Section_J_Attachment_6a_2022_CAVE_Geophysical_Investigative_Report.pdf

Geophysical Investigative Report

CAVE Geophysical Survey Services Carlsbad Caverns National Park

Carlsbad, NM

Contract Number: 140P1322P0045P

November 8, 2022 for

U.S. National Park Service Carlsbad Caverns National Park

3225 National Park Hwy Carlsbad, NM 88220

Respectfully Submitted

Doug Laymon, PG Roy Bowling Nathan Collier, PE

Senior Geophysicist / Group Manager

Geophysicist President, Collier Geophysics

Geophysical Investigative Report: Cave Geophysical Survey Services, Carlsbad Caverns National Park, Carlsbad, NM

Contract: 140P1322P0045P

Executive Summary Nov-22 Page | 1

EXECUTIVE SUMMARY

Collier Geophysics (Collier) was contracted (140P1322P0045P) by the National Park Service (NPS) to conduct geophysical surveying over 6 Areas, comprising approximately 35 acres, near the Park Service facilities at Carlsbad Caverns National Park, NM. The survey objective was to assess and map the risk of voids and karst features occurring in the shallow subsurface within each of the 6 Areas. The primary goal of this study was to support the NPS Facilities Management program with planning for future below-grade utility work. Collier met the project objectives by deploying three geophysical methods across the site. Ground Penetrating Radar (GPR) was used to map the shallow karst potential in the paved and other accessible roadway portions of the Areas, Frequency Domain Electromagnetics (FDEM) was used as a supplementary method to shallowly investigate the unpaved portions of each Area, and Electrical Resistivity Tomography (ERT) was used to investigate the deeper karst potential at each Area.

Geophysical anomalies which comprise: 1) high amplitude reflections and diffractions in the GPR data, 2) low conductivity zones in the FDEM data, and 3) highly resistive zones in the ERT data were interpreted in each of the Areas. These anomalous zones and features are indicative of increased potential for near surface karst, including open voids.

The locations, extents, and estimated depths of these geophysical anomalies were integrated into a combined risk assessment for each Area. The combined risk maps present zones where the geophysical results would suggest an either high or more moderately elevated risk of karst features as related to the project objectives.

For Areas 1 - 4 near the park service buildings and infrastructure, as well as Area 6 at the quarry site, several high risk zones were interpreted, along with multiple moderate risk zones. For locations where dense infrastructure and existing subsurface utilities occur, geophysical data quality was reduced, and risk mapping was less confident in those locations. Area 5 proved to contain (at least at the ground surface) large amounts of dense carbonate rock which caused difficulties in data acquisition, and in turn lowered confidence in results. However, broader geophysical trends as well as onsite geologic observations led to the interpretation of several risk zones in Area 5 as well.

Contract: 140P1322P0045P

Table of Contents Nov-22 Page | 2

TABLE OF CONTENTS

EXECUTIVE SUMMARY

TABLE OF CONTENTS

LIST OF FIGURES

LIST OF TABLES

1.0 INTRODUCTION

1.1 PROJECT OVERVIEW

1.2 GEOLOGIC BACKGROUND

1.3 FIELD CONDITIONS

2.0 METHODS

2.1 GROUND PENETRATING RADAR (GPR)

2.1.1 GPR Method

2.1.2 GPR Survey Design

2.1.3 GPR Data Processing and Interpretation

2.2 FREQUENCY DOMAIN ELECTROMAGNETICS (FDEM)

2.2.1 FDEM Method

2.2.2 FDEM Survey Design

2.2.3 FDEM Data Processing and Interpretation

2.3 ELECTRICAL RESISTIVITY TOMOGRAPHY (ERT)

2.3.1 ERT Method

2.3.2 ERT Survey Design

2.3.3 ERT Data Processing and Interpretation

2.4 COMBINED RISK MAPPING

3.0 RESULTS

3.1 AREA 1 RESULTS

3.2 AREA 2 RESULTS

3.3 AREA 3 RESULTS

3.4 AREA 4 RESULTS

3.5 AREA 5 RESULTS

3.6 AREA 6 RESULTS

4.0 CONCLUSIONS

5.0 LIMITATIONS DECLARATION

6.0 REFERENCES CITED

7.0 DIGITAL DELIVERABLES

APPENDIX A: COMBINED RISK ZONE MAPS

APPENDIX B: DETAILED GEOPHYSICAL RESULTS

Contract: 140P1322P0045P

List of Figures Nov-22 Page | 3

LIST OF FIGURES

Figure 1: Geophysical survey areas Figure 2: Geologic map of the survey areas at Carlsbad Caverns National Park (modified from Skotnicki (2021)) Figure 3: A teepee structure located in the parking lot of the Carlsbad Caverns Visitor Center (from Skotnicki, 2021) Figure 4: Limited/no access portions of Areas 3 and 5 Figure 5: Examples of site conditions during data acquisition Figure 6: (a) The ImpulseRadar CrossOver CO1760 GPR System and (b) Emlid Reach RS2 RTK GPS system Figure 7: CMD-Explorer FDEM instrument at the Park Figure 8: Super Sting R8 Deployment Figure 9: APWA uniform color code for utility locating

Contract: 140P1322P0045P

List of Tables Nov-22 Page | 4

LIST OF TABLES

Table 1: GPR Processing and Interpretation Workflow Table 2: FDEM Processing and Interpretation Workflow Table 3: ERT Processing and Interpretation Workflow

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 5

1.0 INTRODUCTION

1.1 Project Overview

This report provides a summary of geophysical surveys that were conducted by Collier Geophysics, LLC (Collier) within Carlsbad Caverns National Park (the Park), for the National Park service (NPS), under contract number 140P1322P0045P. Collier performed geophysical surveys at 6 Areas in the vicinity of the Park Visitor Center and Park maintenance and service buildings (Figure 1). The objectives of these surveys were to assess the potential for near subsurface karst and void features that may impact future subsurface utility work at the Park.

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 6

Figure 1: Geophysical survey areas

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 7

The primary targets for this investigation are karst features/voids 6 feet diameter or larger down to a depth of at least 20 feet below ground surface (bgs), as well as any voids larger than 50 feet in diameter down to a depth of 50 feet bgs.

To achieve the survey objectives, Collier employed a multi-method geophysical approach. Ground penetrating radar (GPR) was utilized to image the shallow subsurface in the paved portions of the survey Areas. Frequency Domain Electromagnetic (FDEM) terrain conductivity measurements supported the GPR surveys in the paved locations and provided shallow coverage outside of paved sites. For deep imaging of larger features, Electrical Resistivity Tomography (ERT) profiling was performed at each of the 6 Areas.

The following sections include a regional summary of the geology, followed by a methodological description of the geophysical surveying and theoretical aspects of the geophysical techniques utilized. Results of the geophysical surveying are then presented along with relevant discussions of the data and identified anomalies.

1.2 Geologic Background

The geophysical survey is mainly over rocks of the Tansill formation, and partly over the Capitan Formation (Figure 2). The Permian Tansill Formation consists of light gray dolomite and minor thin dark tan siltstone beds (Skotnicki, 2021). The dolomite beds are mostly thick-bedded and massive. The formation was deposited behind the Capitan Reef in a shallow-water/lagoonal setting, as were the underlying Yates and Seven Rivers Formations. Close to exposures of the Capitan Formation the unit contains some teepee structures, but fewer than in the underlying Yates Formation. Teepee structures are sedimentary structures interpreted to represent formation in peritidal environments.

Teepees are largely the result of evaporation of water and subsequent precipitation of minerals within sediment, resulting in expansion and buckling to form a teepee-like shape. They can be relatively large is size, as shown in (Figure 3) below, photographed in the parking lot of the Carlsbad Caverns Visitor Center. The Tansill Formation directly behind the Capitan Formation is approximately 200 to 400 feet thick.

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 8

Figure 2: Geologic map of the survey areas at Carlsbad Caverns National Park (modified from Skotnicki (2021))

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 9

Figure 3: A teepee structure located in the parking lot of the Carlsbad Caverns Visitor Center (from Skotnicki, 2021)

The Capitan Formation formed in Permian time as part of a massive “Great Barrier Reef” type structure composed of reef front limestones and dolomites (King, 1948; Pray, 1985). The Capitan displays weakly-developed inclined layering that dips southeastward between ~15° and 30°. These are thought to be bedding surfaces that dip down into the Delaware Basin. In outcrop, most exposures appear massive and structureless. A faint brecciated texture is visible locally where angular clasts of dolomite of all sizes are strongly cemented by different generations of carbonate. Coarse-grained light yellow calcite spar commonly fills dissolution fissures and cracks (Skotnicki, 2021)

The Whites City Fault is visible as a linear feature that parallels the bottom of the slope of the Capitan Formation, (Figure 2). Besides the lineation, no topographic scarp is visible, indicating that there has been no recent significant movement along this structure. As a result, most of the fault itself is shown as being obscured by younger Quaternary alluvial deposits to the south. Skotnicki (2021) considers the fault to be late Tertiary or early Quaternary in age. Back reef rocks of the Tansill and Yates Formations are gently folded along several east-west trending anticlines and synclines located north of the fault. The fold axes are roughly parallel to the folds and thrusts within the Marathon fold belt to the south, which suggests that the two fold belts may have been created at the same time.

Karst features in the project site include Carlsbad Caverns and associated caves, along with shallower features caused by meteoric dissolution. The caverns primarily occur at depths of approximately 500 to 1,000 feet below ground surface, and formed during the late Miocene and early Pliocene (~12-4 Ma ago). They originated mainly from sulfuric

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 10 acid dissolving limestone of the Capitan Formation and backreef strata represented by the Seven Rivers, Yates, and Tansill Formations (Kirkland, 2014).

For this project, the primary geophysical targets are the near surface karst features in the Tansill Formation, and in Area 5, the Yates and Capitan Formation as well. Smaller near-surface features are likely related to the same processes that created the larger, deeper caves in the area. These types of karst features are likely caused by past rising groundwater and sulfuric acid dissolution, similarly to Carlsbad Caverns. Heretofore unknown, and even deeper caverns may exist in the survey Areas, but with lower probability of detection with the geophysical survey methods used in this project.

Structural jointing as well as sedimentary structures such as the aforementioned teepees (Figure 2) in the Tansill formation may be regions additionally predisposed to classical shallow meteoric karst formation.

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 11

1.3 Field Conditions

Field work for this investigation was performed at the Park between August 2 and 16, 2022. Access and coordination of data collection activities amongst Park operations were organized with Park staff. Of the 6 survey Areas, portions of Areas 3 and 5 had limited or no access for geophysical surveying during the time of the field work. These portions included (Figure 4):

The central portion of Area 3, which was occupied by onsite construction staging.

Limited GPR surveying was possible in this location, but no access was possible for FDEM or ERT surveying.

The central portion of Area 5, where the crew encountered steep slopes, loose rock and cliff faces. No ERT or FDEM surveying was performed in this location for crew safety.

Figure 4: Limited/no access portions of Areas 3 and 5.

Contract: 140P1322P0045P

1.0 Introduction

Nov-22 Page | 12

Outside of the above mentioned locations, data for each of the three geophysical methods were collected as defined in the Technical Approach portion of Collier’s proposal for this contract. Ground conditions varied from paved/gravel roadways to unimproved desert landscapes containing variable densities of vegetation (Figure 5).

Weather consisted of seasonal tempera -precipitation.

Figure 5: Examples of site conditions during data acquisition.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 13

2.0 METHODS

The following section describes the concepts, data acquisition parameters, and processing and interpretation workflows for of each of the three geophysical methods (GPR, FDEM, and ERT) utilized for this project.

2.1 Ground Penetrating Radar (GPR)

2.1.1 GPR Method

The ground penetrating radar (GPR) method is based on the recording of reflected electromagnetic waves that are acquired using a transmitter (TX) and receiver (RX) antenna pair. The reflections from subsurface interfaces are stored as a time series (known as traces) representing reflection “strength” or amplitude as a function of travel-time. Travel-time is the time required for the transmitted energy to travel down into the subsurface, reflect off an interface of contrasting electromagnetic impedance, and travel back up to the surface to be recorded by the GPR receiver. Therefore this time is often referred to as two-way travel-time (TWT).

As the GPR instrument is moved to different locations on the surface, different reflection series are recorded which represent changes in the subsurface EM impedance/reflectivity distribution. GPR signals are sensitive to the presence of a variety of subsurface materials including: buried objects (metallic and nonmetallic), voids, water saturated sediments, and geologic boundaries.

For this project, the targets for the GPR surveys were karst features in the shallow subsurface. Karst features such as fractures, vugs, and open voids, represent a dielectric permittivity ( ) contrast for GPR signals. The relative dielectric permittivity ( ), is the variation in permittivity from the vacuum of free space. For limestone rocks, typically ranges from 4-8, (Davis and Annan, 1989) and in the air space within any present karst features = ~1. This permittivity contrast provides the electromagnetic impedance contrast to reflect GPR signals from the rock-air boundaries that exist in karst regimes. For larger open voids that could exist, GPR is most sensitive to the tops of these features, as a larger percentage of the incident GPR energy is reflected off the tops of voids, while a smaller portion transmits through to the bottom of the void-space.

Therefore, the depths and extents of the tops of karst features are most often interpreted from GPR data, as the bottom/base of features are often poorly resolved.

In cases where near surface soils or fill materials are present however, the depth of penetration of GPR signals can be limited. This is because soils and some fill materials

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 14 contain significant clay fractions, which increase the bulk electrical conductivity of the material and attenuate GPR signals. In zones of poor penetration (i.e. less than 3 feet) the interpretation of karst features can be difficult or not possible.

2.1.2 GPR Survey Design

The GPR surveying was performed using an ImpulseRadar CrossOver CO1760 GPR system (Figure 6). The CO1760 antenna is a dual frequency antenna which is able to transmit and receive GPR signals at both 170 MHz and 600 MHz simultaneously.

Therefore both shallow, higher resolution, GPR data with 600 MHz signals and deeper, but lower resolution, data with 170 MHz signals was able to be acquired at the same time. GPR surveys were performed over all the accessible paved and gravel roadway portions of Areas 1-4. Additionally, GPR surveying was performed in two accessible non-roadway locations adjacent to Park Service buildings in Area 1. GPR data were collected in parallel scanlines at nominally 3-foot spacing.

Geospatial positions of the GPR data were collected using an Emlid Reach RS2, multi-band Real-time Kinematic (RTK) Global Positioning System (GPS) system, achieving sub-foot horizontal precision. Real-time positons of the GPR antenna were streamed directly into the CrossOver CO1760 and synced to provide positions for each GPR trace.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 15

Figure 6: (a) The ImpulseRadar CrossOver CO1760 GPR System and (b) Emlid Reach RS2 RTK GPS system.

2.1.3 GPR Data Processing and Interpretation

GPR processing was performed using a set of in-house software routines in a workflow designed to improve the signal-to-noise ratio of the data and combine the low (170 MHz) and high (600 MHz) datasets for efficient interpretation. Interpretation of the GPR data was performed using Geolitix, by Geoltiix Technologies Inc. Table 1, below describes the processing and interpretation workflow for the GPR data.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 16

Table 1: GPR Processing and Interpretation Workflow

Processing Step

Process Details

1 Time Zero Correction

Parameters (170 MHz)

Parameters (600 MHz) Trimming of top of traces in each scanline to align first arrival with time zero

21 Sample Trim

21 Sample Trim

Exponential Gain

= 1

= 2

= 1

= 2

An along-trace scaling to compensate for the effects of amplitude loss due to signal attenuation losses and geometrical spreading

Background Subtraction

Average Trace Removal

Average Trace Removal

Removes the average of all traces from each trace in a scan line.

Performed to mitigate the effects of the TX/RX interference and antenna self-coupling

Frequency Filtering

Bandpass Frequency Filter:

Low-Cut:

400MHz

High-Cut:

255MHz

Bandpass Frequency Filter:

Low-Cut:

250MHz

High-Cut:

900MHz

Filtering in the Fourier domain to suppress signals occurring at frequencies outside of the center-frequency range of the GPR antenna, including signals caused by inductive effects and sensor amplitude saturation

Merging of 170 MHz and 600 MHz Data

For each scanline, combining the 600 MHz data and 170 MHz data into a single profile, by seamlessly integrating the 170 MHz data traces into the 600 MHz data traces at the signal noise floor travel-times of the 600 MHz data

Import of processed and merged GPR data into Geolitix for visualization and anomaly interpretation

Visualization and interpretation of minor and significant GPR anomalies from each frequency-combined scanline for each Area.

Combined frequency scanlines allowed for the interpretation of both shallow and deep anomalies concurrently.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 17

GPR anomaly interpretation consisted of identifying relatively high-amplitude reflections and diffraction features within the scanlines. Diffractions are parabolic-shaped reflections which occur at local EM impedance contrasts in the subsurface such as the edges of voids, fractures, vugs, ect. and also at the tops of buried object such as pipes, boulders, cobbles, etc. In karst regions, dense sets of fractures and rugose tops of dissolution voids produce complex overlapping sets of diffractions. These diffractions have relatively higher amplitudes compared to surrounding features, due to the EM impedance contrast between the air-filled fractures or voids and the limestone host rock.

Additionally, and in particular for the low-frequency, lower resolution data, subsurface voids can be presented in GPR data as lateral amplitude contrasts where signal ring-down is observed. In the case of a void or open fracture whose thickness is below the signal wavelength, no clear diffractions may be observed, but the EM impedance contrast produces high relative amplitudes, and signal reverberations from the top and bottom of the void produce ring-down in the GPR scanlines.

GPR anomalies were interpreted as either minor or significant. Significant anomalies are features which had signal characteristics which would be associated with open voids and dense karst features, such as high amplitude diffractions, and strong signal ringing.

Minor anomalies exhibit signals which can be attributed to karst features as well, but are either lower in amplitude, smaller in extent, and/or are more isolated than features which would compose significant anomalies. Concisely, interpreted significant anomalies are indicative of open, dense, and interconnected fractures, as well as open voids, and minor anomalies are indicative of fractured rock, vugs, and possibly small–scale voids.

There were some locations however, where interpretation of features at depth was made difficult due to limited GPR signal penetration. Some portions of the roadway and other paved regions were observed to have low signal penetration. These locations are likely underlain by engineered fill with some clay fraction, causing attenuation of the GPR signals, thereby limiting the depth to which features could be identified.

The minor and significant GPR anomalies were interpreted as a function of geospatial position as well as signal travel-time. To convert travel-time to depth below ground surface (bgs), a bulk GPR signal velocity of 0.299 ft/ns (0.091 m/ns) was used. This velocity was determined by assessing the geometry of isolated diffractions. The 0.299 ft/ns GPR velocity is an average from assessing multiple diffractions across the various survey Areas, and is representative when compared to published values for limestone lithologies (Cassidy, 2009, Davis and Annan 1989). It is important to note that local

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 18 velocity variability likely occurs across the site, but it is not expected that the variations would significantly impact the general depth estimates of the interpreted anomalies.

2.2 Frequency Domain Electromagnetics (FDEM)

2.2.1 FDEM Method

A frequency domain electromagnetic (FDEM) instrument consists of at least one pair of transmitting and receiving coils. A primary magnetic field of a constant frequency is generated using an alternating current in the transmitter coil, and a secondary magnetic field is detected in the receiving coil as a result of the interaction of the primary field with the subsurface. The FDEM instrument allows for simultaneous measurements of both the in-phase and quadrature (orthogonal phase) components of the secondary magnetic field. The in-phase component is measured in parts-per-thousand (ppt) of the amplitude of the primary magnetic field. The in-phase response is primarily sensitive to magnetic susceptibility, generally due to the presence of metallic or ferromagnetic material in the subsurface. The quadrature component (90-degrees out of phase with the primary signal) is primarily sensitive to electrical conductivity (the reciprocal of resistivity), due to changes in lithology, karst features, moisture, and/or fines (clay) content. The quadrature response is calibrated and measured as apparent bulk conductivity in millisiemens per meter (mS/m). Note that these are the primary sensitivities, but that both components can be affected by buried objects and other geologic features. The effective depth of sensitivity of the FDEM method is a function of the separation between the transmitter and receiver, the antenna orientation, the frequency of the primary field, and the bulk electromagnetic properties of the subsurface.

For this project the FDEM target was near-surface conductivity variations that may be attributed to karst features. Open air-space has very low conductivity (~10 / ).

Near-surface soils and limestones at this site are likely relatively more conductive than air (0.1 5 / ) (Davis and Annan, 1989). The increase in void space through fractures, vugs, and open voids in karstified limestones will decrease the relative bulk electrical conductivity and provide a contrast to the FDEM survey. Very dense and dry limestones, however, may also have very low conductivities, and FDEM surveying may not be able to differentiate karstfied limestones from those which are dense or dry.

It is important to note that apparent conductivities measured from FDEM surveying are very sensitive to interference from nearby metallic objects, as these objects effectively have infinite conductivity and violate several assumptions of the FDEM method.

Therefore, in locations adjacent to buildings, cars, signposts, reinforced concrete, Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 19 metallic subsurface utilities, etc. FDEM measurements of soil and rock apparent conductivity become unreliable.

2.2.2 FDEM Survey Design

FDEM surveying was performed using a CMD-Explorer, by GF Instruments (Figure 7).

The CMD-Explorer consists of a boom with three sets of FDEM coil pairs, at three separations; 1.48 m (4.86 ft), 2.82 m (9.25 ft), and 4.49 m (14.73 ft) using a common transmitter coil. The effective depth of sensitivity of the FDEM method is a function of the antenna spacing between the transmitter and receiver, the antenna orientation, the frequency of the primary field, and the bulk electromagnetic properties of the subsurface.

Data were acquired using a vertical dipole configuration, which results in the greatest depth of investigation. The depth of investigation is not precise, but as a rule of thumb when using a vertical dipole orientation, is approximately equal to the antenna separation but can extend to depths of 1.5 times the antenna separation. In the case of the CMD-Explorer the values recorded would correspond to the bulk electromagnetic properties in the upper (approximately) 5 feet, 10 feet, and 15 feet, of the subsurface respective to each antenna separation. FDEM data were collected at a rate of 10 Hz, using a primary field frequency of 10 kHz at all three receiver antenna separations simultaneously.

Figure 7: CMD-Explorer FDEM instrument at the Park

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 20

FDEM surveys were performed over all the accessible portions of Areas 1-6. Surveying was completed at nominally 10-foot line spacing where possible. As the CMD-Explorer instrument has a width of 4 meters, many locations required non-organized surveying between buildings, around obstacles and through vegetation. Therefore some portions of the survey Areas have variable FDEM line spacings.

Geospatial positions of the FDEM data were collected using an additional Emlid Reach RS2 rover in the same RTK system as the GPR data. Real-time positons were streamed directly into the CMD-Explorer and synced to provide positions for the center of each FDEM measurement

2.2.3 FDEM Data Processing and Interpretation

Raw FDEM data were exported in tabular format using CMD Data Transfer, version 1.6.2, by GF Instruments. Positions for each measurement were interpolated for each record from GPS positions using the data transfer software. The data were then processed using Geosoft Oasis Montaj, version 10.0 by Seequent (Geosoft), a processing and data visualization software suite used for analysis of geophysical data sets. The data were processed to remove dropouts and sources of noise. Additional normalization and leveling was also selectively applied to ensure continuity between measurements. Final maps of the Coil 1, 2, and 3 apparent conductivity datasets were imported into QGIS version 3.10.11 and interpreted for conductivity anomalies. Table 2 below describes the processing workflow performed on the FDEM data.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 21

FDEM anomaly interpretation consisted of identifying zones of relative low apparent conductivities in the three coil datasets for each site. Again, Coil 1 data represents the shallow (~5 ft) investigation depth, Coil 2 the medium (~10 ft) and Coil 3 the deep (~15

ft) investigation depth. Anomalously low apparent conductivities are zones where shallow subsurface karst features may be present which lower the effective bulk apparent resistivity. As was noted in Section 2.2.1, however, interference from metallic objects near the instrument prevented interpretation of subsurface karst-related anomalies. For the FDEM datasets at each Area, portions which were obviously affected by metallic interference were masked. Unfortunately, in many of the locations adjacent to buildings and parking lots (Areas 1, 2, and 3 in particular) showed signs of interference.

Table 2: FDEM Processing and Interpretation Workflow

Processing Step

Process Details

Reading the raw FDEM data into a Geosoft database

Data for each site were loaded into its own unique database and separated by conductivity (quadrature) for each of the three coil separations

2 Cleaning data Manual removal of erroneous data transects, and editing of select data values that represent data recording errors

3 Leveling of data

When needed, each component for each coil separation was individually leveled to have consistent mean values between separately collected data files for each site

4 De-spiking/Smoothing

Removing high frequency noise by applying first a non-linear filter, then a b-spline filter to each coil.

This process suppressed high amplitude spikes and small-scale variations final gridded data maps

Two-dimensional gridding of the data

Minimum curvature gridding method applied

Import of processed and gridded data into QGIS for visualization and anomaly interpretation

Plotting and displaying apparent conductivity (quadrature) component of FDEM data for each coil separation as plan view maps and interpretation of anomalous low conductivity zones.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 22

In unaffected locations, the FDEM interpretations were used to support the coincident GPR and/or ERT interpretations.

2.3 Electrical Resistivity Tomography (ERT)

2.3.1 ERT Method

The ERT method is used to characterize subsurface lithology and/or materials in terms of electrical resistivity. ERT incorporates the injection of an electrical current into the ground through a pair of electrodes (current electrodes) while simultaneously measuring the potential or voltage between an offset electrode pair (potential electrodes). The subsurface apparent resistivity is then calculated from the measured voltages, according to electrode geometry. This measured apparent resistivity represents the bulk resistivity of earth materials where the majority of injected current flows.

The geometry between 2 current electrodes and 2 or more potential electrodes defines an array. The distance between the potential electrodes is directly related to resistivity measurements with depth. The amount of current injected and distance between the current electrodes determines the investigation depth, i.e., larger spacing forces more available current to flow at depth. Calculated apparent resistivity values for a set of measurements are then used in a tomographic inverse modeling scheme to build a best fit model of the true resistivity distribution of the subsurface.

Electrical resistivity (the reciprocal of conductivity) is a material property which is diagnostic of the type of geologic material present. Unsaturated soils have higher resistivity (lower conductivity) than saturated soils. Sand and gravel with minimal silt/clay content have higher resistivities than soils with high silt/clay content. Sandstone, limestone, and granite typically have higher resistivity values than shale and siltstone.

Materials saturated with saline or brine waters have very low resistivity values and can be analogous to highly conductive clays.

For this investigation, karst features will have higher resistivity values due to the increased presence of open void space (air), which has very high resistivity (~10 ). In the case of soil/clay filled voids however, karst features may have relative lower resistivity due to the conductive nature of soils and clays. Similar the FDEM measurements, very dense and dry limestones can also have very high resistivities, which not only affects the ability of the ERT instrumentation to inject sufficient currents for stable measurements, but may make karst feature identification difficult in such locations.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 23

ERT data are susceptible to interference from objects that act as subsurface conductors and draw injected current away from the ERT array in ways not related to geologic structures. Buried subsurface utilities are the primary source of noise in ERT measurements.

2.3.2 ERT Survey Design

The ERT survey was performed using an Advanced Geophysical Systems Inc. (AGI) Super Sting R8 8-channel multiple electrode resistivity imaging system (Sting R8). The survey equipment consisted of a transmitter/receiver and cables capable of utilizing up to 56-takeouts for electrodes (Figure 8).

Figure 8: Super Sting R8 Deployment

The ERT lines for this survey consisted of an electrode array with 56 evenly spaced electrodes placed at 10 ft or 20 ft intervals. Line 1 in Area 1, Line 10 in Area 2, and Lines 1, 2, 3, 4, and 5 in Area 5 used 20 ft electrode spacing and all other lines used 10 ft electrode spacing. The electrodes and cables for each line were connected to the Sting R8, which was positioned between electrodes 28 and 29. Some lines required extending the array beyond 56 electrodes using a “roll-along” survey. When “rolling” the survey line, after the first 56 channel array is complete, the first 28 electrodes are moved ahead of the existing array in a leapfrog manner and a new set of readings is taken to extend the length of the line as needed. The ERT surveying for this project utilized a dipole-

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 24 dipole array in combination with a strong gradient array. The dipole-dipole array is more sensitive to horizontal changes in the subsurface such as would be expected from the presence of karst features. The strong gradient array has high signal to noise ratios, and was used to provide more stable readings at greater depths.

The ERT lines were positioned across Areas 1 – 6 at nominally 50 foot line spacing, and in accordance with the planned line layout in the Technical Approach portion of Collier’s proposal for this contract. In roadways, parking lots, and near buildings and concrete sidewalks, ERT lines were moved and positioned to locations that were accessible. In locations where survey lines crossed outcropping limestone, pre-drilling of electrode placements was performed using an impact hammer-drill. For the lines in Area 5, nearly all electrode placements were on outcrop and were pre-drilled. Final array geometry for each line was determined in the field by satisfying four requirements: time efficiency, data quality, vertical/horizontal resolution capacities, and desired investigative depth.

All placed electrodes were watered with saltwater before data collection. This was done to reduce contact resistance in highly resistive soil and outcrop locations. A contact resistance test was performed before each data acquisition. Electrodes with very high contact resistance were replanted, moved, and/or re-watered. This was done until the contact resistance between electrodes was consistent, or no improvement was observed. It should be noted that in locations where electrodes were drilled into outcrop, very high contact resistances remained, even after liberal dosing with saltwater.

Geospatial positions of each electrode for each line were recorded using a GNS3 Geode GNSS receiver made by Juniper Systems achieving sub-meter accuracy. Precise elevation data for each electrode, which is a necessary input into the tomographic modeling, was derived from a 1-m resolution digital elevation model (DEM) of the site obtained though the USGS 3D Elevation Program (USGS 3DEP).

2.3.3 ERT Data Processing and Interpretation

The collected ERT data were processed in EarthImager by AGI using a standard workflow to remove erroneous and noisy data and prepare the raw apparent resistivity values for tomographic inversion. Inversion was performed to maximize resolution and provide good fit between modeled and measured data. The resulting 2D ERT cross sections obtained for each line were interpreted for anomalous resistivity zones using Surfer version 17, by Golden Software, and Geosoft. The table below details the ERT processing and interpretation workflow.

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 25

Table 3: ERT Processing and Interpretation Workflow

Interpretation of the ERT data was based on contour plots of inverted resistivity data for each line in each Area. Anomalously high resistivities for this project were considered to be over 30,000 Ohm-m. Modeled resistivities falling at or above this contour value were interpreted to be indicative of subsurface karst features such as open voids, large vugs, and large fracture sets. Moderately high resistivity zones were considered to be those falling between 10,000 Ohm-m and 30,000 Ohm-m. These zones may also indicate the presence of karst features but may be of smaller extent and may have less open void space. These anomalous zones were then binned into categories comprising those features within the upper 20 ft bgs and those between 20 and 50 ft bgs. While some anomalous zones exist deeper than 50 ft bgs, they were not specifically categorized in

Processing Step

Process Details

1 Remove noisy data.

High contact resistance resulted in noisy data. Data were removed to improve inversion stability

2 Georeferenced data

Lat, Long, and Elevation data was tied into the resistivity data to get accurate survey geometry and apparent resistivity calculations

3 Invert apparent resistivity data

Resistivity data was inverted in EarthImager to acquire a 2-D resistivity cross-section for each line in each Area

4 Identify anomalous resistivity zones

Areas modeled to have resistivity values greater than 10,000 ohm-m and 30,000 ohm-m were interpreted as moderate and highly anomalous zones, respectively

5 Georeference resistivity anomalies

Anomalies in the 2-D cross sections were exported and georeferenced in Geosoft in order to display the anomalies in plan-view maps for each Area

Categorize anomalies into depth zones.

The georeferenced resistivity data was binned into two zones. One for depths up to 20ft bgs, and another for depths between 20 and 50 ft bgs

Contract: 140P1322P0045P

2.0 Methods

Nov-22 Page | 26 the interpretation of the ERT data. Collier understands that features greater than 50 ft bgs are of less risk to the planned utility improvements that this study was designed to address. It must be noted that the presence of dense, very low porosity limestone rock has the ability to register more than 30,000 Ohm-m. As such, in certain locations (as detailed in Section 4) broad zones with highly anomalous resistivities may be due to the occurrence of such dense limestone rather than exceedingly large voids in the shallow subsurface.

In addition to the anomalously high resistivities, modeled resistivities below 50 Ohm-m were interpreted, and are presented in the plan views of the resistivity anomalies. These low modeled resistivity values, where they occur in the shallow subsurface may be associated with soil and/or clay filled voids, or heavily weathered limestones which may be undergoing dissolution.

Lastly, while the data processing workflow was designed to remove data contaminated by noise, some features in the ERT sections are likely spurious and artificial. In particular where ERT lines are adjacent to, or cross buried subsurface utilities, measured apparent resistivities can be contaminated by interference. As such, the resulting modeled ERT sections can display localized extremes which are the result of the presence of a buried utility and not necessarily karst features.

2.4 Combined Risk Mapping

To synthesize the interpretations from all three methods (GPR, FDEM, and ERT) in each of the surveyed Areas, Collier performed a manually interpreted combined risk mapping.

This risk mapping incorporates the geophysical interpretations and their relative importance, as well as direct site observations and external information into a two zone classification scheme. The risk zones are as follows:

Interpreted High Risk Zones – These zones contain one or more significant and/or major anomalies from one or more geophysical methods. These zones have been interpreted to have significant potential for karst features, including open voids. Features may be large (> 6ft in diameter) and/or within the upper 20 ft of the ground surface.

Interpreted Moderate Risk Zones – These zones contain minor or moderate geophysical anomalies. These zones may contain isolated/sparse significant and/or major anomalies. These zones have been interpreted to have the potential for karst, which may include voiding, but may mainly include smaller