All DMMU z-samples passed bioassay testing. The DMMP agencies reviewed the previous testing results and determined the results were still valid for deep sediments that have been undisturbed since the 2014 testing.

Therefore, the sediment to be exposed by dredging is not considered to be degraded relative to the currently exposed sediment surface. On this basis the DMMP agencies conclude that this project is in compliance with the State of Washington anti-degradation policy.

December 3, 2019

10. Suitability Determination. This memorandum documents the evaluation of the suitability of sediment proposed for dredging from the Kenmore federal navigation channel for open-water disposal at the Elliott Bay non-dispersive disposal site. The approved sampling and analysis plan was generally followed.

With respect to bioaccumulation, after full consideration of all lines of evidence, KEN01 and the Supercomposite passed bioaccumulation testing for dioxin.

With respect to benthic toxicity the results are summarized are Table 21. KEN02 and KEN07 had an incomplete suite of valid bioassays from which to make a determination. Therefore, in the absence of all the information needed to make a decision on bioassays, these two DMMUs are not suitable for open-water disposal. The remaining seven DMMUs had detected SL exceedances of total chlordane and no bioassays were run. Bioassay testing is needed to determine if these seven DMMUs are suitable, without this information they are not suitable for open-water disposal.

In summary, based on the results of the previously described testing, the DMMP agencies conclude that all 34,350 CY are unsuitable for open-water disposal at the Elliott Bay non-dispersive site based on the bioassay results. Bioassay re-testing of all DMMUs would be needed to make a final determination on the suitability of the proposed dredged material from the Kenmore navigation channel.

A determination regarding the suitability of the material for upland disposal at any location other than a permitted landfill must be coordinated with the local Health Department.

This suitability determination does not constitute final agency approval of the project. A final decision will be made after full consideration of agency input, and after an alternatives analysis is done under section 404(b)(1) of the Clean Water Act.

December 3, 2019

11. References.

DMMP, 2018. Dredged Material Evaluation and Disposal Procedures (Users Manual). Prepared by the Seattle District Dredged Material Management Office for the Dredged Material Management Program, December 2018.

DMMP, 2017. Determination Regarding the Suitability of Proposed Dredged Material from the Port of Everett Marina, Evaluated under Section 404 of the Clean Water Act, for Placement at the Dredged Materila Management Program’s Unconfined Open-Water Disposal Site in Port Gardner.

March 30, 2017

DMMP, 2015. Modification to Ammonia and Sulfide Triggers for Purging and Reference Toxicant Testing for Marine Bioassays. DMMP Clarification Paper prepared by Laura Inouye (Ecology), Erika Hoffman (EPA), and David Fox (USACE) for the Dredged Material Management Program, August 2015.

DMMP, 2013. Bioassay Endpoint Refinements: Bivalve Larval and Neanthes Growth Bioassays.

Clarification paper prepared by David Kendall (USACE), Russ McMillan (Ecology), and Bill Gardiner, Brian Hester and Jack Word (NewFields, LLC) for the DMMP and SMS Programs, June, 2013.

DMMP, 2011. Marine Sediment Quality Screening Levels: Adopting RSET Marine SLs for Use in DMMP. A Clarification Paper prepared by Laura Inouye (Ecology) and David Fox (USACE) for the Dredged Material Management Program, June 2011.

DMMP, 2010. Dredged Material Management Program New Interim Guidelines for Dioxins.

December 6, 2010.

DMMP, 2009. Clarifications to the DMMP Bioaccumulation Protocol. A DMMP Clarification Paper and SMS Technical Information Memorandum prepared by David Kendall (USACE) and Russ McMillan (Ecology) for the DMMP/SMS agencies. September 10, 2009.

DMMP, 2008. Quality of Post-Dredge Sediment Surfaces (Updated). A Clarification Paper Prepared by David Fox (USACE), Erika Hoffman (EPA) and Tom Gries (Ecology) for the Dredged Material Management Program, June 2008.

Ecology, 2017. Sediment Cleanup User’s Manual II (SCUM II): Guidance for Implementing the Cleanup Provisions of the Sediment Management Standards, Chapter 173-204 WAC. December 2017.

Ecology, 2013. Sediment Management Standards – Chapter 173-204 WAC. Washington State Department of Ecology, February 2013.

EPA, 2015. ProUCL Version 5.1 User Guide. Stasticial Software for Environmental Applications for Data Sets with and without Nondetect Observations. U.S. EPA Office of Research and Development, October 2015.

December 3, 2019 Kapsenburg L, Miglioli A, Bitter MC, Tambutte E,…