Basis_of_Design_Report_Attachment_8.pdf

25+79 Cross Section Results

25+79 Initial Results

Model Run SF Results Drained Undrained –

Clay with Uniform Shear Strength

Undrained – Clay with Shear Strength Profile

Existing Conditions 1.35 1.00 1.03 2H:1V slope from vertical extent of dredging at navigational channel edge

1.21 1.02 --

3H:1V slope from vertical extent of dredging at navigational channel edge

1.12 1.01 --

Combination Slope 1.11 1.02 0.97 3H:1V extending from toe of slope 1.35 1.04 0.97 3H:1V extending downward, offset 10 feet 1.44 1.05 0.98 Rock Toe Buttress 1.62 0.90 0.91

25+79 Existing Conditions ‐ Drained

Grid Radius Wedge Entry/Exit

SF = 1.35 SF = 1.20 SF = 1.26

25+79 Sloping Conditions‐ Drained 2:1 from Vertical Extent at Navigation

Channel Edge

SF = 1.21 SF = 1.12

3:1 from Vertical Extent at Navigation Channel Edge

25+79 Combined Slope‐ Drained

SF = 1.11 SF = 1.21 SF = 1.22

25+79 Additional Design Slopes – Drained

3:1 Offset from toe of slope 3:1 no offset from toe of slope

SF = 1.44 SF = 1.35 SF = 1.62

Rock Toe Buttress

25+79 Existing Conditions – Undrained with constant clay shear strength

SF = 1.00 SF = 1.10 SF = 0.93

25+79 Sloping Conditions‐ Undrained with constant clay shear strength

2:1 from Vertical Extent at Navigation Channel Edge

SF = 1.02 SF = 1.01

25+79 Combined Slope – Undrained with constant clay shear strength

SF = 1.02 SF = 1.03 SF = 1.04

25+79 Additional Design Slopes – Undrained with constant clay shear strength 3:1 Offset from toe of slope 3:1 no offset from toe of slope

SF = 1.05 SF = 1.04 SF = 0.90

25+79 Existing Conditions ‐ Undrained with clay shear strength increasing by 1.4 psf/ft for depths up to 25 ft and 7.4 psf/ft for depths below 25 ft

SF = 1.03 SF = 1.07 SF = 0.92

25+79 Combined Slope ‐ Undrained with clay shear strength increasing by 1.4 psf/ft for depths up to 25 ft and 7.4 psf/ft for depths below 25 ft

SF = 0.97 SF = 1.01 SF = 0.89

25+79 Additional Design Slopes ‐ Undrained with clay shear psf/ft for depths below 25 ft 3:1 Offset from toe of slope 3:1 no offset from toe of slope

SF = 0.98 SF = 0.97 SF = 0.91

25+79 Existing Conditions ‐ Undrained with clay shear strength increasing by 8.1 psf/ft

SF = 0.86 SF = 0.95 SF = 0.87

25+79 Combined Slope ‐ Undrained with clay shear strength increasing by 8.1 psf/ft

SF = 0.87 SF = 0.88 SF = 0.79

25+79 Updated Data Results

Drained Undrained –

Clay with Uniform

Shear Strength

Undrained

– Clay with Shear Strength Profile

Existing Conditions 1.21 1.02 0.93 3H:1V extending from toe of slope 1.14 1.06 0.96 3H:1V extending downward, offset 10 feet 1.26 1.07 0.96

25+79 Updated Parameters ‐ Drained

Existing Conditions 3:1 no offset from toe of slope 3:1 slope offset 10 ft

SF = 1.21 SF = 1.14 SF = 1.26

25+79 Updated Parameters – Undrained Uniform Shear Strength

SF = 1.02 SF = 1.06 SF = 1.07

25+79 Updated Parameters ‐ Undrained Shear Strength Profile

SF = 0.93 SF = 0.96 SF = 0.96

37+23 Cross Section Results

37+23 Initial Results

Drained Undrained –

Clay with Uniform Shear Strength

Undrained – Clay with Shear Strength Profile

Existing Conditions 1.25 1.13 1.09 2H:1V slope from vertical extent of dredging at navigational channel edge

1.12 1.09 --

3H:1V slope from vertical extent of dredging at navigational channel edge

0.97 1.03 --

Combination Slope 0.93 1.06 1.18 3H:1V extending from toe of slope 1.33 1.16 1.04 3H:1V extending downward, offset 10 feet 1.48 1.18 1.03 Rock Toe Buttress 1.04 1.03 1.01

37+23 Existing Conditions – Drained

SF = 1.25 SF = 1.20 SF = 1.24

37+23 Sloping Conditions – Drained

2:1 from Vertical Extent at Navigation Channel Edge

SF = 1.12 SF = 0.96

37+23 Combined Slope – Drained

SF = 0.93 SF = 1.11 SF = 0.99

37+23 Additional Design Slopes – Drained

3:1 Offset from toe of slope 3:1 no offset from toe of slope

SF = 1.48 SF = 1.33 SF = 1.04

37+23 Existing Conditions – Undrained with constant clay undrained shear strength

SF = 1.13 SF = 1.20 SF = 1.27

37+23 Sloping Conditions – Undrained with constant clay undrained shear strength

2:1 from Vertical Extent at Navigation Channel Edge

SF = 1.09 SF = 1.03

37+23 Combined Slope ‐ Undrained with constant clay undrained shear strength

SF = 1.05 SF = 1.06 SF = 1.11

37+23 Additional Design Slopes – Undrained with constant clay undrained shear strength 3:1 Offset from toe of slope 3:1 no offset from toe of slope

SF = 1.18 SF = 1.16 SF = 1.03

37+23 Existing Conditions ‐ Undrained with clay shear psf/ft for depths below 25 ft

SF = 1.09 SF = 1.10 SF = 1.04

37+23 Combined Slope ‐ Undrained with clay shear strength increasing by 1.4 psf/ft for depths up to 25 ft and 7.4 psf/ft for depths below 25 ft

SF = 1.18 SF = 1.17 SF = 1.06

37+23 Additional Design Slopes – Undrained with clay shear psf/ft for depths below 25 ft 3:1 Offset from toe of slope 3:1 no offset from toe of slope

SF = 1.03 SF = 1.04 SF = 1.01

37+23 Existing Conditions ‐ Undrained with clay shear increasing by 8.1 psf/ft

SF = 0.91 SF = 0.92 SF = 0.89

37+23 Combined Slope ‐ Undrained with clay shear increasing by 8.1 psf/ft

SF = 0.84 SF = 0.87 SF = 0.84

37+23 Updated Data Results

Drained Undrained –

Clay with Uniform

Shear Strength

Undrained – Clay with

Shear Strength Profile

Existing Conditions 1.41 1.26 1.19 3H:1V extending from toe of slope 1.32 1.30 1.23 3H:1V extending downward, offset 10 feet 1.46 1.33 1.25 Backfill at 3H:1V (Simulating Post-Temporary Shoring Removal) 1.94 1.37 1.29

37+23 Updated Parameters ‐ Drained

SF = 1.41 SF = 1.32 SF = 1.46

37+23 Updated Parameters – Undrained Uniform

SF = 1.26 SF = 1.30 SF = 1.33

37+23 Updated Parameters ‐ Undrained Shear

SF = 1.19 SF = 1.23 SF = 1.25

37+23 Backfill at 3H:1V (Simulating Post‐Temporary Shoring Removal)

Drained Undrained Constant Shear Strength Undrained Shear Strength Profile

SF = 1.94 SF = 1.37 SF = 1.29

30+50 Cross Section Results

30+50 Initial Results

Model Run Drained SF Undrained SF

Existing Conditions 1.03 0.98

3:1 offset 1.05 0.98

3:1 at toe of slope 1.05 0.99

30+50 Drained

Existing Conditions 3:1 Offset from toe of slope 3:1 no offset from toe of slope

SF = 1.03 SF = 1.06 SF = 1.05

30+50 Undrained

SF = 0.98 SF = 0.98 SF = 0.99

30+50 Updated Results

Existing Conditions 0.91 1.08

3:1 offset 0.92 1.06

3:1 at toe of slope 0.93 1.07

30+50 Drained

SF = 0.91 SF = 0.93 SF = 0.92

30+50 Undrained

SF = 1.08 SF = 1.06 SF = 1.07

31+66 Cross Section Results

Existing Conditions 1.25 1.07

Rock Toe Buttress 1.36 1.05

31+66 Drained

SF = 1.25 SF = 1.36

Existing Conditions Rock Toe Buttress

31+66 Undrained

SF = 1.07 SF = 1.05

Existing Conditions Rock Toe Buttress

31+66 Updated Data Results

Drained Undrained –

Clay with Uniform

Shear Strength

Undrained

– Clay with Shear Strength Profile

Existing Conditions 1.03 1.17 0.98 3H:1V extending from toe of slope 0.99 1.17 0.99 3H:1V extending downward, offset 10 feet 1.00 1.18 1.01

31+66 Updated Parameters ‐ Drained

SF = 1.03 SF = 0.99 SF = 1.00

31+66 Updated Parameters – Undrained Uniform

SF = 1.17 SF = 1.17 SF = 1.18

31+66 Updated Parameters ‐ Undrained Shear

SF = 0.98 SF = 0.99 SF = 1.01

APPENDIX J

CAP DESIGN

http://portal/sites/LRROC/Document Library/100 Percent BODr/Appendicies/Appendix J_Cap Design/Rouge_Appendix J_Cap_Design_20160617.docx

APPENDIX J

CAP DESIGN

LOWER ROUGE RIVER – OLD CHANNEL

BASIS OF DESIGN REPORT

Prepared for

Prepared by Anchor QEA, LLC1

August 2016

1 Anchor QEA, LLC, is the sole member of its affiliate, Anchor QEA of Michigan, LLC.

Basis of Design Report August 2016 Lower Rouge River – Old Channel J-i 130287-04.01

TABLE OF CONTENTS

1 INTRODUCTION ............................................................................................................. J-1

2 DISSOLVED PHASE CHEMICAL ISOLATION ANALYSIS ............................................ J-4

2.1 Approach ....................................................................................................................... J-4

2.2 Model Inputs ................................................................................................................. J-6

2.2.1 Chemical Isolation Layer Composition and Thickness ......................................... J-6

2.2.2 Groundwater Upwelling Flux ................................................................................. J-6

2.2.3 Sediment Porewater Concentrations.................................................................... J-10

2.2.4 Diffusion Coefficients ............................................................................................ J-13

2.2.5 Partitioning Coefficients ....................................................................................... J-13

2.2.6 Fraction Organic Carbon ...................................................................................... J-15

2.2.7 Sediment Consolidation ........................................................................................ J-15

2.2.8 Input Summary ...................................................................................................... J-19

2.3 Model Results .............................................................................................................. J-21

3 NAPL CONSIDERATIONS ............................................................................................. J-24

3.1 Sediment Core Logs ..................................................................................................... J-24

3.2 Volume of NAPL per Unit Area for Free Phase NAPL Intervals ............................. J-25

3.3 Potential for Ebullition Facilitated Transport of NAPL ............................................ J-31

4 ARMOR LAYER ANALYSIS .......................................................................................... J-32

4.1 Vessel-induced Propwash ........................................................................................... J-32

4.2 Wind-generated Waves .............................................................................................. J-36

4.3 Vessel-generated Waves ............................................................................................. J-36

4.4 Hydrodynamic Flow ................................................................................................... J-37

4.5 Ice Effects ..................................................................................................................... J-37

4.6 Vessel Anchor Drag..................................................................................................... J-37

4.7 Armor Layer Material Selection and Layer Thickness .............................................. J-37

5 FILTER LAYER EVALUATION ..................................................................................... J-39

6 BIOTURBATION ............................................................................................................ J-41

7 CONSOLIDATION ......................................................................................................... J-42

8 GEOTECHNICAL CONSIDERATIONS ......................................................................... J-43

Table of Contents

Lower Rouge River – Old Channel J-ii 130287-04.01

9 CAP DESIGN SUMMARY .............................................................................................. J-45

10 REFERENCES ................................................................................................................. J-47

List of Tables Table J-1 Model Simulation Scenarios .............................................................................. J-12

Table J-2 Chemical-specific Model Input Parameters ..................................................... J-14

Table J-3 Input Parameter Values for the Chemical Isolation Cap Model ..................... J-19

Table J-4 Model Results ..................................................................................................... J-23

Table J-5 NAPL Observations, Volumes, and Saturations ............................................... J-27

Table J-6 Representative Vessel Specifications ................................................................ J-34

Table J-7 Representative Vessel Operating Conditions ................................................... J-35

Table J-8 Propwash Analysis Stable Median Armor Stone Results ................................ J-36

Table J-9 Armor Layer Material Specifications ................................................................ J-38

Table J-10 Graded Gravel Filter Layer Material Specifications ......................................... J-40

Table J-11 Cap A (DU 41) Design Summary ...................................................................... J-45

Table J-12 Cap B (DU 32-34) Design Summary ................................................................. J-46

List of Figures Figure J-1 Cap Configuration and Processes Simulated Figure J-2a Cap A With Representative Sample Locations Figure J-2b Cap B With Representative Sample Locations Figure J-3 Upwelling Velocity Estimation-Cross Section Station 35+32 Figure J-4a Depth Profile for Core OC14-SC-58 in Cap A Figure J-4b Depth Profile for Core OC14-GT-09 in Cap B Figure J-4c Depth Profile for Core OC14-SC-40 in Cap B Figure J-4d Depth Profile for Core OC14-GT-15 in Cap B Figure J-4e Depth Profile for Core OC14-SC-41 in Cap B Figure J-4f Depth Profile for Core OC14-SC-71 in Cap B Figure J-5a Cap A Cross section 57+84 Figure J-5b Cap B Cross Section 33+73 and 34+17

Table of Contents

Lower Rouge River – Old Channel J-iii 130287-04.01

Figure J-5c Cap B Cross Section 35+32 and 36+71 Figure J-6 Settlement and Sediment Consolidation Estimates for Cap A Figure J-7 Settlement and Sediment Consolidation Estimates for Cap B

List of Attachments Attachment J-1 Initial Porewater Concentrations Attachment J-2 Model-predicted Average Porewater Concentrations Within the

Bioturbation Zone Attachment J-3 Cap Armor Analysis – Propeller Wash Attachment J-4 Bearing Capacity Calculations

Lower Rouge River – Old Channel J-iv 130287-04.01

LIST OF ACRONYMS AND ABBREVIATIONS

ΣIWTU sum of target interstitial water toxicity units µg/L microgram per liter

ABM articulated block mat

ARCS Assessment & Remediation of Contaminated Sediments cm centimeter cm/hr centimeter per hour cm/s centimeter per second cm/yr centimeter per year cm2/s square centimeter per second cm2/yr square centimeter per year

DU dredge unit dynes/cm dynes per centimeter

FS Feasibility Study g/cm3 gram per cubic centimeter g/mL gram per milliliter

H:V horizontal to vertical

L/kg liter per kilogram

LWD low water datum m3/m2 cubic meter per square meter

LRROC Lower Rouge River Old Channel

NAPL non-aqueous phase liquid

NAVD88 North American Vertical Datum of 1988

ODOT Ohio Department of Transportation

ORP oxidation reduction potential

PAH polycyclic aromatic hydrocarbon

RAL remediation action level

RAO remedial action objective

List of Acronyms and Abbreviations

Lower Rouge River – Old Channel J-v 130287-04.01

PDI Pre-Design Investigation

PRG preliminary remediation goal

SIC Seepage-induced consolidation

USACE U.S. Army Corps of Engineers

USCS Unified Soil Classification System

USEPA U.S. Environmental Protection Agency yr-1 per year

Lower Rouge River – Old Channel J-1 130287-04.01

1 INTRODUCTION

This report describes the basis of design for an in situ containment cap to chemically and physically isolate contaminated sediments located within designated capping areas of the Lower Rouge River Old Channel (LRROC). Capping was selected in these designated areas to meet remedial action objectives (RAOs) developed for LRROC sediment (Basis of Design Report) by providing long-term protection of the sediment biologically active zone (top 10 centimeters [cm] of sediment) from elevated concentrations of dissolved phase polycyclic aromatic hydrocarbons (PAHs) and non-aqueous phase liquid (NAPL).

The design analyses conducted in support of the 100% Design were performed in accordance with the U.S. Environmental Protection Agency’s (USEPA’s) Guidance for In-Situ Subaqueous Capping of Contaminated Sediments (Palermo et al. 1998) as well as other technical guidance documents referenced where appropriate, which are based on consideration of the following five cap thickness components:

• Chemical isolation of contaminants (Ti)

• Bioturbation (Tb)

• Consolidation (Tc)

• Erosion (Te)

• Operational considerations (e.g., placement inaccuracies, geotechnical filtering; To)

Conceptually, the total required thickness of the cap (T) is calculated for each individual layer based on the design parameters and can be summarized in the following equation (Palermo et al. 1998):

T = Ti + Tb + Tc + Te +To (Equation 1)

However, not every component described in Equation 1 adds to the total cap thickness. For example, as noted in Palermo et al. (2002), the erosion and bioturbation components may be a combined thickness, not requiring independent analysis if it is assumed that the armor layer can resist erosion as well as accommodate bioturbation. In addition, the chemical isolation layer, which is evaluated through contaminant transport modeling, explicitly accounts for bioturbation as well as the effects of consolidation processes on chemical

Introduction

Lower Rouge River – Old Channel J-2 130287-04.01 transport. Therefore, this appendix focuses on the chemical isolation and armor layers of the cap.

The chemical isolation layer of the cap was designed to limit and attenuate long-term transport of PAHs and to address visual NAPL impacts. This layer was designed using a mathematical model of chemical transport within the cap; specification of input parameters for the capping model was based on site-specific and literature-derived information such as the following:

• Sediment and porewater PAH concentration data

• Groundwater upwelling flux calculations and a site-specific surface water quality monitoring survey used to assess preferential flow zones

• Chemical diffusion and partitioning coefficients

• Fraction of organic carbon of sediments

• Consolidation properties of the underlying sediments

A review of available NAPL observations from sediment investigations and an evaluation of NAPL quantity and mobility potential within the capping areas was also conducted and considered in the chemical isolation layer design.

The armor layer of the cap was designed to withstand the natural- and human-generated erosive forces that could potentially affect the stability of the cap. The dominant erosional force anticipated for this cap is propeller wash (propwash) from passing vessels that operate in the area.

The remainder of this appendix has been organized into the following sections, addressing each of the cap design components:

• Section 2: Dissolved Phased Chemical Isolation Analysis

• Section 3: NAPL Considerations

• Section 4: Armor Layer Analysis

• Section 5: Filter Layer Evaluation

• Section 6: Bioturbation

• Section 7: Consolidation

Introduction

Lower Rouge River – Old Channel J-3 130287-04.01

• Section 8: Geotechnical Considerations

• Section 9: Cap Design Summary

• Section 10: References

Lower Rouge River – Old Channel J-4 130287-04.01

2 DISSOLVED PHASE CHEMICAL ISOLATION ANALYSIS

Chemical transport modeling was conducted to design the chemical isolation layer of the cap. This modeling was performed consistent with USEPA guidance (Palermo et al. 1998).

The primary goal of this modeling was to simulate the transport of PAHs within the chemical isolation layer of the sediment cap and specify the composition and thickness of cap material that will provide long-term protectiveness. The modeling evaluation considered a chemical isolation layer of sand amended with organoclay (to increase sorptive capacity) that will be placed as a bulk mixture.

2.1 Approach

The one-dimensional model of chemical transport within sediment caps developed by Dr. Danny Reible (currently at Texas Tech University) was used to predict the time variable transport of PAHs for this evaluation. The model simulates the fate and transport of chemicals (dissolved and sorbed phases) under the processes of advection, diffusion/dispersion, biodegradation, bioturbation/bioirrigation, and exchange with the overlying surface water. This model has been used to support the evaluation and design of sediment caps at numerous sites across the United States, including River Raisin in Michigan (Anchor QEA and Mannik and Smith 2015) and Onondaga Lake in New York (Parsons and Anchor QEA 2012). Details on the model structure and underlying theory and equations are provided in Lampert and Reible (2009), Go et al. (2009), and the USEPA and USACE capping guidance document (Appendix B of Palermo et al. 1998).

For this effort, the model was used to predict porewater concentrations of 33 PAH compounds within the cap by simulating each compound independently. The model results for the 33 PAH compounds were combined to calculate the sum of target interstitial water toxicity unit (∑IWTU)2 values (USEPA 2003) for comparison to the site remediation action level (RAL) of a ∑IWTU value of 5 (Basis of Design Report); these comparisons were made within the bioturbation zone of the simulated cap, which is the upper portion of the cap surface where the majority of benthic activity and exposure would occur.

2 ∑IWTU is the sum of the ratio of individual PAH porewater concentrations to the chemical-specific final chronic value (FCV)

Dissolved Phase Chemical Isolation Analysis

Lower Rouge River – Old Channel J-5 130287-04.01

As discussed in Section 1, the cap design consists of an armor layer placed atop a layer of chemical isolation material, which in turn overlays the native sediment; due to grain size differences between these two materials, a geotechnical filter layer will be placed between them (see Section 5). However, for the purposes of chemical isolation layer modeling, the armor layer was conservatively assumed not to contribute to chemical isolation. Therefore, the cap profile simulated in the model consisted of a minimum of 6 inches of chemical isolation material (i.e., sand/organoclay bulk mixture), overlain by the 6-inch filter layer. In the model, the upper 10 cm of the filter layer was taken to represent the bioturbation zone, which is the depth over which benthic mixing processes are represented to occur. This 10-cm bioturbation zone was incorporated into the model to represent native material that would deposit within the interstitial spaces and on top of the physical containment layer sometime after cap placement. This model approach assumes that the bioturbation zone would be fully established over a relatively short period, compared to the time required for PAHs to be transported through the cap. PAHs in the underlying sediments were simulated based on vertical concentration profiles from sediment cores collected within the capping areas, taking into account the range of dredging depths that will occur prior to cap placement. Figure J-1 illustrates the processes simulated and cap configuration represented in the model.

The overall LRROC sediment remediation project is divided into dredge units (DUs); four of these DUs have been designated as “dredge and cap” areas where some dredging is required to accommodate cap placement to meet USACE navigation requirements. DUs where capping will occur were combined into two capping areas, as shown in Figure J-2 and described as follows:

• Cap A: within DU 41, adjacent to the Superior Materials Cement Plant along the west bank (i.e., Mainland bank) between Stations 57+39 and 58+32 (Figure J-2a)

• Cap B: within DUs 32 to 34, adjacent to the AK Steel water intake structure, along the north bank (i.e., Mainland bank) between Stations 33+31 and 37+01 (Figure J-2b)

These two areas were evaluated separately based on differences in porewater PAH concentrations and groundwater flow characteristics.

Lower Rouge River – Old Channel J-6 130287-04.01

The model was used to identify an appropriate composition and thickness of cap material that would be required to provide long-term protectiveness of the site’s benthic community by maintaining a ∑IWTU within the surface layer of the cap below the target value of 5. For the purposes of this design, long-term protectiveness was defined based on conducting transient (time-variable) model simulations for each of the 33 chemicals contributing to the IWTU over a duration of 100 years, consistent with the evaluation period used at several other capping sites.

2.2 Model Inputs

Specification of input parameters for the chemical transport model was based on site-specific data and information derived from literature and experience with cap designs at other sites.

A description of the approach used to develop the key model input parameters is provided in the following subsections, and a full listing of model inputs is provided in Section 2.2.8.

2.2.1 Chemical Isolation Layer Composition and Thickness

The chemical isolation layer composition (i.e., relative amounts of sand and organoclay in bulk mixture) and thickness were design parameters for this evaluation. Modeling was conducted by initially representing a cap profile consisting of 6 inches of chemical isolation material overlain by a 6-inch filter layer. The percent by weight of organoclay included in the chemical isolation material started with 1% and was increased, as necessary, during model simulations to achieve the target ∑IWTU value of less than 5 at the cap surface (i.e., within the 10-cm bioturbation zone) for more than 100 years. If organoclay percentages became high enough that constructability became a concern, the thickness of the chemical isolation layer was increased as well.

2.2.2 Groundwater Upwelling Flux

Groundwater in the bedrock that underlies the site is under artesian conditions, based on observations during geotechnical investigations performed by the Detroit Water and Sewerage Department for the Modified Detroit River Outfall No.2 project, which crosses the LRROC near the southern end of Zug Island (DWSD 2007). Groundwater flow through a cap is a key process that impacts the effectiveness of a cap, and therefore, it was important to understand if any preferential artesian flow conditions were present in the capping areas, Lower Rouge River – Old Channel J-7 130287-04.01 given that flowing artesian springs have been observed in the area of Zug Island (e.g., EA 2016). To better understand if significant preferential artesian flow areas may be present in the capping areas, an upwelling survey was conducted in August 2015 (EA 2016).

Groundwater is typically characterized by temperature, conductivity, oxidation reduction potential (ORP), and water quality parameters that differ significantly from those of surface water (i.e., groundwater has higher salinity, lower temperature, and higher ORP than surface water). Therefore, surface water quality measurements were made along transects through the capping areas to evaluate whether there were any localized changes in water quality that could signify a preferential artesian flow condition. These measurements were compared with water quality measured from known bedrock groundwater seep locations. Survey results did not indicate any evidence of significant artesian upwelling within the capping areas; details of the water quality probe-based survey are discussed in the Pre-Design Investigation (PDI; EA 2016). Therefore, preferential artesian flow pathways within the capping areas were not considered in the design, and quantification of upwelling through the sediments in the capping areas was performed using a desktop hydrogeological evaluation, as discussed in the remainder of this subsection.

Groundwater upwelling fluxes in the capping areas were estimated using available hydrogeological information in the area of Cap B and the adjacent upland. Only limited adjacent upland data (i.e., one geotechnical boring from the PDI in 2015 [EA 2016]) were available for the Cap A area; therefore, the reasonable assumption was made that groundwater upwelling conditions in that area are similar to those in the area for which the hydrogeological analysis was conducted (the one PDI boring in the Cap A area was generally consistent with the information used for the Cap B area). A representative cross section through the Cap B area is shown in Figure J-3. In the river channel, the sediment is underlain by silty clay. In the upland, overburden, which consists of fill and silty sand, is also underlain by silty clay, which is underlain by bedrock (not shown in Figure J-3).

Groundwater in the bedrock is under artesian conditions, as described previously. As shown in Figures J-2a and J-2b, most of the Cap B area is located directly in front of the AK Steel water intake structure, which prevents the overburden (and any associated groundwater flow) from contacting the sediment based on the intake structure’s drawings (Ford Motor Company 1959), as shown in Figure J-3. Furthermore, a permanent sheetpile wall that will be built as part of the design (LRROC Basis of Design Report) will connect the water intake

Lower Rouge River – Old Channel J-8 130287-04.01 structure with an existing bulkhead, further preventing groundwater within the overburden from flowing through the sediments in the capping areas.

Groundwater flow that could contribute to upwelling within the sediments to be capped includes lateral flow in the overburden, lateral flow in the silty clay, and upward flow from bedrock across the silty clay and sediment. Literature values of hydraulic conductivity for silty clay ranges from 10-9 to 10-6 centimeters per second (cm/s) and from 10-6 to 10-3 cm/s for silty sand (Fetter 2001). Because hydraulic conductivity for silty clay is expected to be 2 to 3 orders of magnitude lower than that of the overburden; lateral flow in the silty clay unit is expected to be negligible compared to lateral flow in the overburden. Therefore, only lateral flow in the overburden and vertical flow from bedrock were considered in this upwelling flux calculation.

Information used in this upwelling flux calculation was obtained from the following sources:

• Sediment/silty clay contact elevations in the channel were obtained from in-channel sediment cores. The average top of silty clay elevation in the proposed capping areas is estimated to be 540 feet (North American Vertical Datum of 1988 [NAVD88]).

• Hydraulic gradient and saturated thickness in the overburden were estimated from Phase 2 PDI geotechnical borings collected adjacent to the cap areas (EA 2016) and upland information at the Honeywell Detroit Tar Site (MACTEC 2009) and Honeywell Detroit Coke Site (MACTEC 2011), which is approximately 1120 feet from the Cap B area. Overburden/silty clay contact elevations were estimated from upland borings and extrapolated from in-channel data.

• The top elevation of bedrock and the potentiometric head in the bedrock was obtained from the geological information in DWSD (2007). The top elevation of bedrock (490 feet) is similar to what was reported at the upland sites in the vicinity, including the Honeywell Detroit Tar Site (MACTEC 2009) and Honeywell Detroit Coke Site (MACTEC 2011). The potentiometric head in the bedrock is 596 feet, reflecting the artesian conditions (i.e., the observed water levels during bedrock drilling recorded in the DWSD [2007] borehole logs were reported to be several feet above ground surface and as artesian conditions).

• The elevations and dimensions of the AK Steel water intake structure were obtained from its drawings (Ford Motor Company 1959).

Lower Rouge River – Old Channel J-9 130287-04.01

Across the Cap B area, groundwater upwelling is considered to consist only of vertical flow from bedrock to surface water, because the AK Steel water intake structure and the proposed permanent sheetpile wall effectively eliminate the potential for lateral flow in the overburden to reach the sediments to be capped. In the Cap A area, however, groundwater upwelling could be a result of both vertical flow from bedrock and lateral flow in the overburden, because no such structures/barrier walls are present in that area; therefore, both the lateral and vertical flows were calculated for that area.

Figure J-3 presents the basis for estimating the upwelling flux within the capping areas. The vertical upwelling flux for the Cap B area was calculated as a Darcy flux (q) using Darcy’s Law (𝑞𝑞 = 𝐾𝐾 × 𝑖𝑖). The hydraulic conductivity (K) value was assigned as the upper range of literature values for clay (10-6 cm/s). This value is higher (i.e., more conservative) than previous laboratory test result of 2×10-7 cm/s performed on a clay sample collected from the Honeywell Detroit Tar Site (MACTEC 2009). The vertical hydraulic gradient (i) was calculated by dividing the difference between the bedrock potentiometric head and the low water datum [LWD] elevation by the average thickness of the silty clay beneath the river channel. Based on these inputs, the calculated upwelling flux is 15.6 centimeters per year (cm/yr). Because hydraulic conductivity is an order-of-magnitude estimate, a Darcy flux of 20 cm/yr was used in the model for the Cap B area.

For the Cap A area, upwelling flux could also be due to discharge of lateral flow in the overburden to the river channel. The upwelling flux due to lateral flow for this area was calculated using Darcy’s law by distributing the lateral flow along a sloped surface

(𝑞𝑞 = 𝐾𝐾×𝑖𝑖×𝑏𝑏

𝑊𝑊

). The hydraulic conductivity (K) value was assigned as the upper range of literature values for silty sand (10-3 cm/s). This value is similar to the reported hydraulic conductivity value of 9×10-4 cm/s obtained from a pumping test at the nearby Honeywell Detroit Coke Site (MACTEC 2011). Conservatively, the horizontal hydraulic gradient (i) was estimated to be 0.01 ft/ft, based on the potentiometric surface at the Honeywell Detroit Tar Site (MACTEC 2009). Saturated thickness in the overburden (b) was estimated to be 20 feet based on the observations at the Honeywell Detroit Coke Site and Detroit Tar Site (MACTAC 2009); this value is conservative because the upland geotechnical borings collected as part of the PDI in 2015 (EA 2016) indicate a higher elevation for the silty clay

Lower Rouge River – Old Channel J-10 130287-04.01 unit (which results in a saturated thickness of overburden of approximately 10 feet, or half the assumed value). The width of the slope surface (w) was estimated to be 40 feet, as shown in Figures J-2a and J-2b. The resulting calculated upwelling flux due to lateral flow is 160 cm/yr. As an approximation, the total estimated upwelling flux in the Cap A area is taken as the sum of the Darcy fluxes due to vertical flow (20 cm/yr) and horizontal flow (160 cm/yr), which results in 180 cm/yr. Inaccuracies associated with the process of adding the vertical and lateral flow is considerably smaller than the uncertainty in the parameter values used in the Darcy flux calculations. Because hydraulic conductivity is an order-of-magnitude estimate, a Darcy flux of 200 cm/yr was used in the model for the Cap A area. Based on the information available, these estimates of Darcy flux are likely conservative, and could be lower, based on the greater thickness of silty clay identified by the upland PDI geotechnical borings (EA 2016).

2.2.3 Sediment Porewater Concentrations

The source term in the model is represented by the concentration of contaminants in the porewater within the sediments beneath the cap. As discussed in Section 2.1, model simulations were conducted based on PAH concentrations measured in sediment cores within each capping area, based on the range of dredging depths that will occur prior to cap placement. Because PAH concentrations varied horizontally and vertically, modeling was conducted by simulating several individual sediment cores. For this evaluation, the full thickness of PAH-impacted sediments based on the measured concentration profile was simulated at each coring location. Bulk sediment data were used to calculate PAH concentrations in sediment porewater throughout the sediments beneath the cap. In general, porewater samples were collected from the top 6 inches of sediments only. Because the full concentration profile is necessary for this analysis (especially because dredging will occur), porewater concentrations in the subsurface were calculated from sediment samples collected in the sub-surface using equilibrium partitioning theory. In the few cases where both sediment and porewater samples were collected from the surface, the calculated porewater concentration from the surface interval was used instead of the measured porewater sample to maintain consistency with the sediment core samples from the sub-surface. Using the calculated porewater is conservative, because calculated porewater concentrations are higher than measured porewater concentrations. Figures J-4a through

Lower Rouge River – Old Channel J-11 130287-04.01

J-4f present PAH concentration depth profiles for each of the coring locations simulated with the model. The left-hand panels of these figures show the bulk sediment total PAH concentrations, the center panels show the total PAH concentrations in porewater, and the right-hand panels show corresponding ∑IWTU values. For some samples, the calculated total PAH concentrations in porewater (center panels) indicate the presence of NAPL (i.e., they exceed effective solubility for PAHs), which is consistent with the visual observations of NAPL (sheens and free phase) noted in the core logs (EA 2016). Simulating these concentrations with the model is conservative. Based on calculations presented below, it is unlikely that NAPL will migrate upward as a separate phase and enter the base of the cap.

Dredging in portions of the capping areas is necessary to maintain navigation depth requirements. Therefore, sediment porewater concentration profiles used in the model were adjusted to account for this dredging (i.e., the porewater concentration profile simulated in the model begins at the depth corresponding to the post-dredge surface). Cores collected within the dredge areas are at a relatively low spatial density compared to the variability in dredge depths (Figures J-5a through J-5c). Thus, the area represented by each core consists of a range of dredge depths. To capture this range of dredge depths in the model evaluation, a bounding approach consisting of modeling sediment porewater concentrations starting at a sampling depth equal to the minimum dredge depth and the maximum dredge depth within the area near a core was taken. The specific dredge depths evaluated for each core were as follows:

• One core was collected in the Cap A area (OC14-SC-58; Figure J-4a). Dredging depths in this area range from 5 to 10 feet as shown in Figure J-5a. As such, the concentrations from a depth of 5 feet (post-dredge surface) to the bottom of the core at 8 feet were represented in the model (simulated 3 feet of sediment). Scenarios of deeper dredge depths were not conducted, because 5 feet of dredging was conservative (i.e., concentrations in sediment decrease significantly with depth after 5 feet).

• Concentrations from five core sampling locations were used to evaluate the range of conditions within the Cap B area:

− Dredge depths at OC14-GT-09 (Figure J-4b) range from no dredging to 10 feet of dredging (Figures J-5b). As such, the porewater PAH concentrations represented

Lower Rouge River – Old Channel J-12 130287-04.01 in the model were specified based on the existing profile (no dredging) and 10 feet (maximum post-dredge surface) based on cross section 33+73 and 34+17 (Figure J-5b).

− Dredge depths at core locations OC14-SC-40 (Figure J-4c), OC14-GT-15 (Figure J-4d), OC14-SC-41 (Figure J-4e), and OC14-SC-71 (Figure J-4f) range from no dredging to 5 feet of dredging based on cross sections 34+17 (Figure J-5b), 35+32 (Figure J-5c), and 36+71 (Figure J-5c). As such, initial concentrations in porewater represented in the model for these cases were specified based on the existing profile (no dredging) as well as from a depth of 5 feet (post-dredge surface) to the bottom of each core. In each core, the range of assumed dredge depths encompass a conservative case. The conservative case is considered when a sample contains high concentrations, or contains a higher proportion of the more mobile chemicals that would transport into the cap faster.

A summary of dredge depths and coring locations modeled is provided in Table J-1.

Table J-1 Model Simulation Scenarios

Cap Area Core Location Core Depth

(feet) Dredge Depth

(feet) Simulated Sediment

Thickness (feet) Cap A OC14-SC-58 8 5 3

Cap B

OC14-GT-09 24

0 24

10 14

OC14-GT-15 20

0 20 5 15

OC14-SC-40

0 12 5 7

OC14-SC-41 16

0 16 5 11

OC14-SC-71 11

0 11 5 6

Attachment J-1 lists the individual PAH porewater concentrations used as model inputs for each coring location within the capping areas evaluated with the model, for all 33 PAH compounds simulated. The source of chemicals within the sediment was specified as finite in

Lower Rouge River – Old Channel J-13 130287-04.01 the model, indicating that the mass of chemicals can deplete from the sediments as the chemicals migrate out of the sediment into the cap over the long-term periods simulated by the model. This approach is appropriate, because concentrations decrease significantly at the bottom of the cores (i.e., the full inventory of PAH mass is represented by the model and there is no additional source of PAHs beneath the sediment).

2.2.4 Diffusion Coefficients

The molecular diffusivity of each PAH compound is a required model input parameter. To obtain values for the full suite of PAH compounds included in the modeling, the correlation identified from Schwarzenbach et al. (1993), relating diffusivity to a compound’s molecular weight, was used. The values derived from this approach were found to be consistent with values tabulated in or derived from other literature sources (e.g., Hayduk and Laudie 1974;

USEPA 1996). Diffusivity values for each PAH compound are presented in Table J-2.

2.2.5 Partitioning Coefficients

Partitioning of chemicals between the dissolved and sorbed (i.e., cap material) phases is described in the model by the chemical-specific equilibrium partition coefficient (Kd). This approach assumes that sorption follows linear isotherms and is instantaneous (not rate-limited) and reversible. For nonpolar organic compounds, such as PAHs, the partition coefficient is calculated in the model based on the customary Kd = fOC*KOC approach (e.g., Karickhoff 1984), where KOC is the compound’s organic carbon partition coefficient and fOC is the organic carbon content of the solid phase (i.e., sediment or cap material). Separate partitioning coefficients were developed for the different cap isolation layer materials as follows:

• For sand and sediment, KOC was set to a literature-based value for each of the 33 PAH compounds modeled and Kd was calculated by the model, as described above based on the KOC and the fOC of the cap material or sediment (see below for details on fOC).

• For modeling partitioning of PAH compounds to organoclay-amended sand (i.e., the chemical isolation layer), the Kd was calculated as a function of the compound’s octanol-water partition coefficient (KOW), using the relationship developed by Reible et al. (2011), which was based on laboratory studies of sorption of PAHs onto organoclay. The equation was modified to account for the percent weight of

Lower Rouge River – Old Channel J-14 130287-04.01 organoclay within the bulk mixture (which essentially ignores the sorptive capacity of the sand in the mixture), as described by Equation 2:

𝑙𝑙𝑙𝑙𝑙𝑙 𝐾𝐾𝑑𝑑 = %Organoclay ∗ (1.20 ∗ 𝑙𝑙𝑙𝑙𝑙𝑙 𝐾𝐾𝑂𝑂𝑊𝑊 – 0.467) (Equation 2)

Site-specific partitioning coefficients were not available; therefore, conservative values of KOC and KOW for each of the 33 PAH compounds modeled were obtained from USEPA (2003); these values are also listed in Table J-2.

Table J-2 Chemical-specific Model Input Parameters

PAH Water Diffusivity (1E-06 cm2/s)1 log KOC (L/kg)2 log KOW (L/kg)2

1-Methylnaphthalene 8.0 3.8 3.8

2-Methylnaphthalene 8.0 3.8 3.9

Acenaphthene 7.5 3.9 4.0

Acenaphthylene 7.6 3.2 3.2

Anthracene 6.8 4.5 4.5

Benz(a)anthracene 5.7 5.6 5.7

Benzo(a)pyrene 5.3 6.0 6.1

Benzo(b)fluoranthene 5.3 6.2 6.3

Benzo(e)pyrene 5.3 6.0 6.1

Benzo(ghi)perylene 5.0 6.4 6.5

Benzo(k)fluoranthene 5.3 6.2 6.3

C1-Chrysenes/benz(a)anthracenes 5.5 6.0 6.1

C1-Fluoranthenes/pyrenes 5.9 5.2 5.3

C1-Fluorenes 6.8 4.6 4.7

C1-Phenanthrenes/anthracenes 6.4 5.0 5.0

C2-Chrysenes/benz(a)anthracenes 5.2 6.3 5.5

C2-Fluorenes 6.4 5.1 5.2

C2-Naphthalenes 7.5 4.2 4.3

C2-Phenanthrenes/Anthracenes 6.1 5.4 5.5

C3-Fluorenes 6.1 5.6 5.7

C3-Naphthalenes 7.0 4.7 4.8

C3-Phenanthrenes/Anthracenes 5.9 5.8 5.9

C4-Naphthalenes 6.7 5.2 5.3

Lower Rouge River – Old Channel J-15 130287-04.01

PAH Water Diffusivity (1E-06 cm2/s)1 log KOC (L/kg)2 log KOW (L/kg)2 C4-Phenanthrenes/Anthracenes 5.6 6.2 6.3

Chrysene 5.7 5.6 5.7

Dibenz(a,h)anthracene 5.0 6.6 6.7

Fluoranthene 6.2 5.0 5.1

Fluorene 7.1 4.1 4.2

Indeno(1,2,3-cd)pyrene 5.0 6.6 6.7

Naphthalene 8.6 3.3 3.4

Perylene 5.3 6.0 6.1

Phenanthrene 6.8 4.5 4.5

Pyrene 6.2 4.8 5.0

Notes:

cm2/s = square centimeters per second L/kg = liters per kilogram 1 Calculated for each PAH based on the molecular weight of the compound using the correlation identified in

Figure 9.7b of Schwarzenbach et al. (1993).

2 Based on values reported in USEPA (2003).

2.2.6 Fraction Organic Carbon

The fOC is used in the model for both the underlying sediments and the bioturbation zone.

The fOC of the sediments was set to 4.14%, which is the average organic carbon content measured in the samples from the six sediment cores collected within the capping areas. The fOC in the bioturbation zone was specified based on the assumption that sediments having an organic content similar to that of the existing surface sediments would settle on the surface of the cap over time. Within the LRROC, foc in surface sediments (generally top 1 foot) ranges from 0.1 to 25%; within the Cap A and Cap B areas the fOC measured in the one sample collected from the top 1 foot from was 1.3%. Based on limited data, the bioturbation zone fOC was conservatively set to 1%. For modeling organoclay-amended sand, the fOC was not used, because partitioning was based on a Kd value that depends on the compound’s KOW and the percent organoclay in the bulk mixture, as described in the previous subsection.

2.2.7 Sediment Consolidation

Consolidation of the underlying sediment due to the weight of the cap layers was considered in the model, because the consolidation process causes the expulsion of porewater from the

Lower Rouge River – Old Channel J-16 130287-04.01 sediment voids into the cap’s chemical isolation layer, where it can be transported upward to the surface water. To estimate the effects of such sediment consolidation, a cap-induced settlement analysis was performed.

The consolidation characteristics of the sediments were based on the results of consolidation tests performed on sediment samples collected during the PDI, including traditional oedometer tests (ASTM D2435) conducted on samples from DUs 13, 15, and 25 as well as seepage-induced consolidation (SIC) tests (Znidarci, et al. 1992) conducted on samples from DU 34. Consolidation estimates were evaluated for both Cap A and Cap B, assuming the maximum allowable cap thickness, including the full armor layer and over-placement allowances (63 inches and 47 inches, respectively).

Settlement calculations were performed by developing a representative subsurface profile to depict the anticipated post-dredge conditions. The Cap A settlement evaluation focused on the subsurface conditions found in DU 41, which were evaluated based upon the observations from OC14-SC-58. The generalized existing subsurface conditions observed at OC14-SC-58 include a surficial layer of sand, gravel, and silt to a depth of 6 feet, overlying soft clay (lacustrine clay) with trace amounts of sand and gravel that extended to the bottom of the exploration (9.8 feet). Based on deeper explorations performed across the site, the clay layer was assumed to extend to a depth of 45 feet. For DU 41, between 5 and 10 feet of dredging is anticipated prior to cap placement. No oedometer or SIC test data were available within DU 41. The bearing subgrade of the cap in the vicinity of DU 41 is believed to be lacustrine clay based on the anticipated dredge depth and subsurface observations from OC14-SC-58. Therefore, oedometer data collected from other sampling locations at varying depths within the lacustrine clay were used to estimate potential settlement.

The Cap B settlement evaluation focused on the subsurface conditions found in DU 34, specifically represented by the subsurface conditions observed in OC14-GT-15. The existing conditions at OC14-GT-15 include a surficial layer, approximately 4 feet thick, consisting of a mixture of very soft, clay, and fine silty sand. Beneath the surficial layer is a deposit of very soft, fat clay with some fine sand, extending to a depth of approximately 18 feet. Beneath the deposit of fat clay, a lacustrine deposit of soft lean clay was observed. The lacustrine deposit

Lower Rouge River – Old Channel J-17 130287-04.01 extends to the bottom of the exploration. Sediment removal in DUs 32, 33, and 34 is anticipated to vary and may range from 0 to 10 feet prior to cap placement.

The available oedometer and SIC test samples were selected to be representative of the anticipated range of parameters for a given soil stratum. An estimate of the primary consolidation settlement was then calculated for each stratum based upon the assumed unit weight and thickness of the capping material. The total primary settlement was then estimated by summing the calculated settlement for each soil stratum.

In areas where dredging will be performed prior to cap placement, the reduction in stress on the subsurface sediments will be dependent on the thickness of the dredge cut and unit weight of the material removed. For DU 41, the dredge cut thickness was assumed to range from 5 to 10 feet to accommodate the 63-inch-thick cap. Because the reduction in stress from dredging is nearly equal to the induced stress expected from cap placement, settlement estimates for Cap A are much less than that of Cap B.

For the settlement estimates presented for Cap B, no dredging was…