S01_2 Conneautville Final Geotechnical Report Rev 1.pdf

FINAL GEOTECHNICAL REPORT (REVISION 1)

Conneautville Dam NRCS Task Order AG-SPEC-D-16-0006 Conneautville, Pennsylvania

Schnabel Reference 15715008.00 February 28, 2018

Prepared For:

U.S. Department of Agriculture – Natural Resources Conservation Service

AG-3A75-C-12-0008

NRCS Nationwide A&E Services (IDIQ)

Golder Associates Inc. Schnabel Engineering, LLC

Freese and Nichols, Inc. Wilson & Company, Inc., Engineers & Architects

Mr. Donald R. Murray USDA – NRCS – Bloomsburg Technical Office 702 Saw Mill Road, Suite 205 Bloomsburg, PA 17818-7728

Subject: NRCS Task Order AG-SPEC-D-16-006, Conneautville Dam, Final Geotechnical Report (Revision 1), Conneautville, Pennsylvania (Schnabel Reference 15715008)

Dear Mr. Murray:

GSFW ENGINEERING JV (GSFW) is pleased to submit our Final Geotechnical Report (Revision 1) for this project. This report was prepared in accordance with our approved Plan of Work, dated March 25, 2016. We submitted the Detail Design package on August 28, 2017, and we received comments from NRCS on December 6, 2017. We submitted a response to NRCS on January 10, 2018, stating that we would submit a revised report. This revised report incorporates agreed upon responses to NRCS review comments.

We appreciate the opportunity to be of service for this project. Please call us if you have any questions regarding this report.

Sincerely, GSFW Engineering JV

Gerald C. Robblee, PE Senior Associate

GCR:JRY:PIW:hcf

Distribution:

NRCS (email) Attn: Mr. Donald R. Murray

FINAL GEOTECHNICAL REPORT (REVISION 1)

Conneautville Dam NRCS Task Order AG-SPEC-D-16-0006 Conneautville, Pennsylvania

Schnabel Reference 15715008.00

Gerald C. Robblee, PE PA Professional Engineer PE084781

GSFW Engineering JV 1380 Wilmington Pike, Suite 100, West Chester, PA 19382 Phone (610) 696-6066; Fax (610) 696-7771

PA NRCS

Conneautville Dam – Final Geotechnical Report (Revision 1)

February 28, 2018 Page i GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

FINAL GEOTECHNICAL REPORT (REVISION 1)

CONNEAUTVILLE DAM

NRCS

CONNEAUTVILLE, PENNSYLVANIA

TABLE OF CONTENTS

1.0 INTRODUCTION

1.1 Background

1.2 Purpose

1.3 Scope of Services

1.4 Elevation Datum

2.0 SITE AND PROJECT DESCRIPTION

2.1 Site Description

2.2 Project Description

2.3 Regional and Site Geology

2.4 Regional and Site Seismicity

3.0 FIELD INVESTIGATION PROGRAM

3.1 General

3.2 Previous Investigations

3.3 2016 Design Rehabilitation Design Explorations

4.0 GEOTECHNICAL LABORATORY INVESTIGATIONS

4.1 General

4.2 Previous Investigations

4.3 2016 Design Rehabilitation Design Laboratory Testing

5.0 SUBSURFACE CONDITIONS

5.1 Geologic Profile

5.2 Hydrogeologic Conditions

6.0 GEOTECHNICAL DESIGN CONSIDERATIONS

6.1 General

6.2 Dispersive Soils

6.3 Auxiliary Spillway Erosion

6.4 Seepage and Static Slope Stability

6.5 Seismic Performance

7.0 LIMITATIONS

8.0 REFERENCES

February 28, 2018 Page ii GSFW Engineering JV

LIST OF TABLES

Table 3.1: Summary of 1978 Field Investigations Table 6.1: Summary of Pre- and Post-Lime Augmented Atterberg Limits Table 6.2: Summary of Strength Properties for Stability Analyses Table 6.3: Summary of Static Slope Stability Analyses Table 6.4: PGA for the Conneautville Dam Site Table 6.5: Summary of Pseudo-Static Slope Stability Analyses

LIST OF FIGURES

Figure 1: Location Map Figure 2: Site Vicinity Map Figure 3: Local Geological Map Figure 4: Soil Map Figure 5: Earthquakes Within 100Km of Conneautville Dam Figure 6: Seismic Zones near Conneautville Dam Figure 7: Subsurface Exploration Plan Figure 8: Dam Cross Section 1 Figure 9: Dam Cross Section 2 Figure 10: PGA Curve Figure 11: 2008 PSHA Deaggregation

APPENDICES

Appendix A: Subsurface Exploration Data Appendix B: Soil Laboratory Test Results Appendix C: Headcut Erodibility Index Calculations Appendix D: Seepage and Static Slope Stability Analysis

February 28, 2018 Page 1 GSFW Engineering JV

1.0 INTRODUCTION

1.1 Background

Conneautville Dam is located in Conneautville Borough, Crawford County, Pennsylvania. Conneautville Dam is situated on Thatcher Run about 30 miles southwest of Erie, Pennsylvania. A Location Map is included as Figure 1. The original dam was constructed between 1983 and 1984. A project to improve the dam is currently in the design phase and is being funded through the Natural Resources Conservation Service (NRCS). The objectives of this project are to rehabilitate the dam to meet or exceed state and NRCS dam safety criteria, and to maintain or enhance the current level of flood protection. Overall this project intends to increase the life expectancy of the dam by 50 years.

1.2 Purpose

The purpose of this report is to present the results of the geotechnical field and laboratory explorations conducted by GSFW Engineering JV (GSFW) and to provide pertinent information for the rehabilitation design of Conneautville Dam. This report describes field and laboratory investigation methods and presents the results of those investigations. Note that this report combines several other reports required by the Design Specifications as described in our Plan of Work, dated March 25, 2016. We updated the Final Geotechnical Report, dated March 8, 2017, to address NRCS comments, and revised text is shown in red.

Descriptive nomenclature for dams is based upon one looking downstream in the direction of flow.

“Right” and “left” are referenced in this manner. The reservoir side is known as the upstream side with the opposite slope referred to as the downstream slope.

1.3 Scope of Services

The geotechnical investigation procedures were described in detail in the plan titled Final Geotechnical Investigation and Outlet Inspection Plan, prepared by GSFW and dated May 6, 2016. The laboratory testing program was described in detail in the plan titled Laboratory Investigation Plan, prepared by GSFW JV and dated September 2, 2016. Sections 3.0 and 4.0 provide more details on the field and laboratory investigations, respectively. The results of the investigations are summarized in Section 5.0 and were used as input into our geotechnical analyses performed during this phase of the project. Our analyses and results are described in Section 6.0.

1.4 Elevation Datum

The original design used the National Geodetic Vertical Datum of 1929 (NGVD29) as its datum. Elevations in this report are in feet and referenced to the North American Vertical Datum of 1988 (NAVD88). A vertical datum shift of -0.564 ft was used to convert original design elevations to the NAVD88 datum. The acronym “EL” represents elevation.

February 28, 2018 Page 2 GSFW Engineering JV

2.0 SITE AND PROJECT DESCRIPTION

2.1 Site Description

The Conneautville Dam is an earthen embankment with internal chimney and blanket drains. The dam is about 650-ft long with maximum height of about 40 ft. The drainage area is about 261 acres with its headwaters in Spring Township. Flows pass through the spillways and drain through the center of Conneautville Borough to the Conneaut Creek. The dam has an auxiliary spillway excavated onto natural soils in the left abutment. The natural soils, used to construct the embankment, are considered to potentially be dispersive clays and silts. The foundation consists of silts and clays. The top of the dam is varied from about EL 1001 to EL 1002 (NAVD88), and the pool is maintained empty except during storms. A Site Vicinity Map is included as Figure 2.

2.2 Project Description

The proposed project includes auxiliary spillway repair, embankment dam surface treatment, impact basin improvements, and filter diaphragm installation. The current dam does not meet all applicable NRCS and Pennsylvania Department of Environmental Protection (PADEP) dam safety and performance criteria, and therefore improvements are necessary. These improvements will extend the useful life of the dam to 50 years and improve the safety for the local population.

2.3 Regional and Site Geology

The Conneautville Dam is located in the Northwestern Glaciated Plateau Section of the Appalachian Plateau Province. Periods of glaciation occurred in this area from about 780,000 to 15,000 years ago.

The Laurentide ice sheet advanced and retreated over the Northwestern Glaciated Plateau Section during that time period. In particular, Northwestern Pennsylvania was glaciated by the Erie lobe of the Laurentide ice sheet and one of its sublobes, the Grand River lobe. The Grand River lobe advanced and spread out from the lowland of the Grand River valley in Ohio. The Kent and especially the Defiance Moraines outline the lobe very well. These periods of glaciation carved the landscape of Northwestern Pennsylvania resulting in the present day topography (PADCNR & Fleeger et al., 2005).

The Northwestern Glaciated Plateau Section consists primarily of broad, rounded uplands cut by long, linear valleys. The uplands have a southeast-oriented linearity that is pronounced in eastern Erie and central Crawford Counties. Elsewhere, upland linearity is obscure to absent. The uplands are cut by many flat-floored, narrow to wide valleys that are separated from adjacent uplands by steep slopes on one or both sides of the valley. These U-shaped valleys were cut by glaciers during periods of glaciation. The valleys are very linear and are oriented northwest-southeast for the most part, although some valleys are normal to this orientation. The valley floors are often wetlands. There is frequently a considerable depth of unconsolidated material beneath the valley floor. Glacial tills and lacustrine deposits are common throughout the region and were formed during glaciation. Local relief between valley floor and the top of an adjacent upland may be up to 600 ft, but is generally less than 100 ft. Elevation ranges from 900 to 2,200 ft. The drainage pattern is dendritic. Bedrock, which is largely covered by glacial deposits, consists of a variety of sandstones, siltstones, and shales, as well as some conglomerates and coal. Devonian Shale and Devonian Siltstone are present around the city of Conneautville, and throughout the region.

Rock dated from the Mississippian and Pennsylvanian, subdivisions of the Carboniferous period, is also present. Bedding in the rocks is horizontal. Many of these formations are relatively soft and were easily

February 28, 2018 Page 3 GSFW Engineering JV eroded into linear landforms by the continental glaciers. The local geology is shown on Figure 3 and a portion of the NRCS soils survey map is shown on Figure 4 (PADCNR & Fleeger et al., 2005).

2.4 Regional and Site Seismicity

2.4.1 General

A literature review of regional seismicity was performed to identify seismogenic sources in the area and locations of significant past earthquakes. Figure 5 shows the locations of magnitude 4 (M4.0) or greater earthquakes within a 100-km radius of the dam site that have been recorded since 1900 and are documented by the USGS. These events include:

M5.0 – 1/31/1986 – Pease Mill, Ohio M4.8 – 9/25/1998 – Adamsville, Pennsylvania M4.0 – 12/31/2011 – Youngstown, Ohio

All of the above earthquakes can be attributed to the Northeast Ohio Seismic Zone which is shown in Figure 6. Although we have not identified any surficial faults in the area that have been active during the Quaternary Period, there are several nearby source zones. A source zone is an area where earthquakes have been recorded in the past, but where they cannot be attributed to any specific fault. This type of earthquake hazard is common in the Central and Eastern United States and intraplate seismic regions.

Figure 6 shows two source zones near the site that have had a history of seismic activity: the Northeast Ohio Seismic Zone and Niagara-Attica Seismic Zone.

The Northeast Ohio Seismic Zone stretches from Cuyahoga County, Ohio, to the Pennsylvania border and from Lake Erie down to Akron, Ohio. Over 100 earthquakes have been recorded in the vicinity of the source zone from 1951 to the present. The largest being the January 31, 1986, earthquake which caused the Perry Nuclear Power Plant to shut down. The seismicity can be attributed to a system of ancient basement faults with no surficial expressions. Although it has not produced structurally damaging earthquakes, given the large population and industry in the area, it is a source zone that seismologists continue to study.

The Niagara Attica Seismic Zone is beyond 100 km from the Conneautville Dam but is worth mentioning given its past seismic activity. The source zone underlies much of Lake Ontario and spans from southern Ontario to western New York. The largest acknowledged event occurred in 1929 near Attica, New York, and is thought to be an M4.9 event. Fault lineaments are known to exist in the area, but it is difficult to determine if they are the cause of the earthquakes.

The recent rise of induced seismicity should also be mentioned when discussing the seismic hazard at the Conneautville Dam. Although the link between deep wastewater disposal and increase in seismic activity is debatable, an increase in seismic activity has been recorded in northeastern Ohio. Many scientists believe the series of 2011 Youngstown earthquakes and the 2004 Ashtabula earthquake may have been the result of fluid injection activities.

February 28, 2018 Page 4 GSFW Engineering JV

3.0 FIELD INVESTIGATION PROGRAM

3.1 General

Field investigations were performed during the original design and construction of the dam (circa 1978- 84), during 2004 and 2005 evaluations performed by Gannett Fleming, during the 2010 planning studies performed by Bioengineering Group, and as part of the current GSFW scope of services. A brief summary of each investigation plan is provided below.

3.2 Previous Investigations

3.2.1 Original Design & Construction Phase Explorations

During the original design phase of the project in 1978, field and laboratory geology and soils investigations were performed. The field investigation included both test borings and test pits within the dam footprint, the auxiliary spillway footprint, proposed borrow areas, the reservoir area, and downstream of the dam. The number and type of investigation performed as part of the original design work are summarized below. Details of the field investigations were obtained from the NRCS geology reports included in the Design Folder provided to GSFW in PDF format. The geology reports were referenced as Geology Investigation – Bruce Benton 7/8/78 and Geology Drawings Sheets 1 of 9 – 9 of 9.

Table 3.1: Summary of 1978 Field Investigations

Location No. of Test Borings

No. of Test Pits

Embankment 17 5

Auxiliary Spillway 2 7

Reservoir 5 4

Downstream of Dam 0 8

Borrow Areas 0 11

The test borings and test pits primarily encountered fill soils, glaciofluvial soils, and glaciolacustrine soils.

The design of the project included excavation and removal of the fill soils and glaciofluvial soils from the dam footprint for seepage, settlement, stability, and seismic liquefaction concerns. The glaciolacustrine deposit consists of medium stiff to stiff, low to medium plasticity Lean Clay (CL) and low plasticity to non-plastic Silt (ML) soils.

During construction of the embankment, some embankment cracking and settlement of the principal spillway conduit were observed. To evaluate the impacts of the settlement and cracking on embankment stability, three test borings were performed into the recently compacted fill prior to completing fill placement. The result of the exploration indicated the fill soils were primarily medium plasticity lean clay (CL) soils with estimated percent compaction values often below the specified minimum value of 98% of the soil’s maximum dry density as determined by ASTM D698. Details of the field investigations performed during construction are summarized in the NRCS trip reports and internal memos included in the Design Folder provide to GSFW in PDF format.

February 28, 2018 Page 5 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

3.2.2 2004 & 2005 Gannett Fleming Investigations

Gannett Fleming used a “plug sampler” to collect six samples of embankment soils in 2004 during a site visit with staff from the PADEP (Gannett Fleming, 2004). In 2005, Gannett Fleming performed four test pits in the downstream slope to collect samples of embankment fill, blanket drain filter sand, and drain fill (Gannett Fleming, 2005).

3.2.3 2010 Planning Study Explorations

Three test borings were performed during the 2010 planning studies by the Bioengineering Group to evaluate the soil conditions in the auxiliary spillway. Bioengineering Group Test Borings B-1 through B-3 (BIO B-1 through BIO B-3) ranged in depth from 36 ft to 52 ft. The soils encountered were described as either silt clays or clayey silt. Sandy cohesionless soils were encountered in BIO B-1 and BIO B-2 at depths of 20 ft and 25 ft, respectively. The zone of sandy soils was logged as 15-ft thick in BIO B-1 and 7-ft thick in BIO B-2.

3.3 2016 Design Rehabilitation Design Explorations

3.3.1 General

The field investigation program conducted by GSFW included 4 test borings and 24 test pits performed by GSFW subcontractor Parratt-Wolff of Lewisburg, Pennsylvania, from May 31 to June 8, 2016. Test borings were advanced using a CME-850 track-mounted drill rig, and test pits were excavated using an IHI-33 NX Mini Excavator. None of the test borings reached rock and instead terminated in foundation soils. The locations of the test borings and test pits are shown on Figure 7.

3.3.2 Test Pits

Twenty four test pits were excavated to collect samples from the upstream and downstream embankment slopes and from the auxiliary spillway for laboratory testing. Test pits were about 2-ft deep and from each a 1-gallon bag sample and 5-gallon bucket sample were taken.

The test pits were backfilled by placing about 1 ft of excavated soil back in the pit, tamping with the back of the backhoe bucket, and repeating until the test pit was filled. Topsoil was placed in the uppermost layer and was supplemented with additional topsoil as needed. Each test pit area was fine graded and seeded with straw placed as mulch prior to demobilizing from the site.

3.3.3 Test Borings

Two arrays of two borings were advanced through the dam and into the foundation. Drilling through soil was performed using a 3¼-inch inside diameter hollow stem auger. Two borings were located at the crest of the dam and two borings were located about 30-ft downstream at about mid slope. Three of the borings (B-1, B-3, and B-4) were advanced from 19 ft to 28.5 ft into the foundation, and B-2 was advanced 103 ft into the foundation. GSFW field staff logged the borings and described soils using the Unified Soil Classification System descriptive system, in general accordance with ASTM D2488.

February 28, 2018 Page 6 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

Standard Penetration Tests (SPTs) were performed at the ground surface, at a depth of 5 ft and every 5 ft thereafter. The SPTs were performed in general accordance with ASTM D1586 to collect soil samples.

The number of blows of a 140-pound hammer falling 30 inches required to drive the split spoon sampler four consecutive six-inch increments was recorded. The total number of blows required to drive the sampler one foot from a penetration value of 6 inches to 18 inches is referred to as the N-value. The SPT hammer blow counts presented on boring logs and profiles are the blow counts as recorded in the field and have not been corrected for hammer efficiency or other factors. Continuous sampling was performed through the blanket drain. Shelby tube (undisturbed) sampling was performed in embankment materials and in the foundation. In general, each undisturbed sampling interval consisted of pushing two Shelby tubes to increase the likelihood of obtaining sufficient samples for laboratory testing.

Water levels were measured in each boring during and after completion of drilling operations. Measured water levels are summarized in Section 4.0 and are presented on the boring logs in Appendix A.

Boreholes were grouted upon completion using tremie grouting methods and a cement/bentonite grout.

When the internal drain system was encountered, the borehole was backfilled with fine drain fill. Details on borehole abandonment are included on the boring logs.

February 28, 2018 Page 7 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

4.0 GEOTECHNICAL LABORATORY INVESTIGATIONS

4.1 General

Laboratory investigations were performed during the original design and construction of the dam, during the 2010 planning studies performed by Bioengineering Group, and as part of the current GSFW scope of services. A brief summary of each laboratory investigation is provided below.

4.2 Previous Investigations

4.2.1 Original Design & Construction Phase Laboratory Testing

The laboratory testing performed during the original design phase of the dam included:

Atterberg Limits Moisture content Particle size distribution Crumb Dispersion Moisture-density relationship with standard energy One-dimensional consolidation Consolidated Undrained triaxial compression w/pore pressure measurements (CU')

Details of the testing can be found in three soils mechanics reports either partially included or included in the Design Folder provided to the GSFW by the NRCS. The soils mechanics reports were referenced as:

Original Soils Report (NRCS, May 4, 1979) Soils Report (NRCS, Aug. 30, 1979) Supplemental Report (NRCS, Oct. 23, 1979)

During construction, three test borings were performed in the partially constructed embankment as discussed above. Results are in the Design Folder. Laboratory testing on drive tube (split spoon) samples and tube samples collected from the test borings included:

Moisture content Density Particle size distribution Atterberg Limits Crumb Dispersion One-dimensional consolidation Unconsolidated Undrained triaxial compression (UU) Consolidated Undrained triaxial compression w/pore pressure measurements (CU')

4.2.2 2004 & 2005 Gannett Fleming Lab Testing

The laboratory testing performed by Gannett Fleming in 2004 on the samples collected with the plug sampler included crumb dispersion tests. The lab testing performed by Gannett Fleming in 2005 included four particle size distribution tests of fine drain fill, four particle size distribution tests on coarse drain fill, February 28, 2018 Page 8 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved three crumb dispersion tests, and three no erosion filter tests. The results of the testing are included in the appendices of the 2004 and 2005 Gannett Fleming reports (Gannett Fleming, 2004; and Gannett Fleming, 2005).

4.2.3 2010 Planning Study Laboratory Testing

As part of their planning study, Bioengineering (BIO, 2010) performed the following types of laboratory tests:

Moisture content Density Particle size distribution Atterberg Limits Unconsolidated Undrained triaxial compression (UU)

4.3 2016 Design Rehabilitation Design Laboratory Testing

GSFW performed laboratory tests on soil samples obtained from samples collected during the field investigation. The results of geotechnical index testing and mechanical testing of soils are located in Appendix B.

Laboratory testing was performed on samples obtained from the 24 test pits to characterize the variability of the near surface soils on the dam and within the auxiliary spillway. The laboratory testing included:

Performing 24 sets of index tests to include particle size distribution (with hydrometer), water content, and Atterberg Limits to evaluate the variability of the index properties of surficial site soils.

Performing 6 specific gravity tests.

Performing moisture density relationship testing on 6 soil samples to estimate required remolding densities to be used in pinhole dispersion tests.

Performing 6 pinhole dispersion tests on untreated surficial soils to evaluate the level and variability of soil dispersion of the untreated surficial soils.

In addition to the above testing, additional testing will be performed as part of the design phase of the project. The additional testing includes:

Performing 4 pinhole dispersion tests on lime-treated soils. The soil samples that were lime-treated were of the same soil layer that was classified as potentially dispersive based on the results of the pinhole dispersion tests on test pit samples described above. The 4 pinhole tests on the lime-treated soils were intended to measure the effect of the lime treatment on this potentially dispersive material.

Performing up to 3 sets of agricultural soil tests to evaluate if additional soil amendment will be needed to enable the lime-treated soils to support vegetation growth.

Performing a limited number of Atterberg Limit tests on lime-treated soils to characterize how the lime treatment affects a treated soil’s plasticity.

February 28, 2018 Page 9 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

Laboratory testing of samples obtained from the four test borings to characterize the embankment, drain, and foundation materials included the following:

Performing 24 sets of index tests to include Atterberg Limits, particle size distribution (with hydrometer), specific gravity, and crumb tests.

Performing 28 water content tests on split spoon samples collected as part of SPT tests.

Performing 4 sets of particle size distribution tests to evaluate the gradation of the blanket drain.

Performing 3 sets of consolidated undrained triaxial compression tests (CU') with pore pressure measurements on Shelby tube samples of embankment materials. Each set of specimens also had Atterberg Limits, particle size distribution (with hydrometer), water content, and specific gravity tests performed.

Performing consolidated undrained triaxial compression tests (CU') with pore pressure measurements on a Shelby tube sample of foundation materials and a direct shear test on a Shelby tube sample of foundation materials. Each set of tests also had Atterberg Limits, particle size distribution (with hydrometer), water content, and specific gravity tests performed.

February 28, 2018 Page 10 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

5.0 SUBSURFACE CONDITIONS

5.1 Geologic Profile

Several geologic materials have been identified at the site during the 2016 field investigations. Previous investigations encountered discontinuous glaciofluvial soils in the central valley portion of the site and the abutments. The glaciofluvial soils in the central valley were classified during the original design phase as loose and potentially subject to cyclic liquefaction. The design and construction of the dam included over-excavation of the foundation in the central valley portion of the dam to remove the loose glaciofluvial soils.

The geologic units encountered in 2016 that have been identified have characteristics that are distinct enough to identify them in the field based on visual descriptions and field testing. Each material is described below, with a brief summary of characteristics and details from laboratory testing, field testing, and in situ testing. Two geologic cross sections located through the dam where the 2016 test borings were performed are shown on Figures 8 and 9.

The geologic materials observed are as follows:

Fill material consisting of both silt and clay embankment fill and poorly graded sand that forms the seepage control system for the existing dam.

Foundation material that is comprised of galciolacustrine silt and clay.

The characteristics of these units are presented below. The soil classifications noted below and elsewhere in this report are visual classifications, except at intervals where laboratory testing has been performed. Detailed descriptions of the subsurface materials encountered in the subsurface explorations are included in the test boring logs (Appendix A). Standard Penetration Test results (N values) are listed as recorded in the field.

5.1.1 Fill Material

5.1.1.1 Embankment Fill

The embankment fill material comprises the majority of the dam. Both the embankment above grade and the over-excavated section below original grade appear to be fairly homogeneous with the soils varying between LEAN CLAY (CL) and SILT (ML). In terms of Atterberg Limits (plasticity), most samples were clustered around the A-line and no sample had a liquid limit above 50. Plasticity was typically low to medium and the combined coarse fraction was often <15% based on index testing. Color was typically brown above the groundwater table and then transitioned to gray below the groundwater table. SPT N-values ranged between 9.3 and 37.5 blows per foot and averaged 20 blows per foot.

5.1.1.2 Filter and Drainage Materials

The filter and drainage materials in the blanket drain were encountered in Borings B-2 and B-4. Through visual classification and laboratory testing, the filter and drainage materials were classified as POORLY GRADED GRAVEL WITH SILT (GP-GM) to POORLY GRADED SAND WITH SILT (SP-SM). The filter and drainage materials were typically moist and brown in color. In terms of density, the material was medium dense to dense. The percent passing the #200 sieve varied from 5.3% to 10.8% for the four

February 28, 2018 Page 11 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved samples tested. The fine drain fill was originally designed to have PennDOT No. 1 gradation limits while the coarse drain fill was designed to meet PennDOT No. 2B limits. The fine drain fill was originally designed to have less than 3% passing the #200 sieve, and 10 to 30% passing the #100 sieve. During construction, the allowable gradation for fine drain fill was changed to allow less than 6% passing the #200 sieve, and 8 to 30% passing the #100 sieve. The change was documented in the as-built drawings, and in Reports 297 and 299 of the job diary. The coarse drain fill was designed to have less than 3% passing the #200 sieve. The drainage fill sampled failed to meet either of these criterion for the #200 sieve.

The design details for the blanket drain included with the original design drawings showed fine drain fill above and below a layer of coarse drain fill. The original field logs from our investigation indicated that Borings B-2 and B-4 intercepted the blanket drain, and a discrete layer of fine drain fill was not recovered from the split spoon samples collected through the blanket drain. Rather, the upper portion of the blanket drain materials recovered was gravelly and sandy soils that were initially visually classified as POORLY GRADED GRAVEL WITH SILT (GP-GM). Upon re-evaluation and discussion with the field representative, the field interpretation likely misrepresented the strata encountered. The driller did not use a center plug in the augers, so the LEAN CLAY (CL) observed between the filter layers was likely left in the augers above. The split spoons for both B-2 and B-4 showed 60 to 65% recovery at the transition from shell fill to the top of the blanket drain; this indicates that part of the sample was lost, or the sampler became plugged during driving and the top of the blanket drain was not effectively sampled. Based on our re-evaluation, we revised the drill logs of Borings B-2 and B-4 to show a 0.9- and 0.8-ft thick layer of filter sand above the drain fill, respectively. The reinterpreted strata changes are more consistent with the as-built drawings and construction documentation.

5.1.2 Foundation Material

The foundation material consists of glaciolacustrine soils. The glaciolacustrine soils were primarily classified as SILT (ML) or LEAN CLAY (CL), but layers of SANDY SILT (SP-SM/SM) were also present.

Clay pockets were observed in SPT samples, but thin alternating layers of silt and clay were not observed. In previous investigations, the glaciolacustrine deposit was sometimes described as a varved soil deposit. Varved soils are commonly characterized as soils with thin interbedded layers of varying soil types such as clay and silt. Often the interbedded layers are horizontal and can be very thin (⅛-inch thick or less). Thin interbedded soils layers were not observed in the samples of the foundation soils taken from the test borings performed in 2016.

The foundation materials often had plasticity indexes below 10. The color of this stratum was gray. The coarse fraction was typically lower than five percent. SPT N-values in the top 20 ft of the foundation ranged from 4 to 40 blows per foot. SPT N-values more than 20 ft into the foundation ranged from 11 to 63 blows per foot.

Rock was not encountered in the 2016 borings at a maximum depth of exploration of 139.5 ft.

February 28, 2018 Page 12 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

5.2 Hydrogeologic Conditions

5.2.1 Water Level Conditions

Water level measurements were taken during and after drilling. In Borings B-1, B-3, and B-4, water levels were observed between approximately EL 941 to 953. Boring B-2 exhibited a higher water level ranging between approximately EL 965.0 and 970.5. Water level measurements obtained during our subsurface exploration are shown on the boring logs in Appendix A. Longer term water level readings were not obtained because the borings were backfilled upon completion to avoid borehole collapse.

Reported water levels should not be considered globally representative because it is anticipated that the measured water levels could have been affected by drilling procedures and would also be expected to vary due to variations in precipitation and surface runoff.

February 28, 2018 Page 13 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

6.0 GEOTECHNICAL DESIGN CONSIDERATIONS

6.1 General

The following sections address previously identified deficiencies, including those presented in the Gannett Fleming reports (Gannett Fleming, 2004; and Gannett Fleming, 2005).

6.2 Dispersive Soils

The glaciolacustrine soils were reported to have “high dispersion” in the hydrometer dispersion tests performed during the original design of the dam. Additional dispersion testing included pinhole dispersion testing and pore water chemistry evaluations during the original design phase of the project. The results of three of the four pinhole dispersion tests and the pore water chemistry evaluations indicated that soil dispersion was not an issue. The NRCS considered the pinhole dispersion and pore water chemistry evaluations more reliable than the hydrometer. It does not appear that soil dispersion was addressed as part of the original design.

Following construction, cracks and erosion features such as jug holes and tunnels were observed in the embankment with most of the features being observed on the downstream slope. In 2004 and 2005, Gannett Fleming collected several samples for crumb dispersion testing with about two-thirds of the samples being classified as either slightly dispersive or highly dispersive.

In 2005, a program to remediate the observed erosion features (jug holes, tunnels, etc.) was performed, including locally excavating each erosion feature to about one foot beyond its limits and then backfilling each excavation with lime-treated soils. The backfill was a similar clay soil obtained from a downstream borrow site and that was mixed with 2% lime. The lime used was a blend of ⅓ hydrated lime and ⅔ agricultural lime.

Since 2005, additional erosion features such as jug holes, cracks, and tunnels have been observed at the site, although at a much lesser degree than what had been observed prior to the 2005 lime treatment program. As part of the GSFW field and laboratory investigation, six pinhole dispersion tests were performed on the surficial soils collected from the test pits. Four of the six test results were classified as ND3 – Slightly Dispersive, and two of the test results were classified as ND2 – Non Dispersive. GSFW also performed 24 crumb dispersion tests on samples of both embankment fill soils and foundation soils collected from the test borings. The result of all the crumb dispersion tests performed by GSFW were Category 1 Non Dispersive.

Given that: (1) false negative results can occur in the crumb dispersion test, (2) four of the pinhole dispersion test results were considered ND3 – Slightly Dispersive, and (3) the observed jug hole and erosion features on the site, we conclude that slightly dispersive to dispersive soils likely exist within the limits of the existing embankment and the auxiliary spillway. It is also likely that the surficial soils with a high degree of soil dispersion were located in the areas where the pre-2005 erosion features were observed and that the lime treatment in 2005 addressed areas of the embankment where highly dispersive soils were present.

The goal of the current remediation is to develop a plan to address dispersive soils such that frequent maintenance or filling of erosion features is not required. The plan proposed in the 2010 planning study

February 28, 2018 Page 14 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved was to treat the surficial soils of the dam and auxiliary spillway to significantly reduce the potential for development of dispersive soil erosion features and reduce the likelihood that frequent erosion maintenance would be required. Note that during the recent site inspections, the maximum depth of jug holes observed on site was less than 24 inches. If deeper jug holes are encountered during construction after surficial soils are removed, additional remediation efforts should be evaluated.

GSFW performed a lime treatment mix design program to evaluate the effectiveness of using hydrated lime to reduce the dispersion potential of the on-site soils. The four samples that were classified as ND3

– Slightly Dispersive in the pinhole dispersion tests were augmented with 2% hydrated lime (by dry weight of soil), allowed to mellow for at least 48 hours, and then remolded and pinhole dispersion testing was performed. The result of the pinhole dispersion test on each lime-augmented soil was ND1 – Non Dispersive. The pinhole dispersion test results on the lime-augmented soils demonstrated the effectiveness of the hydrated lime to reduce the potential for soil dispersion.

To evaluate the impacts of the lime augmentation on the clayey soils, GSFW performed Atterberg Limits tests on three of the lime-augmented soils. The results showed that both the liquid limit (LL) and the plastic limit (PL) were raised slightly by adding the hydrated lime. A comparison of the pre- and post-lime augmented Atterberg Limits is summarized in Table 6.1.

Table 6.1: Summary of Pre- and Post-Lime Augmented Atterberg Limits

Sample Designation Liquid Limit Plastic Limit Plasticity Index

TP-09 30 23 7

TP-09 (LIME) 40 30 10

TP-17 49 25 24

TP-17 (LIME) 60 34 26

TP-20 47 26 21

TP-20 (LIME) 64 36 28

To evaluate the impacts of lime augmentation on the ability of the lime-augmented soils to sustain grass vegetation, three samples of lime-treated soils were sent to the Agricultural Analytical Laboratory at State College, Pennsylvania, for agricultural soil tests (agronomy). Lime-augmented soils from TP-09, TP-17, and TP-20 were tested after an aging time of five to six weeks. The results of the tests included:

Soil pH Phosphorus (P in ppm) Potassium (K in ppm) Magnesium (Mg in ppm)

The analysis reports also include recommended fertilizer rates for the following constituents:

Nitrogen (lb N/A) Phosphate (lb P2O5/A) Potash (lb K2O/A)

February 28, 2018 Page 15 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

The results are included in Appendix B.

6.3 Auxiliary Spillway Erosion

The potential for the auxiliary spillway to erode will be evaluated as part of the preliminary design.

Geotechnical parameters required for the analyses include the headcut erodibility index, Kh. GSFW estimated Kh for an idealized cross-section that could be entered into the NCRS SITES computer program to perform the spillway integrity analysis.

The soil profile of the auxiliary spillway primarily consists of lean clay cohesive soils based on the results of NRCS Test Boring DH-251 and DH-252 as well as Bioengineering Test Borings B-1 through B-3 (BIO B-1 through BIO B-3). Sandy soil layers were encountered downstream of the control section in BIO B-1 and BIO B-3 at elevations ranging from about EL 970 to EL 960 ft. A sandy layer was also encountered upstream of the control section in DH-251 from about EL 980 to EL 970. GSFW evaluated Kh values for surficial soils that will be treated with lime and potentially subject to freeze-thaw conditions, deeper clay soils, and the isolated seams of the sandy soils as listed below:

Surficial lime-treated clay soils, 0 to 2 ft, Kh = 0.02 Clay soils 2 ft or deeper, Kh = 0.05 Sandy soils in discontinuous seams, Kh = 0.04

The calculations used to estimate Kh are included in Appendix C. Plasticity Index (PI), Dry Density, and Percent Clay for the Auxiliary Spillway samples can be found in Appendix C.

6.4 Seepage and Static Slope Stability

6.4.1 Historic Concerns Regarding Seepage, Settlement, and Stability

The performance of the embankment dam with respect to settlement and stability was not predicted during the design phase of the project. Technical issues identified during construction included:

1. During the initial placement of the principal spillway conduit on concrete blocks and prior to construction of the concrete cradle, the concrete blocks settled into the compacted fill subgrade overnight and the pipeline had to be re-set.

2. Settlement of the principal spillway conduit, approaching about 6 inches, occurred after it was placed and resulted in joints opening.

3. Cracks and bulges were observed in the downstream slope fill soils during placement. The embankment fill was placed to within about 10 to 12 ft of the top of the dam when fill placement was halted during construction. Fill placement resumed after a field and laboratory investigation was performed by the NRCS. Fill placement was reportedly completed without further incident.

The dam was placed into service and maintained a permanent pool until 2004 when it was drained due to concerns related to embankment seepage, internal erosion, and stability.

February 28, 2018 Page 16 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

PADEP engaged Gannett Fleming to evaluate the above issues circa 2004-2005. Gannett Fleming issued a report titled Conneautville Dam Technical Review in October 2004 (Gannett Fleming, 2004) and a subsequent report titled Conneautville Dam Stability and Compatibility Evaluation in June 2005 (Gannett Fleming, 2005).

The reports issued by Gannett Fleming discussed the ability of the existing chimney and blanket drain to serve as filter seepage and leakage, and presented the results of their evaluations of dispersive soil issues (as discussed above) and slope stability.

Gannett Fleming reviewed grain size distribution curves for samples of fine and coarse drain fill collected from their 2005 test pits. Based on their evaluations, Gannett Fleming concluded the filter system in the dam does not meet current NRCS criteria for filters. However, they also stated that given consideration for dispersive soils and the nature of the deviations from the current NRCS filter criteria, “the final determination of acceptability or non-acceptability of the filter drain system can only be made on the basis of direct testing of representative materials from the dam using the No-Erosion Filter Test.” (Gannett Fleming, 2005). Gannett Fleming performed four No-Erosion Filter Tests (NEF Tests) on embankment and drain system samples. The results of the testing showed some material loss during water flow was observed along with turbid discharge during the tests. But in each case, a filter cake developed and the discharge eventually ran clear, indicating that the filters self-healed and adequate performance was achieved.

GSFW evaluated filter performance during the preliminary engineering stage of the project based on: (1) the results of the 2016 borings which indicated a discrete upper fine drain fill layer may not be present in some portions of the blanket drain, (2) the results of both Gannett Flemings’s gradation tests and their No Erosion Filter tests, and (3) the results of gradation tests performed on samples of the blanket drain recovered in 2016 by GSFW. The results of the 2005 Gannett Fleming filter analysis and laboratory (No Erosion Filter) testing were reviewed by GSFW and our technical consultant, Mr. James Talbot, PE. We concluded that the results of the 2005 Gannett Fleming analysis suggested that the existing filter and drain system may not meet current NRCS standards in all locations, but that the existing filter is considered adequate since it is outside the “allowable” gradation band (developed with criteria published in NRCS Chapter 26 of the NEH) by a small amount, and the NRCS filter design methodology has a high level of conservatism. The results of the 2016 GSFW investigation also suggest that the 2005 Gannett Fleming analysis may represent an evaluation of anomalous conditions that could be considered a “worst case” analysis given the history of care the NRCS has given to filter construction since the early 1980s.

Results of additional filter analysis are represented in Appendix D.

The existing drainage system does not include a filter diaphragm below the principal spillway conduit, which is required under current NRCS criteria. Construction of filter diaphragm below the conduit is recommended; however, since the existing blanket chimney drain performs the function of a filter above the conduit, extending the proposed diaphragm above the top of the existing drain is not necessarily required. However, it is typical for a filter diaphragm around a pipe penetration to be thicker than the existing drain; therefore, we recommend extending the filter diaphragm for conservatism and to address the potential for concentrated seepage along the existing conduit.

Gannett Fleming developed five static normal pool load cases that included:

including "no erosion filter" leads me to think this test was performed and results are available.

is this the case?

my recollection of GF report is a no erosion filter test is recommended to evaluate the suitability of filter against piping

Donald.Murray Highlight

February 28, 2018 Page 17 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

1. Global circular failure

2. Global block failure through potential weak layer in the varved foundation (i.e., thinly bedded weak clay layer)

3. Circular failure with a water filled crack (pre-existing crack in the embankment)

4. Circular failure through “pre-existing” shear zone that is saturated due to stormwater runoff

5. Block failure through “pre-existing” shear zone that is saturated due to stormwater runoff

The result of the Gannett Fleming analyses indicated the calculated factors of safety were all above 1.5 except for Load Cases 2 and 4, where the calculated factors of safety were 1.44 and 1.49, respectively.

Given the available data in 2005 when the analyses were performed, the above analyses are considered conservative but representative of a worst-case scenario. Since then, additional data collection and remediation have been performed at the site that, in the opinion of GSFW, result in Load Cases 2 through 4 being considered overly conservative and not representative of the conditions revealed through the 2016 exploration. The rationale for our assessment includes:

A weak layer due to varved soils was not observed in the foundation.

The extensive embankment cracking observed in 2004 and early 2005 that would potentially allow rain water to locally saturate zones of the downstream slope was remediated in 2005 with the lime treatment. Also, additional surface treatment is planned to reduce the potential for future cracking to develop.

The dam is expected to remain as a dry dam and only impound water for a short duration during large storms.

GSFW reviewed the results of the previous field and laboratory investigations and evaluated the seepage and stability of the embankment as described below in general accordance with TR-60.

6.4.2 Evaluation of Current Conditions (Static Loading)

6.4.2.1 Seepage Analyses

The reservoir is expected to remain dry, which reduces potential issues associated with internal erosion, the condition of the conduit, and other factors. However, GSFW performed preliminary analyses assuming steady state seepage while the reservoir impounds water to EL 981.25 ft, which corresponds to the principal spillway crest. We also performed a preliminary transient seepage analysis using reservoir

vs. time data for the 6-hr freeboard hydrograph to evaluate the potential for the embankment to become saturated during the spillway design storm. The steady state seepage analyses conservatively assumed the horizontal permeability of the embankment and foundation soils were both 1x10-6 cm/sec, and the vertical to horizontal permeability (Kv/Kh) was 0.5. We estimated properties required for transient seepage analyses based on the results of one-dimensional consolidation tests performed on samples taken from the partially constructed embankment in 1983. The seepage analyses details and results are presented in Appendix D.

February 28, 2018 Page 18 GSFW Engineering JV Project 15715008.00 ©2018 All Rights Reserved

The results of the steady state seepage analyses show that both the upstream and downstream slopes of the embankment essentially remain unsaturated. The results of the steady state seepage analyses were used in our slope stability analyses to estimate pore water pressures. The results of the transient seepage analyses used to model the migration of a wetting front into the embankment during the spillway design flood showed that the wetting front (depth of saturation) moved into the dam less than one foot.

Since the depth of saturation during the SDF is less than one foot (based on transient analysis), assuming steady seepage from the crest of the auxiliary spillway is considered overly conservative and therefore was not evaluated.