J 35 73 20.00 29 Attachment (2) Geotechnical Relief Well Pump Testing Report-Amendment #1.pdf

REPORT

Final Geotechnical Relief Well Pump Testing Chena River Lakes Flood Control Project, North Pole, Alaska

Contract: W911KB-17-D-0002

Task Order: W911KB17F0047

Submitted to:

U.S. Army Corps of Engineers, Alaska District Attn: Robert T. Weakland, PE

PO Box 6869

JBER, AK 99506

Submitted by:

Golder Associates Inc.

2121 Abbott Road, Suite 100 Anchorage, Alaska, USA 99507

+1 907 344-6001

1780780

December 20, 2018

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

CHL023

Section 35 73 20.00 29 Attachment (2) Page 1 of 672

Distribution List 2 Copies - U.S. Army Corps of Engineers

1 Digital Copy - U.S. Army Corps of Engineers

Section 35 73 20.00 29 Attachment (2) Page 2 of 672

Table of Contents

1.0 INTRODUCTION

1.1 Scope of Work

1.2 Subsurface Conditions

1.3 General Relief Well Construction

1.4 Relief Well Locations

2.0 FIELD METHODS

2.1 Inspect and Document Initial Conditions

2.1.1 Surface Survey

2.1.2 Down-hole Video Survey

2.2 Conditioning

2.2.1 Brushing

2.2.2 Air Surging (Development)

2.2.3 Sediment Removal

2.3 Step-Drawdown Testing

2.3.1 Sand Testing

2.4 Equipment Used

2.5 Protection of Site

2.6 Fieldwork Challenges

3.0 RESULTS

3.1 Surface Survey

3.2 Down-Hole Video Survey

3.3 Step-Drawdown Testing

3.3.1 Transmissivity and Hydraulic Conductivity

3.3.2 Storativity

3.3.3 Well Efficiency

3.3.4 Relief Well Performance

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 3 of 672

3.3.5 Sand Testing

4.0 DISCUSSION AND CONSIDERATIONS

5.0 LIMITATIONS AND USE OF REPORT

6.0 CLOSING

7.0 REFERENCES

TABLES

Table 1: Summary of Relief Well Surface Evaluations

Table 2: Summary of Relief Well Down-Hole Video Surveys

Table 3: Summary of Relief Well Step-Drawdown Testing

Table 4: Results of Observation Well Data Analyses

Table 5: RMSE for the Best-Fit Model and Step One

Table 6: Observed Information from Pumping Tests for Use in Future Relief Well Performance Evaluations

Table 7: Timeline of Step-Drawdown Testing and Surging

FIGURES

Figure 1: Vicinity Map

Figure 2: Key Map

Figure 3: Relief Well Location Map – Wells 147+40 to 148+00

Figure 4: Relief Well Location Map – Well 171+00

Figure 5: Relief Well Location Map – Wells 270+80 to 286+50

Figure 6: Relief Well Location Map – Wells 287+25 to 302+30

Figure 7: Relief Well Location Map – Wells 304+00 to 319+25

Figure 8: Relief Well Location Map – Wells 319+65 to 334+15

Figure 9: Relief Well Location Map – Wells 338+50 to 339+00

Figure 10: Relief Well Location Map – Wells 352+50 to 364+50

Figure 11: Relief Well Location Map – Wells 369+60 to 378+07

Figure 12: Relief Well Location Map – Well 418+00

Figure 13: Relief Well Location Map – Wells 434+50 to 446+20

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 4 of 672

Figure 14: Relief Well Location Map – Wells 473+00 to 478+00

Figure 15: Field Investigation Timeline

Figure 16: Typical Well Head Construction

APPENDICES

APPENDIX A

Photo Log

APPENDIX B

Step-Drawdown Test Results

Section 35 73 20.00 29 Attachment (2) Page 5 of 672

Acronyms bgs below ground surface ft feet

EIS Environment Impact Statement

GFI government furnished information gpm gallons per minute

HP horse power

INW Instrumentation Northwest Inc.

kW kilowatt ml milliliter

N/A not applicable ppm parts per million psi pounds per square inch

PVC polyvinyl chloride

RMSE root mean square error

RST Rossum Sand Tester

SOW scope of work

TOC top of casing

USACE U.S. Army Corps of Engineers

Section 35 73 20.00 29 Attachment (2) Page 6 of 672

1.0 INTRODUCTION

Golder Associates Inc. (Golder) was contracted by the U.S. Army Corps of Engineers (USACE), Alaska District to perform the services outlined in the revised Statement of Work (SOW) for the “Geotechnical Relief Well Pump

Testing, Chena River Lakes Flood Control Project, at North Pole, Alaska,” dated July 6, 2017 and modified on

August 14, 2017 (Contract No. W911KB-17-D-0002, Task Order W911KB17F0047).

Relief wells are located on the west (downstream) side of the Moose Creek Dam as part of the Chena River

Lakes Flood Control Project in North Pole, Alaska (Figure 1). The relief wells provide pressure relief to the dam during high-water events by allowing water to free flow from the well heads in a controlled manner. The water then flows into a seepage collection channel to the west and is directed away from the project site.

The relief wells have been installed in groups starting in 1980 up to 2017; however, records of initial installation, development or conditioning, and initial conditions are absent. USACE has requested that baseline information for

152 select relief wells (Figures 2 through Figure 14) be established so that potential changes can be documented for use in planning and developing future effort.

Baseline information was collected over two field seasons from 2017 to 2018 (Figure 15) and includes surface observations near the well heads, downhole video survey of the well screens, conditioning, and testing of the relief wells to convey water from the surrounding aquifer. This report describes the field methods used, the results of the field observations and testing, and further considerations to maintain and improve the conveyance of groundwater during high-water events.

1.1 Scope of Work

Work was accomplished in the following four tasks as outlined in the SOW:

Task 1 – Project Management and Plan Development. Work completed for this task included review of government furnished information (GFI 1 through 7)

▪ GFI 1 – Project Location Vicinity Map

▪ GFI 2 – Relief Well Locations

▪ GFI 3 – Relief Wells to be Tested

▪ GFI 4 – 2011 Drawdown Test Report

▪ GFI 5 – 2016 Relief Well Testing Report

▪ GFI 6 – Relief Well Pumping Test Report

▪ GFI 7 – Relief Well Inventory

Task 2 – Inspect and Document Initial Conditions

Task 3 – Relief Well Conditioning and Pumping Tests

Task 4 – Relief Well Maintenance Pumping Reporting and Project Files Report

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 7 of 672

1.2 Subsurface Conditions

Shallow subsurface conditions across the site mainly consist of quaternary floodplain alluvium and swamp deposits. The alluvium is described as having well-stratified layers and lenses of unconsolidated soil that includes pebbles, cobbles, and boulders. Swamp deposits are described as peat and silt layers that can be more than five feet thick in some places (Pèwè et al.,1966). A recent geotechnical report by Golder (2017) further describes the site as having silt with varying levels of sand overlying alluvial sand and gravel deposits, based on geotechnical boreholes advanced at select sites near the relief wells. Permafrost was not identified on the site in that particular investigation; however, it is mentioned in GFI 4 as being present during relief well installation in the 1980s.

1.3 General Relief Well Construction

The majority of the relief wells are nominally 12 inches in diameter with the exception of six that are eight inches in diameter. The relief wells are constructed with continuous well screen from near the ground surface down to the bottom of the relief well. Above the well screen is blank steel casing. Relief well head construction is generally one of two types that this report will refer to as a culvert-type and a pipe-branch type. Examples that indicate the differences in the well head construction are provided in Figure 16.

1.4 Relief Well Locations

The relief wells are located relatively close to Laurance Road or adjoining project access roads, and they are identified by a station numbers ranging from 147+40 to 478+00. The eight-inch diameter relief wells are located as a group north of Lake Park Road (Figure 8). Locations for 138 of the 152 relief well locations, shown in Figures 3 through Figure 14, were provided by USACE; and the remaining relief well locations were field located based on stationing and photographs. Coordinates for each relief well and the method used to locate the relief well are provided in Table 1.

2.0 FIELD METHODS

Field efforts were conducted during two separate field seasons, from September 11, 2017 to October 10, 2017 and May 8, 2018 to July 24, 2018. A timeline of the field effort is presented in Figure 15. Photographs of the field effort are presented in Appendix A and in the digital Project Files Report. Prior to beginning fieldwork, a Work

Plan, Abbreviated Accident Prevention Plan, Quality Control Plan, and a Traffic Control Plan were developed by

Golder and approved by USACE.

A two to three-person crew was on-site for the majority of the field effort; the crew consisted of staff from Golder and Arctic Drilling Inc. (Arctic) of Fairbanks, Alaska. Golder directed the field efforts and was responsible for the collection and reporting of information. Recorded data were reviewed following the procedures described in the

Quality Control Plan submitted to and approved by USACE prior to performing work. Arctic performed the conditioning of the relief wells and was responsible for the movement and maintenance of the equipment and tools used to perform the step-drawdown testing.

Individual relief well characteristics restricted or increased effort to complete the services outlined in the SOW.

Section 2.6 describes these field challenges.

2.1 Inspect and Document Initial Conditions

2.1.1 Surface Survey

Golder inspected both the condition of the visible portion of the relief well above the ground surface and the immediate area approximately five feet around each relief well. Visual observations were also documented for:

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 8 of 672

Vegetation, trash, debris, and visible damage or vandalism

The condition of the painted relief well station identification

The condition of the snow flag marker

Measurements collected at each relief well include:

Vertical distance of relief well invert lip to the ground surface

Initial sounding depth (depth of relief well)

Initial water level relative to the ground surface and top of casing (TOC)

Location of the closest relief well

Sounding depths were measured using a flat-bottomed weighted cloth tape. A cloth tape and steel tape measure were used to collect the additional measurements. Various common hand-held water level meters were used to measure the depth to water relative to TOC. Distance to the nearest relief well for 171+00 and 418+00 was not measured due to the distance (>1000 feet) and error that would have been accrued with the measurement method.

2.1.2 Down-hole Video Survey

Golder performed a down-hole video survey of each relief well installed before 2017, for a total of 143. The survey was performed prior to conditioning by a two-person Golder crew from September 11 to 21, 2017. The one exception was Relief Well 309+75, which was buried beneath the sidewalk next to Lake Park Road during the

2017 field season. Golder received access to the well during the 2018 field season and performed the down-hole video survey on June 12, 2018.

The video surveys were recorded with a Geovision Standard Color SSM1 Camera paired with the Geovision

Heavy Duty Econo System (Appendix A, Photo 1). Real-time viewing was used during recording for quality assurance as well as camera depth display that is also included in each video. The camera angle was pointed straight down hole during the survey because it was found to capture the most information due to the advantage of reflection from the well screen. Audio was not recorded. Items noted during the down-hole video surveys include:

Depth from the top of casing to the interface between the steel riser and stainless-steel well screen

Location and extent of biofouling described qualitatively as minor, moderate or extensive, and an interpretation of color

Location and extent of mechanical blockages

Location and extent of other screen abnormalities such as foreign objects and breaks in the screen

All measurements were recorded from the TOC.

2.2 Conditioning

Following the down-hole video survey, relief well conditioning was conducted in the form of brushing, surging with air, and removing sediment. Conditioning was performed to increase water flow into the relief wells prior to collecting baseline hydraulic properties through step-drawdown testing. A total of 46 relief wells were conditioned

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 9 of 672 during the 2017 field season. The remaining 104 relief wells were conditioned during the spring and summer of

2018.

2.2.1 Brushing

Brushing was performed on 135 of 152 relief wells. Brushing was not part of the SOW for the six eight-inch diameter relief wells or the nine new relief wells that were installed by USACE in the summer of 2017. In addition, two relief wells (Relief Well 355+64 and Relief Well 377+57) were not brushed because the inner radius had been modified by USACE. Additional discussion is presented in Section 2.6.

The brushing operation was performed using a nylon brush that was attached to drill rods (Appendix A, Photo 2) and pushing it to the bottom of each casing and then removing it.

2.2.2 Air Surging (Development)

Development by air surging was performed on 150 of 152 relief wells after they were brushed. Relief Wells

355+64 and 377+57 were not developed as described in Section 2.6. Air surging was performed by first injecting air at a low rate just below the static water level to allow uninhibited flow through the relief well screen. Equipment was then advanced down-hole at the rate of five feet every five minutes. At each interval, compressed air was injected at low rates lifting the water in the relief wells to near the top of the casing for a short duration (Appendix

A, Photo 3 and 4), and then turning off the air to let the water fall back down into the well. This operation helped clean and develop the well by rapidly inducing flow into the well through the screen, and then reversing the flow direction in an alternating fashion. One full cycle of air surging through the screened portion of each relief well was completed.

2.2.3 Sediment Removal

After air surging was performed, accumulated sediment was removed from the bottom of 150 relief wells by either air lifting the accumulated sediment or by a type of reverse circulation using compressed air. A flat bottom tape measure was used to measure the depth of the relief wells after surging. Compressed air was then injected consistently at a low flow rate at the bottom of the well for approximately 15 minutes. The depth of the relief well was then sounded again, and any change was noted. If the sounding depth changed more than 0.1 feet from the measurement recorded after surging, reverse circulation was performed by setting the air intake lower on the inductor pipe column which created a suction or “vacuum” at the bottom of the relief well (Appendix A, Photo 5).

Reverse circulation was performed for an additional 15 minutes and the relief well was then sounded again and recorded as the final depth. Measurements were recorded from TOC.

2.3 Step-Drawdown Testing

After conditioning, a step-drawdown test was performed on 147 relief wells. Five of the relief wells were not step-drawdown tested as discussed in Section 2.6. Thirty relief wells were tested during the 2017 field season, and the remaining wells were completed in 2018.

The step-drawdown testing was performed to estimate baseline hydraulic parameters of the relief wells. These hydraulic parameters include: hydraulic conductivity, the effective pumping rate at 65 percent efficiency and the theoretical specific capacity for the optimum pumping rate. Testing was performed in general conformance with the methods described in Engineering Manual 1110-2-1914, Design, Construction, and Maintenance of Relief

Wells and ASTM D 4043-17, Standard Guide for Selection of Aquifer Test Method in Determining Hydraulic

Properties by Well Techniques.

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 10 of 672

Each relief well was pumped for a minimum of four hours. Approximately every 60 minutes, the pumping rate was increased by 25 percent creating a “step” increase each time for a total of four steps. The target pumping rates for each step-drawdown test was estimated based on the results of pumping tests conducted in 2010 provided in GFI

3, Table 1. Water level drawdown in the relief wells was recorded every 15 to 30 seconds using an

Instrumentation Northwest (INW) PT2X vented pressure transducer (30 psi). The water level drawdown was checked at approximate 15-minute intervals using a hand-held water level meter. Prior to the start of pumping, a static water level was verified and recorded.

Target pumping rates ranged from 20 to 2,000 gallons per minute (gpm). A nominal 10-inch diameter submersible pump manufactured by National Pump Company (model SH10MC-1) was used in combination with a 60 HP motor manufactured by Franklin Motors for the majority of the step-drawdown testing (Appendix A, Photo 6).

However, target flow rates were not always achievable with the 10-inch submersible pump described above.

Therefore, several smaller pumps were also used to perform the step-drawdown tests, as discussed in Section

2.4. A gate valve was used to vary the pumping rates, and a flow meter with a gauge and totalizer was used to establish and monitor pumping rates during the step-drawdown test. The pumps were set approximately 30 feet below the TOC on a steel pipe column with a 90-degree elbow at the top to connect to the surface discharge system. Examples of the surface portion of the discharge system are presented in Appendix A, Photos 7 through

9.

Discharge water from the pump testing was released in a controlled manner to allow the water to flow into the seepage collector channel. Water extracted and discharged from the relief wells during testing was authorized under the initial construction Project Environmental Impact Statement (EIS) acquired by USACE for the project per the SOW, and no other permits were acquired for extraction or discharge prior to executing the field work.

2.3.1 Sand Testing

During each step-drawdown test, sanding rates were measured in the discharge water using a Rossum Sand

Tester (RST) manufactured by Roscoe Moss Company (Appendix A, Photo 10). The RST was located after a 90-degree bend in the discharge system where flow was estimated to be turbulent. Prior to running the test, the inlet valve to the RST was opened and calibrated to ensure a target flow of 0.5 gpm. Once achieved, the inlet valve was closed, and the glass tube was cleaned and replaced. Sand collected in the glass tube was measured in millilitres (ml) per minute and then converted to parts per million (ppm). A measurement was recorded approximately four times during a step-drawdown test; one measurement was collected during each of the four pumping rates.

2.4 Equipment Used

A pump hoist truck, boom truck (model 1000 JBT) (Appendix A, Photo 11) and tracked drill rig (Yanmar E50 D, Appendix A, Photo 12) were used to lower and lift equipment in and out of the relief wells. The pump hoist truck was primarily used for the brushing and surging effort. The boom truck and tracked drill rig were primarily used to set and extract a pump and associated discharge pipe in a relief well. Additional support vehicles were used to haul and tow support equipment such as a compressor (Ingersoll Rand 400), tool trailers, and generators

(approximately 300 kW capacity).

The following pumps and flow meters were used throughout the duration of the field program to accommodate the wide range of pumping rates.

Submersible Pumps (flow ranges are approximations)

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 11 of 672

10-inch National SH10MC-1: 250 to 2000 gpm

Six-inch Grundfoss 800: 100 to 1100 gpm

Six-inch National 500: 50 to 600 gpm

Four-inch National 45: 5 to 70 gpm

Surface Pump

Six-inch Prime Aire PA6C60-4045D surface pump: 200-2000 gpm

Flow Meter

Eight-inch McCrometer Model LP32-D: 250 to 2000 gpm

Four-inch McCrometer Model LP21: 80-500 gpm

Two-inch Blue-White Model RT-200ml-GPM4: 20-200 gpm

2.5 Protection of Site

To protect the site from damage, erosion control methods were used during water discharge, and special consideration to equipment movement was implemented. To protect side slopes along Laurance Road from sloughing and slope failure, access points with shallow slopes were utilized to mobilize equipment when possible.

To control and reduce erosion from water discharge, the ground surface around relief well heads was protected with tarps and plywood during surging, and energy dissipation techniques were utilized from the point where water discharged to where the discharge water entered the seepage collection channel. Visual monitoring of the site occurred during discharge to ensure that damage was not occurring.

Rutting caused by equipment in the gravel along the shoulder of Laurance Road were raked and restored to original condition prior to demobilizing from the site. Erosion occurred around some of the well heads during surging. Areas where erosion occurred were backfilled with imported gravel as needed and regraded to conditions similar to those noted prior to beginning work. Upon completion of the work, the surface conditions and restoration efforts were verified and accepted by the Chena River Lakes Flood Control Project Dam Tender, Gary Wollangur.

In addition to the above-mentioned effort, well heads were visually inspected for ground settlement during conditioning and testing. These observations were the result of approximately one cubic foot of ground settlement that was observed around the head of Relief Well 285+75, after completion of the surging effort during the 2017 field season. Compared to other relief wells included in the SOW, a significant amount of fine sand was produced in the water during the surging/development effort for Relief Well 285+75 (Appendix A, Photo 13). Additional relief wells produced similar amounts of sand but visible ground settlement was not observed.

2.6 Fieldwork Challenges

While performing the field effort, individual relief well characteristics were identified that either shortened step-drawdown testing or restricted conditioning and step-drawdown testing from being performed. The outcomes described below are a result of decisions made after discussion with USACE.

Conditioning and step-drawdown testing was not completed for Relief Wells 355+64 and 377+57. USACE had installed a 10-inch casing nested inside the 12-inch well casing at an earlier date due to separation that had occurred between the screened section of the well and the blank surface casing. Performing

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 12 of 672 conditioning and step-drawdown testing would have risked damage to the well and conditioning/pump testing equipment placed in the well should any tooling have been “hung up” on the lower end of the inner casing.

Step-drawdown testing was attempted twice but not completed for Relief Wells 360+94, 441+30, and

441+97. An attempt was made to set the 10-inch submersible pump in the relief wells; however, the pump would not advance past a safe working distance below the TOC due to a potential bend in the near surface casing/well screen. Step-drawdown tests were then attempted with the available 6-inch diameter submersible Grundfoss pump. Sufficient drawdown was not achieved to successfully complete a four-hour step-drawdown test.

The last pumping rate,“fourth step,” was not performed for the full 60 minutes for Relief Well 171+00. Two attempts were made to perform the step-drawdown test; however, discharge water flowed back into the relief well during both attempts, artificially elevating water levels in the relief well. During the second attempt, discharge water was rerouted and the relief well head was not inundated until 12 minutes after the fourth step had begun. Data from the second attempt is considered appropriate to estimate the parameters indicated in the SOW.

The last pumping rate,“fourth step,” was not performed for the full 60 minutes for Relief Well 285+75. The sand content measured at the initiation of the “fourth step” was over 1,500 ppm. Therefore, the pump test was shut down to risk damaging the 10-inch diameter submersible pump and the well. Data is considered appropriate to estimate the parameters indicated in the SOW.

Multiple attempts were made to step-drawdown test Relief Wells 283+50, 291+00, and 291+75 due to possible errors that occurred during testing in 2010. Flow rates for the three relief wells were reported in GFI

3 as being much higher than was observed in the field. Pumps had to be removed and another, smaller pump, was reset to perform the test.

Corrections to Relief Well Stationing Provided in SOW

Relief Well 320+84 was incorrectly identified in GFI 3, Table 1. The correct station identification is Relief Well

320+85 and is identified as such in this report.

Relief Well 364+50 is not listed in GFI 3, Table 3; however, it was one of the nine relief wells that was installed in 2017.

3.0 RESULTS

A summary of the results presented in the following sections is tabulated in the attached Table 1 Summary of

Relief Well Surface Evaluations; Table 2, Summary of Relief Well Down-Hole Video Surveys; and Table 3, Summary of Relief Well Step Drawdown Testing.

Additional information that supports both the information provided in the following sections and in Table 1 through

Table 3 is provided in the electronic data deliverables submitted as the Project Files Report. The information provided in the Project Files Report includes:

Completed field sheets that contain the written field recorded information for the surface surveys, down-hole video surveys, conditioning, step-drawdown tests, and sanding rates

Down-hole videos recorded for the survey

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 13 of 672

Photographs of relief wells, equipment, and active work

Original water level data recorded with pressure transducers during the step-drawdown tests

Approved Work Plan, Quality Control Plan, and Statement of Work

Field Daily Reports

3.1 Surface Survey

Near the project site, trash was not prevalent and obvious vandalism was not observed. There were varying amounts of differential movement and scouring of the ground surface relative to relief well heads. The differential movement could be attributed to multiple reasons including frost jacking, ground settlement, and construction methods. Scouring could be the result of previous relief well testing, high-water events, and proximity of some relief wells next to surface drainage paths. A detailed summary of findings from the well surface evaluations are presented in the attached Table 1. Below is a list of the major findings:

9 out of 152 relief wells do not have a relief well identification. Many of these are the relief wells constructed in 2017. 35 additional wells are considered to have a “faint” relief well identification.

13 of 152 relief wells do not have a snow flag.

The distance from the ground surface to the invert lip ranged from -4.7 feet (Relief Well 309+75 is in a vault) to +2.3 feet. The distance from the invert lip to the TOC ranged from -3.3 feet to +3.4 feet.

Initial well depth ranged from 25.1 to 129.6 feet below TOC (the distance from TOC to ground surface is also provided).

Water levels in the relief wells ranged from 1.6 to 18.4 feet below ground surface (bgs) and 3.6 to 15.9 feet below the TOC.

Distances between the nearest relief well ranged from 20.2 to 302.3 feet. This range does not include the distance estimated by station number for Relief Wells 171+00 and 418+00. Depth to water in the nearest relief well is not specifically tabulated in Table 1 because the depth to water is already provided for each relief well and measurements for successive relief wells were recorded over a short time period from

September 21 to 27, 2017.

3.2 Down-Hole Video Survey

The down-hole video survey results show that biofouling, mineral deposits and sediment are present throughout the screened sections in the majority of the relief wells surveyed. Mineral deposition is minor; however, the severity of the biofouling ranged from minor growth to extensive growth, so thick that the well screen was not visible. What is thought to be multiple types of biofouling were observed based on a change of color and texture.

Without further testing, the actual type(s) of biofouling are unknown.

The most prevalent bio-foul is likely an iron-type and is soft, orange, and tends to grow in clumps (Appendix A, Photo 14). The majority of relief wells located on the east side of Laurance Road and those located on the north side of the Chena River have significant sections of well screens that were covered with extensive orange biofoul.

A summary of findings from the down-hole video surveys are presented in the attached Table 2. When a color is not included in the description, it is the orange biofoul described above. Mineral deposition is not specifically

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 14 of 672 identified in the Table 2, but it was suspected to minor extents in the majority of relief wells. It is common for mineral deposits to form when biofouling is present (Sterrett, 2007).

In addition to the biofouling and mineral deposits, sticks, small dead animals such as voles and mice, and construction material were observed in multiple wells. Minor screen damage was also observed in multiple wells.

Voles and mice were observed at the bottom of relief wells and floating on top of the water column. The largest animal observed was a decomposing porcupine in Relief Well 478+00 approximately 24 feet below the TOC. A hammer was observed in Relief Well 339+00 and piece of PVC was observed in Relief Well 300+20. Minor screen damage was observed in Relief Wells 324+70 and 294+00 in the form of moderate bending in the screen;

however, the screen did not appear to be broken. Additional screen damage may be present under the many sections of relief well screen that were covered in thick biofoul.

The majority of relief wells appear to be off plumb based on observation of the camera cable during the down-hole video survey. The camera was placed at the center of the well to start and often contacted the side of the well during the survey, indicating that the well was not plumb. In some videos, the wells were sufficiently off plumb to prevent the camera from being relocated off of the side of the relief well (a centralizer was not utilized on the camera). Wells that have been identified as off plumb are identified in Table 2. The degree to which the relief wells are off plumb was not measured.

3.3 Step-Drawdown Testing

Water levels recorded during the step-drawdown test were processed from the original format to estimate total drawdown of the water level below the static water level, and the elapsed time since the step-drawdown test was started. Drawdown and elapsed time for each relief well were then analyzed to estimate well efficiency by using a best-fit model that is based on theoretical drawdown calculated using Cooper and Jacob (1946) and head losses generated by the well and aquifer (linear and turbulent) estimated by Rorabaugh (1953). A Table 1 and Figure A are provided in Appendix B for each relief well to convey the results of the model. Table 1 for each relief well provides the estimated aquifer and well losses, corresponding efficiency for each pumping rate applied, the theoretical pumping rate at 65 percent efficiency as a range, and the theoretical range of specific capacity for the theoretical pumping rate. Table 3 provides a summary of the information provided in Appendix B.

3.3.1 Transmissivity and Hydraulic Conductivity

Transmissivity was calculated using the slope of the drawdown that was generated during the first pumping rate, “step one,“ plotted on a semi-log plot (drawdown on the y-axis and the log of elapsed time on the x-axis) (Cooper and Jacob 1946). The slope of the drawdown for step one for each relief well is provided graphically as a Figure B for each relief well in Appendix B. The one exception to this is Relief Well 354+50. An upward trend in water levels or recharge effect was observed when the first pumping rate was applied at Relief Well 354+50 that is attributed to development during testing. Drawdown generated during the second pumping rate, “step two”, was used to calculate transmissivity (Figure 354+50-B, Appendix B).

Hydraulic conductivity was then calculated from the transmissivity using the interpreted depth of the aquifer as the depth of the water column in the relief well at the time of pumping measured after sediment removal. Hydraulic conductivity is estimated to range from 162 to 36,842 feet/day throughout the site. This range is considered realistic for the subsurface conditions when compared to published values (100 to 1X106 feet/day) for larger grained subsurface material such as sands and gravels (Freeze and Cherry, 1979). Estimates of hydraulic conductivity for each relief well are summarized in Table 3.

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 15 of 672

3.3.2 Storativity

The storativity is defined as the amount of water released per unit surface area of the aquifer (e.g., one-foot square horizontal area for the full thickness of the aquifer) from a drawdown of one unit (e.g., one foot). This equates to a volume of water per volume of aquifer, and as such storativity is unitless. The storativity varies depending on the aquifer type and the associated hydraulic conditions.

An aquifer that does not have a “confining” barrier above it is classified as an unconfined aquifer. The water released during pumping from unconfined aquifers equates to the amount of water physically drained from the sediment pore spaces as the water table drops. In contrast, a confined aquifer has a low hydraulic conductivity

“confining” barrier above it and yields water to a well from the reduction in pressure and slight associated expansion of water and compaction of the aquifer materials; no physical dewatering of pore spaces occurs.

Consequently, the storativity of confined aquifers is oftentimes 100 to 1,000 times less than that of a confined aquifer.

A default value of 4X10-4 for storativity was initially assumed for each relief well in the best-fit model reflecting confined conditions consistent with aquifer conditions reported in the SOW and associated GFIs and data collected from observations wells in 2017. These values were in the range typical for confined aquifers (1X10-5 to

1X10-3) (Todd, 1980). Low storage coefficients at the site are thought to be a result of the fine surface confining unit reported and in some areas the presence of permafrost.

Based on further analyses of observation well data collected in 2018 and results of the best-fit model storativity was allowed to vary into the semi-confined (leaky) aquifer range (1X10-5 to 1X10-1). Because the storage coefficients used in the analyses were not always derived from actual testing of the subject wells, but rather extrapolated from other onsite well tests, a level of uncertainty exists for the estimated flow range at 65 percent efficiency.

Storativity estimates from observation data ranged generally between 1X10-5 to 1X10-1 and can be found in

Table 4. There were, however, a few storativity estimates below 1X10-5. These very low storativities are considered unrealistic and not reflective of true aquifer conditions; therefore, they were not used in the best-fit model. Scenarios were tested in the best-fit model using the very low storage coefficients for the relief wells that produced the observation well data; however, a better fit to the data were obtained using storativities typical of semi-confined to confined conditions.

3.3.3 Well Efficiency

Estimated head losses attributed to the aquifer and well were used to calculate well efficiency at each pumping rate and the theoretical pumping rate at the relief well’s 65 percent efficiency. Efficiency is the percentage of aquifer losses, which are considered linear, compared to total losses. However, it can be difficult to separate linear aquifer losses from linear well losses; therefore, two definitions of efficiency are commonly used: one that compares all linear (aquifer and well) losses to total losses and one that compares only linear aquifer losses to total losses. The theoretical pumping rates at 65 percent efficiency were initially estimated using the latter;

however, due to the high values of transmissivity observed over the project site, the former has also been provided in the results. Aquifers with high transmissivities will have extremely low linear aquifer head losses, which can in turn magnify the relative linear well loss component, leading to the calculation of low well efficiencies.

The impairment of the functionality of these wells due to well efficiency is negligible in this case.

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 16 of 672

If the definition of efficiency that compares only linear aquifer losses to total losses is applied, estimated efficiency for some relief wells is so low that flows at 65 percent efficiency cannot be attained at any pumping rate. A value of “N/A” has been assigned to approximately 90 relief wells in Table 3 where this is the case. If the definition of efficiency that compares total linear losses (aquifer and well) to total losses is applied, estimated efficiencies for most wells are higher for the applied pumping rates; thus, a pumping rate at 65 percent efficiency can be estimated. The theoretical pumping range for a significant number of relief wells is too high for commonly manufactured pumps that are small enough to fit in the 12-inch diameter relief wells (i.e., 10-inch diameter).

Due to the high transmissivity, the pumping rate at an efficiency of 65 percent is reported as a range to the nearest 10 gpm due to the degree of accuracy afforded by the calculations. It is our understanding that USACE would like to perform a pumping test in the future at the theoretical pumping rate(s) for 65 percent efficiency;

therefore, a range has been provided to the nearest 1 gpm to indicate that there is more than one flow rate that is applicable with the understanding that it is accurate to ±10 gpm. Table 3 contains tabulated results for estimated hydraulic conductivity (ft/day), range of theoretical pumping rates (gpm) at 65 percent efficiency for each relief well, and the associated range of specific capacity (gpm/ft) for each relief well tested.

The best-fit model is considered appropriate for these analyses because the root mean square error (RMSE) of the observed water levels and the best-fit model tend to be relatively low, ranging from 0.025 to 0.37 feet. This estimate was calculated using a sample population of 10 percent of the relief wells (15 relief wells). The relief wells included in the sample were chosen based on their locations throughout the project site. Higher RMSE values are attributed to the model being simplified theoretical estimates of drawdown over the range of flows that do not account for the actual complex environment that exists. Examples of complexity include changes in hydraulic conductivity radially away from the well, boundary effects, changes in storage properties, potential increase in turbulent head losses as water draws down in the aquifer, etc. A summary of the RMSE values for the sample of relief wells is provided as attached Table 5. In addition, results of the best-fit model and analyses of observation well data was checked using AQTESOLV Pro 4.0 modeling software produced by HydroSOLVE, Inc.

3.3.4 Relief Well Performance

USACE would like to estimate the performance of relief wells over time to identify the need for future conditioning.

Engineering Manual 1110-2-1914, Design, Construction, and Maintenance of Relief Wells states that a relief well should be rehabilitated if specific capacity of the relief well drops by more than 20 percent of the specific capacity estimated when it was installed or its assumed baseline condition. A summary of pumping rates applied to each relief well, observed drawdown and the specific capacity for each flow rate is provided in Table 6 for use in future testing. Drawdown was influenced by nearby surging effort or development in approximately 20 of the relief wells during step-drawdown testing. Drawdown in Table 6 is corrected for surging influence by extrapolating the observed trend prior to influence of surging and is identified in Table 6. Drawdown influenced by development is not corrected in Table 6.

3.3.5 Sand Testing

Sanding contents measured in the discharge water ranged from 0.0 to 1584.9 ppm. Relief Wells 285+75 and

273+30 had measured sand contents that exceeded 1000 ppm, and one additional relief well (Relief Well 171+00) had a sand content above 500 ppm. Twenty-one relief wells, including the three mentioned above had a measured sanding rate higher than 100 ppm. A trend in sand content was not observed that corresponded to increasing or decreasing pumping rates. Sand observed in the discharge water and during air surging was observed to be a poorly-graded fine sand but a grain size analysis was not performed.

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 17 of 672

4.0 DISCUSSION AND CONSIDERATIONS

The conditioning efforts performed in 2017 and 2018 significantly increased transmission of groundwater into the relief wells based on a brief comparison between the specific capacity estimated in 2010 (GFI 3, Table 3) and the specific capacity estimated for pump testing at similar flow rates for this report. On average, specific capacity was significantly higher in 2017 and 2018 after conditioning efforts were performed than those recorded in 2010.

If not already performed, consideration should be given to estimating potential flow scenarios from the relief wells based on the maximum height of water that could collect behind the dam. By applying the definition of efficiency that compares all linear (aquifer and well) losses to total losses, high efficiency rates have been estimated for pumping rates much larger than what is considered probable during high-water events, indicating that the conditions inside many of the relief wells and in the nearby aquifer at the time of pumping are not significantly inhibiting groundwater flow. Regardless of efficiency, the estimated hydraulic conductivity for the relief wells indicates that a significant amount of water can be transmitted from the aquifer through the relief wells. Flow meters with data loggers could be attached to some of the pipe branch type well heads to capture actual flows during high-water events to calibrate results.

Further study of the influence of biofouling should be considered to estimate the need for a treatment program and adjustment to the type of well head construction. It was observed that surface water easily entered relief wells that had invert lips close to the ground surface, which could be a source of contamination. Biofouling has the potential to decrease the transmission of groundwater into a well significantly over time and depending on the type can grow back quickly and encourage the formation of mineral deposition (Sterrett, 2007).

If biofouling is found to be a significant impediment to flow, chemical and biological testing is relatively easy to perform to specifically diagnose the cause of the biofouling. A chemical treatment program with various available biocides, for example, could then be implemented to limit performance degradation over time, and to reduce well maintenance/ redevelopment requirements in the future.

The rate of biofouling has the potential to decrease after development because relief wells that are not properly developed are also more likely to develop biofouling. Screen sections that are not transmitting flow are potential locations for enhanced bacterial growth. Most blockages of screens initiate with biological involvement or biofilm formation. If biological involvement is not removed, reformation of mineral blockages can occur further causing a loss of capacity. A biological and chemical analysis should be considered (Sterrett, 2007) to properly diagnose the cause of well efficiency deterioration. Consideration should be given to a comparison between the observed degree of biofouling shown in the down-hole videos to the velocity information provided in GFI 4.

High sanding rates observed at some relief wells may indicate the need for further development. Further analyses of the data could be completed to attempt a correlation between relief well responses to sanding rates. High sanding rates do have the potential to damage pump testing equipment in the future. Measurements regarding subsurface conditions were completed from the TOC for this report. The ground surface, especially in a cold climate, is constantly adjusting with respect to the relief well head. Future measurements for use in comparisons to baseline conditions should be made from the TOC if possible. A permanent mark should also be placed on the

TOC to indicate to future field staff where measurements should be taken.

Pumping Test Water Level Observations

While performing pumping tests on the 147 relief wells, water levels were observed to deviate from an expected downward trend over time in some relief wells. These are attributed to influence from near-by conditioning

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 18 of 672 activities such as air surging, continued development of the relief well during the pumping test and brief adjustments to pumping rates.

Influence from nearby conditioning activities that was observed in water levels during pumping tests is considered to be negligible. Brief fluctuations were observed; however, as the water levels recover when surging was suspended, a clear trend was observed in drawdown that matches with earlier periods of drawdown that were not influenced by surging. As stated in Section 3.3.3, the RMSE between the observed water levels and the best-fit models are acceptable. The analyses included relief wells where surging influence was observed. For future planning purposes, Table 7 is provided to convey when surging effort and step-drawdown testing were conducted commensurately, if there was observed influence from surging in water levels in the pumped well, and the distance between the pumped relief well and the relief well with the surging effort. As the field effort progressed across the project site, variability was observed in the distance between the pumping well and nearby surging that would influence water levels during the pumping test.

Rising water levels were observed in various relief wells during pumping tests, counter to the expected decrease in water levels with increasing pumping time. The rising trend is attributed to continued development of the well during pumping. A brief comparison of observed development in pumping tests to sanding rates was made;

however, no correlation was evident.

Brief adjustments to pumping rates were so frequent that they are therefore not identified specifically. The 10-inch submersible pump would start at a higher pumping rate when flow was initiated for each pump test due to manufacturer design and then quickly adjust down. Field staff, when starting a new flow rate, would often quickly adjust the flow as needed once after initiation, if the new flow was not adequate. Continued adjustment of flow rates complicates analyses of the data. There were two exceptions: in Relief Well 306+50 the sounding tube moved down but was quickly readjusted during the pumping test; an anomalous jump was observed that is not attributed to staff or equipment during the pumping test for Relief Well 316+85.

The range of spread observed in water level measurements recorded during the pumping tests was found to be appropriate based on the RMSE between the observed water levels and a best fit line for drawdown in step one drawdown (RMSE). The RMSE was estimated to range from 0.0031 to 0.072 feet for a sample of the relief wells;

10 percent of the relief wells tested or 15 in total. The wells chosen are the same wells that were used to evaluate the appropriateness of the best fit model in Section 3.3. The step one water level RMSE values for the sample of relief wells is included in Table 5.

5.0 LIMITATIONS AND USE OF REPORT

This report has been prepared exclusively for the U.S. Army Corps of Engineers, Alaska District, for the Chena

River Lakes Flood Control Project in North Pole, Alaska.

The work program was undertaken using industry standard development and testing procedures, standard mathematical models for groundwater flow to a well, and the analysis presented is based on commonly accepted well hydraulic theory. Two different commonly accepted definitions of well efficiency were presented here, for purposes of completeness and discussion. The two methods yield, in some cases, significantly different estimates of efficiency. However, given that both methods are commonly applied in practice, Golder makes no judgement on whether one should be favored over the other for the purposes expressed in this report, and no warranty is expressed or implied.

AM#1 MOOSE CREEK DAM SAFETY MODIFICATION

NORTH POLE, ALASKA

Section 35 73 20.00 29 Attachment (2) Page 19 of 672

CHL023

Section 35 73 20.00 29 Attachment (2) Page 20 of 672

December 20, 2018 1780780

This work follows the standard of care expected of professionals undertaking similar work in Alaska under similar conditions. No warranty expressed or implied is made.

6.0 CLOSING

Golder appreciates the…