Technical Articles, Vol 27,1 Focus on Latin America



Jared D. Abraham, Aqua Geo Frameworks, LLC., Fort Laramie, WY, USA
Theodore H. Asch, Aqua Geo Frameworks, LLC., Georgetown, CO, USA
Paul Ivancie, Ivancie Geosciences, LLC., Denver, CO, USA

Airborne Electromagnetic (AEM) mapping can provide critical information to a Managed Aquifer Recharge (MAR) program. However, for AEM to be useful, the electromagnetic contrast of the materials needs to be such that the imaging of the critical components of the earth materials for MAR (i.e., sand and gravel versus clay and silt) can be achieved. AEM has seen growing use within the United States (US) over the past 20 years and several of these projects have been specifically for MAR programs. The elements of the selection of an AEM system and the way in which those data are used to create a hydrogeological framework are the key steps in the usefulness of the AEM. Prior geological investigations including boreholes and/or any previous geophysical studies are critical for integration into any AEM based mapping program. With potential climate fluctuations and large runoff or flood events, opportunities exist for utilizing surplus water as recharge. However, to take advantage of these large flow events in the South Platte River, the infrastructure and locations need to be preplanned. Utilization of irrigation canals as buffers to high runoff has been used historically to mitigate infrastructure damage. By utilizing the stretches of the canals that will contribute leakage to managed aquifer storage areas, savings can be realized in infrastructure construction costs. MAR projects can be a key to the sourcing of water by municipal groundwater suppliers in times of reduced surface water flows. Three examples will be shown including two in Colorado and one in Nebraska along the South Platte River. These data were acquired with both frequency domain and time domain systems. Pre-existing geological and hydrogeological data have contributed to the construction of 3D hydrogeological frameworks to guide the delineation and management of managed aquifer recharge sites. Key features that were mappable with the AEM included bedrock configuration and formation identification, aquiclude lateral extent and thickness, near-surface permeable sands zones connections to deeper aquifer materials including sands and gravels, and identification of hydraulic connections of MAR zones to the South Platte River.

Capacitive electrical resistivity: an alternative non-invasive method for permafrost monitoring

Sara Bazin, UiO/NGI, now at Institut Européen de la Mer, UBO, Brest, FRANCE

Shakir G. Syed, Oslo University (UiO), Oslo, NORWAY

Graham L. Gilbert, Norwegian Geotechnical Institute (NGI), Oslo, NORWAY

Bernd Etzelmüller, Oslo University (UiO), Oslo, NORWAY

The Svalbard archipelago is located in the high arctic and its mean annual air temperature has increased between 3 °C and 5 °C during the last 40 to 50 years. Understanding how its permafrost may be affected by the current warming trend has become a research priority. Thermal logging in boreholes enables long-term monitoring of the ground thermal regime but is expensive and only allows point measurements. Electrical resistivity tomography (ERT) is very effective in mapping frozen soils due to the strong resistivity contrast between ice and water. We present an example of 2D and 3D resistivity imaging using a capacitive coupling resistivity (CCR) survey method in Svalbard permafrost. Although little work has been published on the mapping of the active layer of permafrost (i.e. the ground layer which thaws annually) with the CCR method, this case study shows its advantage as non-invasive compared to all other investigating methods.

The Adventdalen is a major valley near the city of Longyearbyen, made of flat terrace-like loess deposit. The site had been chosen by the University of Svalbard (UNIS) to monitor the long-term effect of climate change on permafrost and is referred as the UNISCALM site (Circumpolar Active Layer Monitoring). A CCR survey was acquired in July 2019 using an OhmMapper instrument from GEOMETRICS. The CALM site is ideal for towing a CCR streamer as the vegetation is relatively sparse, the ground is flat and the field is free from anthropologic noise. Two types of acquisition were tested: a 2D section (carried out by 10 passes along the same profile) and a 3D map survey (21 parallel lines separated by 5 m). The data was processed with RES2INV and RES3DINV softwares. The validation of resistivity models with other types of data is essential as resistivity is not only dependent on temperature, but varies with soil type, porosity and water content. For this purpose, the depth of the active layer was measured manually along the profiles: a steel rod is pushed into the ground penetrates unfrozen ground soil until it hits the hard frozen layer. The 2D section imagines 3 horizontal layers which concur with the soil stratigraphy observed in boreholes. The conductive layer (r < ~200 Wm) between 0 and 2 m depth corresponds to the active layer of permafrost. Below, the resistivity model images the ice rich zone (r up to ~2 kWm) down to approximately 5 meters depth, then the cryopeg zone (r ~ 10Wm) where freezing is prevented by the salinity of the soil. On the other hand, the 3D grid acquired at the CALM box images the conductive active layer and some elongated resistive ridges that correspond to ice wedges. These features are typical of the Adventdalen valley and were not visible during field work but can be seen on aerial photos and on shaded DEM imagery. These ice wedges are ice-filled cracks that are caused by volumetric expansion of water during freezing. They are often arranged in polygons. The deeper part of the resistivity grid reaches the top of the cryopeg brine and images a conductive patch confirming that the ice-rich zone is not uniform.

This case-study shows that even if the UNISCALM site has a uniform stratigraphy, its thermal structure is far from 1D. Subsequently, point-measurements of ground temperature and/or thaw depth may not be representative of the entire site. Instead, CCR surveying is a suitable tool for mapping the active layer structure. The CCR method is flexible (2D and/or 3D acquisition modes) and relatively easy to deploy. In addition, it overcomes the problem of electrical coupling between the ground and the electrodes, which is a challenge for the traditional galvanic method in cold and resistive environments. The CCR technique is also non-destructive unlike the traditional ERT method that uses stakes as electrodes, and unlike the commonly used method that drives a mechanical probe into the ground.

Voted Best Paper of EAGE- Near Surface Geoscience ’21, Bordeaux France

Cross-borehole electrical monitoring in groundwater remediation projects: understanding the flow path of remediation agents

Léa Lévy, Lund University, Lund, SWEDEN

Thue Bording, Aarhus University, Aarhus, DENMARK

Rasmus Thalund-Hansen, Technical University of Denmark, Lyngby, DENMARK

Kirsten Rügge, COWI, Lyngby, DENMARK

Maria Hag, Region Hovedstaden, Hillerød, DENMARK

Jørgen Fjeldsø Christensen, Region Syddanmark, Vejle, DENMARK

Soil and groundwater contamination has become a world-wide challenge. In 2017, a total of 1.3 million and 2.8 million contaminated sites were estimated in the US and Europe, respectively. Remediation by excavation is expensive and can lead to a significant carbon footprint. The development of cost-effective in situ remediation technologies, where contaminant sources and plumes are directly treated in the groundwater, is a high priority.

With in-situ remediation, adequate delivery of remediation agents in the whole target volume is challenging, whether it is directly the source area or a treatment zone intersecting the contaminant plume. Monitoring the spatial distribution of injected reagents is important for engineers and decision-makers, in order to determine the need of more careful injection in some areas. We use cross-borehole electrical resistivity tomography (XB-ERT) to visualize the progressive spreading of chemical reagents that are aimed at degrading chlorinated solvents in groundwater. Two sites are presented.

Kærgård Plantation is one of the largest polluted sites in Denmark, with over 300.000 tons of pharmaceutical waste dumped in sand dunes in the 1960s. The geology consists of sand and gravel layers of high permeability and the remediation targets the contamination hot spots themselves. Farum parking lot is an experimental site in the suburb of Copenhagen, where waste and spills from a packaging factory led to several groundwater contamination hotspots. The in-situ remediation there aims at creating a treatment zone intersecting the contaminant plume coming from the different hot spots. The injection takes place in a fine-grained sandy aquifer of intermediate permeability.

Resistivity imaging by XB-ERT allows visualizing the remediation cloud, composed of sulfate ions at Kærgård site and a complex mixture of solid iron and aqueous species at Farum site. Overall successful spreading is observed in Kærgård, whereas the treatment zone installation is much more problematic in Farum. Two injection rounds were attempted with different strategies in Farum. Both led to heterogeneous spreading, as well as significant upstream leakage and surface spills.

Based on time-lapse resistivity results, chemical monitoring, as well as solid iron analyses in core sediments, we suggest the creation of preferential pathways during injection at Farum, possibly related to the unintentional creation of fractures.

Voted Best Paper of EAGE- Near Surface Geoscience ’21, Bordeaux France


Ryan F. Adams, U.S. Geological Survey, Nashville, TN, USA
Amanda Whaling, U.S. Geological Survey, Fayetteville, AR, USA
Dan Kroes, U.S. Geological Survey, Baton Rouge, LA, USA

A suspended sediment concentration discrepancy was observed between endpoints of an 85-mile reach of the lower Mississippi River from the Old River Control Structure to Baton Rouge, LA. These differences account for between 10 and 20 percent of the suspended sediment load carried by the Mississippi River. It is unknown where or to what extend floodplain deposition may play in this discrepancy. To investigate potential sinks of the sediment along this reach, in 2020 and 2021, the U.S. Geological Survey (USGS) and the U.S. Army Corps of Engineers (USACE) conducted a series of airborne and waterborne geophysical surveys on Raccourci Old River (ROR), an oxbow lake created by an artificial cutoff of a meander bend beginning in 1845.
Approximately 120 km of airborne Frequency Domain Electromagnetics (AEM) was flown across Raccourci Island (RI, a large island located in the center of the horseshoe shaped oxbow lake) and along the thalweg of ROR. Sixty km of 1-D waterborne electrical resistivity were collected within ROR. These provided a detailed look at the layering of sediments beneath the entire survey area. Direct-push electrical and hydraulic logs in addition to sediment cores were collected on RI and adjacent to ROR to ground-truth the geophysical results. A grain size analysis was conducted on sub-samples from the drill cores to facilitate geophysical to lithologic transforms of the AEM and waterborne resistivity cross sections.
Sediment volumes accumulated in ROR were calculated using accretion rates derived from historical USACE and USGS bathymetric surveys collected in 1961, 1999, and 2013. Sediment volume estimates on RI were based on Cs-137 isotopic dating and were limited to a single time step, 1964-2016, from the peak isotopic deposition to the date of sampling.
Historic maps of the Mississippi River prior to the 1845 cutoff date show that ROR was a sandy bend of the Mississippi River—a sharp contrast to the fine-grained sediment that has filled it since the cutoff. This sand layer was the target for the final calculation of the total volume and mass of sediment accumulated in ROR since the cutoff event in 1845. The sediment accumulation rates from the historical bathymetric surveys were used to peel back layers from the AEM data until the 1845 surface was identified. A volume above this surface was calculated and the lithologic transforms linking the grain size information from the drill cores to the values from the AEM and waterborne resistivity data were used to estimate bulk densities of the sediment contained in that volume. These mass and volume estimates are critical to understanding the transport and deposition of sediment along this reach of the Mississippi River.


K. Chalikakis, Avignon University, UMR 1114 EMMAH (INRAE-AU)

N. Mazzilli, Avignon University, UMR 1114 EMMAH (INRAE-AU)

S.D. Carrière, Sorbonne University, UMR 7619 METIS (UPMC-CNRS-EPHE)

G. Massonnat, Total S.E., CSTJF

C. Danquigny, Avignon University, UMR 1114 EMMAH (INRAE-AU) and Total S.E., CSTJF

A. Legchenko, Grenoble Alps University, Institute of Research for Development, IGE

The unsaturated zone (UZ) of karst aquifers plays an important role in groundwater recharge processes. Comprehensive knowledge of UZ structure and hydrodynamic functioning is a key to better assess and manage groundwater resources in karst. In this extended abstract we present the results from a 2-year long Surface Nuclear Magnetic Resonance (SNMR) monitoring implemented in the Low-Noise Underground Laboratory (LSBB) site located within the Fontaine de Vaucluse karst hydrosystem in south-France. This exceptional 2-years SNMR monitoring demonstrated the efficiency of SNMR firstly to assess and validate the permanent presence of groundwater within karst hydrosystem UZ and secondly to monitor his temporal variations. Compared to boreholes, SNMR provides an integrated but accurate estimation of groundwater dynamics without disturbing the media.

Voted Best Paper of EAGE- Near Surface Geoscience ’21, Bordeaux France


Mohamed Ahmed, Department of Physical and Environmental Sciences, Texas A&M University—Corpus Christi, Corpus Christi, TX, USA

A geophysical test site (GTS) contains subsurface targets of known materials, orientations, and depths. These sites offer unique opportunities for geophysical research, training, and educational activities. GTSs provide platforms to investigate the penetration and resolution of different geophysical techniques for characterizing the shallow subsurface. GTS-based field exercises represent an interesting, motivating, rewarding, and enjoyable experience for both students and instructors. Recently, we have constructed a GTS at Corpus Christi, Texas (located at Texas A&M University—Corpus Christi). Our GTS (dimensions: 50 m × 50 m) contains several targets such as steel drums, plastic drums, plastic buckets, steel pipes, and well covers. The depth from ground surface to the top of the GTS targets ranges from 0.5 to 3 m. The GTS targets were selected to have magnetic, electric, and electromagnetic responses. These targets were chosen to simulate real-life situations. For example, the steel and plastic drums could represent chemical waste contamination, the steel pipes might represent part of a utility network (e.g., water, gas, electricity, telephone), and the well covers represent the heads of regular and/or abandoned wells. These targets were distributed along seven lines and grouped by material type. In this presentation, we provide a thorough description of the site location, subsurface geology, surface topography, and construction methodology, as well as the types, locations, orientations, and depths of the subsurface targets. We also provide lessons learned from the construction process that could serve as reference remarks for researchers interested in construction of new GTSs. Results of our preliminary magnetic and electromagnetic surveys are also provided. The research and education significance and implications of GTS surveys are also highlighted.


Md. Iftekhar Alam, University of Tennessee
William Doll, Collier Geophysics and University of Tennessee
Laura Bartel, Rutherford County Archaeological Society
Savannah Jobkar, University of Tennessee
Anna Cardwell University of Tennessee and
William Piwonka University of Tennessee

At SAGEEP 2021, we reported on our 2019 geophysical survey at the Old City Cemetery in Murfreesboro TN. The site has significant historical importance and had fallen into disrepair. In 2017, the Rutherford County Archaeological Society (RCAS) began a project to restore the site. Current work is a follow-on from the 2019 survey. Based on our previous results, we have acquired new data at five-selected locations, using GPR, magnetometer, and EM61. The 2021 data acquisition was intended to improve our understanding of optimal geophysical methods for cemetery surveys and focused on enhancing the detection of unmarked graves and a possible mass grave, and better definition of a large anomaly that extends through the center of the cemetery, for nearly the entire length of the site. GPR data were acquired in each of the five areas using a GSSI 4000 350MHz hyper stack antenna, generally with 1m line spacing. The GPR data provided significantly improved locations of anomalies (compared to 2019), and these are being correlated with magnetometer results. Importantly, many GPR anomalies are detected in the southern portion of the cemetery, where most of the burials are unmarked and a mass burial might be located. Magnetometer data were acquired in two of the five areas. The Foerster Ferex 4.034 magnetometer array has four vertical components magnetic gradiometers and was configured with 0.5m vertical separation between sensors and 0.25m lateral separation between the gradiometers. The lateral separation was reduced from our 2019 survey and focused on areas of particular interest in order to assess optimal parameters. Using 1m track spacing, we were able to acquire magnetometer data at 0.25m line spacing over both areas. Geonics EM61 data were acquired in three areas for comparison with the 2019 EM data acquired with the Geosensors R6 and to improve our understanding of EM tools for cemetery surveys. Results of the survey will be presented along with an assessment of the suitability of each tool for similar projects.


Abdullah Alhaj, Missouri University of Science and Technology, Rolla, MO, USA
David Rogers, Missouri University of Science and Technology, Rolla, MO, USA
Neil Anderson, Missouri University of Science and Technology, Rolla, MO, USA
Evgeniy Torgashov, Geotechnology, Inc., St. Louis, MO, USA

Geophysical data were acquired using Electrical resistivity tomography (ERT) and multichannel analysis of surface waves (MASW) to investigate a karst terrain site in Greene County, southwestern Missouri. In addition, conventional data were acquired to constrain geophysical data interpretations using aerial imagery, and borehole control. This case study is aiming to achieve two main objectives using an integrated geophysical and conventional approach by first identifying potentially low resistivity zones associated with changes in moisture content due to being in proximity to or underneath natural and man-made activities using the ERT and aerial images. Second, verifying the ERT data interpretations in terms of determining depth to top of rock, and moisture variations using MASW and Borehole control as ground truth.
The outcomes of ERT survey were displayed as 2D resistivity images of the subsurface to a depth of approximately 100 feet while the outcomes of MASW survey were displayed as 1D shear wave velocity profiles of the subsurface to a depth of approximately 50 feet to100 feet. Aerial images were acquired using Google Earth to identified areas of interest of water catchments or impediment on the ground surface and Borehole control were acquired to evaluate earth materials changes in terms of their moisture content, possible presence of voids and fractures as well as verifying depth to top of rock.
The preliminary findings of this research indicate that there is a potential decrease in resistivity of soils and rocks underneath and in proximity to or underneath natural (e.g., natural surface runoff, etc.) and manmade (e.g., roadways drainage ditches, etc.) activities compared to elsewhere along the study site. This could be attributed to the increase in moisture content at these anomalous zones due to the great infiltration and percolation of water downward into the subsurface through porosity of earth materials and drainage networks or conduits formed by the interconnected fractures, voids, and joints in karst bedrock. The identification of these anomalous zones is critical to understand the potential of formation and development of karst features near locations of water impediments for karst mitigation and remediation endeavors.


Muhammad Rizwan Asif, Department of Geoscience, Aarhus University, Aarhus, Denmark
Pradip Kumar Maurya, Department of Geoscience, Aarhus University, Aarhus, Denmark
Anders Vest Christiansen, Department of Geoscience, Aarhus University, Aarhus, Denmark
Jakob Juul Larsen, Department of Electrical and Computer Engineering, Aarhus University, Denmark
Esben Auken, Geological Survey of Denmark and Greenland, Copenhagen, Denmark

Transient Electromagnetic (TEM) methods are routinely used to obtain detailed understanding of the subsurface, which may be used for a variety of applications such as groundwater mapping and mineral exploration. Modern TEM surveys, employing driving or flying during data collection, result in large datasets that may contain thousands of line kilometers of data. Parts of these data will often be contaminated by interference from man-made conductors, e.g. fences, buried power lines, known as “couplings”. If such disturbed data are inverted, the geological interpretation will be severely degraded in most cases. Therefore, couplings must be identified and removed from the data before inversion. The process of removing couplings is a time-consuming and highly sophisticated manual task. Machine learning based methods have been suggested as obvious automation tools, and the general approach has so far been to use large datasets of manually processed TEM data in a supervised learning approach. The problem with this is that it may be biased to local geological conditions and/or biased toward the individuals who perform the manual assessment (for instance a conservative versus optimistic coupling removal approach).
Here, we present an alternative approach where we consider coupling recognition as an anomaly detection problem. We employ an algorithm based on a deep convolutional autoencoder as an expert system to distinguish between coupled and uncoupled data in an automated manner. Since autoencoders are unsupervised learning methods, we make use of synthetic TEM data of a huge ensemble of realistic subsurface models for its training. The autoencoder is trained to learn an encoded representation of synthetic TEM data in a reduced dimensional space. A reconstruction part is also trained that decodes the encoded representation aiming to output a dataset similar to the original input. Therefore, if uncoupled data are observed by the autoencoder, it will reconstruct the data from its encoded representation with low error. However, when dealing with couplings, the reconstruction error will be elevated, indicating a non-geologic anomaly. We define the size of anomaly as the relative error between the input data and the reconstructed output normalized by the data standard deviation.
We test our expert system on data from more than 20 different areas mapped by a towed TEM (tTEM) system and show that the proposed approach displays high quality data processing within a fraction of a second. The proposed method gives most value in areas with significant infrastructure where manual processing is more demanding. Due to the unbiased nature of the training dataset, our algorithm is generalizable to practically any tTEM data set acquired worldwide. Most importantly, the complex data processing task has been eliminated and no highly skilled operator is required, which would make these systems more accessible to non-expert end-users. Here, the deep learning network is designed for the tTEM system, but it can be modified to any TEM system, airborne or ground-based, but it does require the building of a database with forward responses from realistic models using that particular system setup.
In the presentation, we will present the deep learning algorithm and show several case applications where the proposed approach is compared against the traditional manual processing.


Elijah Ayolabi, Mountain Top University, Ibafo, Ogun, Nigeria
Rotimi Akinwale, Mountain Top University, Ibafo, Ogun, Nigeria
Moyo Oluwafemi, Mountain Top University, Ibafo, Ogun, Nigeria

A geophysical investigation involving 2D/3D Electrical Resistivity Tomography and Ground Penetrating Radar methods was carried out at the generator plant house, Ibafo, Southwestern, Nigeria with the aim of assessing diesel/oil waste polluted. The oil/diesel waste has been indiscriminately disposed over the years during servicing and remediation measures is difficult to propose because the depth of the polluted region is largely unknown. 2D electrical resistivity was carried out with Dipole-Dipole on eight survey lines and minimum electrode spacing of 0.3m was used to ensure high near surface resolution. Two 3D electrical resistivity cubes of 3.5 m by 6.5 m were occupied with 0.5 m minimum electrode spacing. Similarly, twenty Ground Penetrating Radar lines with 450MHz antenna were occupied along and perpendicular to the 2D Electrical Resistivity traverses. From the inverted 2D Electrical Resistivity Tomography, possible oil/diesel waste are characterized by relatively high electrical resistivity range of 51 to 1222 Ωm and delineated within the depth range of 0.2 to 2.5 m. Similarly, high resistivity anomalies (62 to 160 Ωm) are suggestive of the polluted regions on the inverted 3D resistivity tomograms. Meanwhile, on the Ground Penetrating Radar, the suspected oil and diesel polluted regions are characterized by amplitude suppression. The mapped anomalous regions correlate with regions that are visibly polluted by oil/diesel waste on the site. These mapped anomalies have been used to generate a depth map showing the spatial variation of suspected polluted region across the investigated area. The result of this study has shown that the diesel/oil pollution is limited to the near surface within the depth of 0.2 to 2.5m. Remediation and mitigation measures can therefore be guided by the generated depth map of oil/diesel pollution.

Resolving hydrogeological parameters through joint inversion of seismic and electric data considering surface conductivity

Matthias Steiner, TU Wien, Vienna, AUSTRIA

Timea Katona, TU Wien, Vienna, AUSTRIA

Nathalie Roser, TU Wien, Vienna, AUSTRIA

Günter Böschl, TU Wien, Vienna, AUSTRIA

Adrián Flores Orozco, TU Wien, Vienna, AUSTRIA

Geophysical methods have proven to overcome the spatial limitations of direct investigations by providing spatio-temporal information about subsurface properties with an adequate resolution in a non-invasive manner. However, the resolved models remain qualitative unless subsequently transformed to the quantitative estimates of the parameters of interest based on a petrophysical model. Petrophysical joint inversion (PJI) approaches permit an improved quantitative estimation of hydrogeological parameters by simultaneously inverting complementary geophysical datasets, e.g., seismic and electric data, related through a common petrophysical parameter. Subsurface models resolved for data collected in fine-grained environments might still be biased if the petrophysical model underlying the PJI framework does not consider the conduction of electric current along the grain-fluid interface. In this study, we present a PJI framework that implicitly takes into account the surface conductivity based DC and instantaneous resistivity data, i.e., the electrical response observed at a low and a high frequency. We apply this PJI approach to data collected in the Hydrological Open Air Laboratory (HOAL; Petzenkirchen, Austria) to solve for hydrogeological parameters relevant for the understanding of surface-groundwater interactions. We discuss the resolved subsurface models with respect to models obtained through a PJI approach neglecting the surface conductivity, demonstrate the good agreement with available direct information and provide an interpretation of the subsurface conditions.

Voted Best Paper of EAGE- Near Surface Geoscience ’21, Bordeaux France


Elijah Ayolabi, Mountain Top University, Ibafo, Ogun, Nigeria
Rotimi Akinwale, Mountain Top University, Ibafo, Ogun, Nigeria
David Abiodun-Herbert, Mountain Top University, Ibafo, Ogun, Nigeria

Buried utilities of different shapes and sizes are characteristics of built-up areas and carry different materials such as water, waste water, electrical signal, communication signals etc. The careful identification of these utilities in order not to constitute danger to life and property when further development is needed is pertinent. To this end, 2D Electrical Resistivity Tomography and 2D/3D Ground Penetrating Radar has been used in mapping buried plastic pipes at three locations within Mountain Top University, Southwestern Nigeria in order to provide information about the location, diameter and depth of buried plastic pipes. 2D and 3D GPR survey were conducted using monostatic antennas with frequencies of 450 and 750 MHz. One 6 m by 6 m and two 3 m by 3 m 3D GPR were carried out with lines pacing interval of 1 m and 0.5 m respectively. 2D Electrical Resistivity imaging was carried out with dipole-dipole array and electrode spacing of 0.1 m in order to ensure high subsurface resolution. On the 2D ERT, the pipeline is represented by relatively high electrical resistivity anomalies (700 to 100k Ωm). The high resistivity contrast is suggestive of air-filled plastic pipes. The pipes are also represented by hyperbolic signatures on the GPR. Interestingly, linear anomalies are observed on some 3D GPR time slices which are associated to the buried pipes thereby providing a robust visualization of the buried utility. Integration of the 2D ERT inverted sections and radargrams shows that depth to the top of the buried pipes ranges from 0.2 to 0.31 m while diameter of the buried pipes ranges from 0.2 to 0.22m. Ground truthing of the buried pipes shows that the diameter of the pipes are 8 inches (0.2032 m). The result of this study further corroborates that integration of 2D ERT and GPR surveys allows for a better identification of buried utilities.


Stuart Bancroft, Jacobs, Pensacola, FL, USA
Chris Eykamp, Jacobs, Portland, OR, USA
Jennifer Weller, Jacobs, Boulder, CO, USA
John Jackson, EM CX USACE, Denver, CO, USA

Jacobs has developed an Automated Quality Assurance (QA) Tool for the U.S. Army Corps of Engineers (USACE) to automate and streamline manual tasks performed by USACE QA geophysicists. The purpose of the Automated QA Tool is to create a user-friendly product to streamline, standardize, and accelerate QA activities specific to munitions response sites (MRSs) using advanced geophysical classification (AGC). The individual modules of the Automated QA Tool streamline the quality control (QC) and QA information inputs and outputs to inform data usability assessments and facilitate the verification of critical remedial response assumptions. The Automated QA Tool components are accessible from a custom menu loaded into the Geosoft interface or through command line prompts. The Automated QA Tool consists of four separate modules: Validation Target Selection, Customizable Target of Interest (TOI) Search, Conceptual Site Model (CSM) Assumption Test, and Site Noise Assessment.
The Validation Target Selection module provides a physics-based method to inform how validation targets are selected. Targets that fit into other parts of the AGC selection process, such as TOIs, verification targets, and analyst calibration digs, are identified with simple user settings and eliminated from consideration. The remaining eligible digs then are prioritized using physics-based characteristics, including log-linear fits to polarizability curves and intrinsic physical characteristics (size/decay). The results are used to provide a basis for selecting targets as validation digs.
The TOI Search module automates the search process for TOIs using various assumptions about TOI type, TOI orientation, and TOI depth throughout the source database. The user interface allows numerous input combinations to filter and extract a set of source targets that meet the user criteria.
The CSM Assumption Test module facilitates the search for TOIs throughout the project database to affirm the CSM assumptions. The user interface records the expected munitions types and vertical boundary for the project. The items recovered in the project database are compared against the items in the site-specific library and any munitions item or munitions and explosives of concern (MEC) item listed in the project database that does not match an item in the site-specific library is flagged. The output report documents the munitions types recovered and the vertical distribution of all munitions items and MEC items recovered, then it lists any unexpected results (munitions or MEC items recovered that were not part of the CSM). This module also identifies munitions items from the complete Department of Defense (DoD) TOI library with high matches not included in the site-specific library
The Site Noise Assessment module provides mapping and analysis capabilities to assess the geophysical noise response throughout a site and identify areas of concern. This tool offers noise assessment functions based on data chip sampling and synthetic seeding. The Model Coherence Threshold function in UX-Analyze has been incorporated to provide spatial distribution of root-mean-square (RMS) noise and model coherence threshold of data chips placed across the site that are below a user input background threshold. Synthetically inserted TOIs can be placed within the data chip to produce hypothetical model coherence threshold values based on the selected TOI, depth, and orientation. The module output includes a Geosoft map with the RMS noise values presented as a grid layer and symbols or polygons flagging areas above the user threshold as an additional map layer. Synthetic TOI model coherence results are grouped by user-defined depth intervals and statistically analyzed to estimate classification depth limitations within the project area.


Paul D. Bauman, BGC Engineering Inc., Calgary, Canada
Colin M. Miazga, BGC Engineering Inc., Calgary, Canada
Chris L. Slater, BGC Engineering Inc., Calgary, Canada
Eric H. Johnson, BGC Engineering Inc., Calgary, Canada

Putumayo is a department (i.e., state) of southwestern Colombia, bordering on both Peru and Ecuador. In the broad narrative of pre-Columbian (i.e., before 1492) archaeology, little is known about the peoples who inhabited Putumayo. The Inca empire reached the southwestern corner of Colombia, and hundreds of fourth millennium monolithic massive stone sculptures exist in Huila Department to the north; however, Colombia’s 57-year civil war and the control of Putumayo by the FARC guerilla group and narco-traffickers made the Department inaccessible to archaeologists. The 2016 peace accord between the FARC and the Colombian government of President Juan Manuel Santos created a window where archaeological investigations were deemed possible.
Following the 2016 peace accords, a resource extraction company carried out a conventional archaeological exploration program over an approximate 5 km2 area in preparation for the construction of an 8-kilometer lease access road. The exploration area is in the Amazonian Piedmont, in the headwaters of the Amazon River. Of the 4000 test pits that were excavated, artifacts were found in only 80 pits. These artifacts included 1100 pottery shards and 34 lithic (stone) artifacts. Artifacts were dated from 2500 years before the present (BP) to 500 years BP. Given that many of the 4000 test pits were excavated to infill areas of known finds, the percentage of test pits discovering areas of human activity was significantly less than 2%.
In September 2018, a two-week archaeo-geophysical program was carried out in this same exploration area with the goal of assessing a relatively rapid, non-intrusive geophysical and UAV mapping approach. The objective was not to replace conventional test pitting, but instead to better focus test pitting, to “sterilize”areas of little or no cultural activity, to provide boundaries of areas of cultural activity, to reduce the number of test pits, and to estimate the maximum depth to which anthropogenic finds may be expected. Surveys included UAV photogrammetry, UAV multispectral imagery, magnetic gradiometer mapping, resistance meter mapping, terrain conductivity mapping, and ground penetrating radar mapping.
Despite the area being very large and difficult to access, all work was required to be completed within two weeks due to significant security concerns. The area was known to have landmines, booby traps, and was still largely controlled by narco-traffickers and renegade elements of the FARC. Coca cultivation was still ubiquitous; kidnapping and the murder of community leaders remained common in the area. Although we received extraordinary support from the Colombian military, we were unable to mobilize with aircraft due to concerns of being shot at. Significant ongoing flooding in the Upper Amazon Basin required our support crew to construct bridges as we moved.
Photogrammetry and multispectral imagery were extremely useful for creating base maps, but failed to identify any ancient landscape modifications such as terracing or irrigation ditches. A total of 900 m2 of high resolution geophysics was surveyed. About 14.5% of this survey area showed what were interpreted as high priority geophysical anomalies. Magnetic gradiometry provided the most useful identification of high confidence anomalies. Nevertheless, all geophysical methods provided useful information. In one small area, a very high resolution 3D GPR survey identified approximately 200 diffractions that are likely related to pottery shards. Possible anthropogenic sources of the various geophysical anomalies include pits, ditches, earthen floors, burning, fire baked earth, fire baked bricks, ferro- and non-ferromagnetic metal, and chemical reactions from bacterial activity associated with anthropogenic activity.
The archaeo-geophysical program described here was executed without any access to the results of the previous intrusive archaeological investigations; in other words, this program was performed as a blind test. The initial results are encouraging, though intrusive investigations are required to fully evaluate the geophysical program. As far as we know, no such multi-method geophysical and UAV prospection program had been previously carried out in Colombia, and certainly not in Putumayo. Other organizations in Putumayo are considering using this same non-intrusive approach to precede conventional and destructive archaeological investigations.


Amanda Beattie, Texas A&M University- Corpus Christi, Corpus Christi, TX, USA
Mohamed Ahmed, Texas A&M University- Corpus Christi, Corpus Christi, TX, USA
Michael Haley, Texas A&M University- Corpus Christi, Corpus Christi, TX, USA
James Rizzo, Texas A&M University- Corpus Christi, Corpus Christi, TX, USA

The Coastal Bend of Texas has been subject to the negative impacts of land subsidence due to the excessive withdrawal of groundwater, oil, natural gas resources, presence of growth faults, and compaction of Quaternary alluvium. As subsidence poses problems to all coastal communities, infrastructure, and wetland habitats, it is important to determine the rates, locations, and factors controlling this phenomenon. The rate of land subsidence can be quantified by collecting time-variable gravity measurements in certain locations. Temporal and spatial variations in Earth’s gravitational field can be affected by the variability in oceanic and solid Earth tides, and because of this, tides must be accounted for and removed from the observed gravity data before converting them into elevations. Tide and drift corrected temporal variations in Earth’s gravity field can then be converted to height variations, and those heights can be used to determine land subsidence rates. Over the course of this study, temporal bi-monthly gravity measurements have been collected at six locations along the Coastal Bend of Texas, including Packery Channel, Bob Hall Pier, Port Aransas, Rockport Harbor, Nueces Bay, and the Lexington. Because this region is underlain by sand, mud, and clay deposits, it is susceptible to sediment compaction, resulting in land subsidence. The gravity survey was conducted using a LaCoste & Romberg G-976 zero-length spring gravimeter. Metadata such as time, elevation, and coordinates are also recorded along with the gravimeter readings. For all stations, changes in gravity measurements were in the range of -0.107 to 0.106 mGal, with Port Aransas having the lowest trend, and Bob Hall having the highest trend. As expected, the changes in height were greatest at the stations with steepest gravity trends. More measurements are currently being collected to refine these land subsidence rates. Additionally, InSAR data have been acquired from 2015 to present and are currently being processed in order to further constrain these results by mapping the spatiotemporal variability in land subsidence within this area. Preliminary results of this study indicate that land subsidence in the Texas Coastal Bend is mainly a vertical motion that varies with both time and location. Based on the locations of the gravity stations, this subsidence is likely due to sediment compaction and growth faults instead of fluid extraction.


Ronald S Bell, Drone Geoscience. LLC

In 2020, a collaboration between Geonics, Ltd. and SPH Engineering, Inc., the developer of the Universal ground Control System (UgCS), resulted in the adaption of the EM61, broadly applied deep looking metal detector, for use with a small UAS. In 2021, the Drone Geoscience (DG) team was given the privilege to test newly developed EM61 Lite, a standard EM61 modified for use with the drone, which in our case is a Matrice 600 Pro hexacopter made by DJI.
Due to the payload limitations of our drone, the first attempt at lifting the EM61 Lite was unsuccessful. After some creative re-engineering of system suspension for the coils and testing different approaches to providing power to the EM61 electronics, the DG Team conducted field trials at ultra-low altitude to successfully map a buried oil and gas pipeline. In this presentation, I will discuss the results of our field trials and share the current thinking regarding achieving a successful implementation of a drone enabled EM61.


Simon Bouteille, IRIS Instruments France
Jacques Marteau, I2I Lyon
Benoit Le Moigne, BRGM
Nicole Hueber, BRGM
Kevin Samyn, BRGM
Catherine Truffert, IRIS Instruments France

La Reunion Island, located East of Madagascar, is composed of three shield volcanoes among with la Fournaise which is still active. The volcanic cirques, subjected to large-scale rock-falls (>10,000 m3), regularly, impose crisis management on public authorities. Rock-falls, which sometimes cause a retreat of several meters at the top of the walls, can occur every 5 to 10 years on average. However, it seems that the effects of climate change (cyclonic events, drought, fire…) increase this occurrence.
On the top of the wall (or Rempart), the decompression cracks are concentrated on a strip often equivalent to 10% of the height of the cliff that can be higher than 1,000 m. These cracks, sometimes more than a meter wide, delimit the rock scales likely to be crumbled. The origin of these cracks, which are almost vertical on the surface, is linked to the natural decompression of the massif by the vacuum. The geometry of these cracks at depth is not well known, but it is likely that they acquire a slightly concave shape, bringing them closer to the wall and cutting out large scales at the crest of the Rempart. The volume of rock falling highly depends on the depth of these cracks.
Our experiment is focused on the Maïdo Rempart overlooking the western part of the Cirque of Mafate where the formations of the ancient volcanic outcrop in 1,000 m high scarps. We have installed a Muons telescope at “l’Ilet de Roche-Plate”, a small village located at the foot of active scree cones at the foot of the Maïdo Rempart. This innovative experiment follows a fire that occurred on the top of the Rempart at the end of 2020, which led to an increase in falling blocks and a potential acceleration in the opening of cracks.
A diagnosis and a monitoring were requested by the local authorities to the national geological survey, the BRGM. One of the issues is to better delimit the volume of “fractured” rocks and if possible, to identify the depth of the decompression cracks that delimit the scales likely to fall. Given the gigantic morphology of the geological feature, the solutions available to geoscientists are few.
Muography was chosen because it allows to access the density variation in time and space, in a passive way, by collecting in the telescope the muons which crossed the rock. A telescope based on scintillators technology has been installed for up to 6 months at the footwall of the Maïdo Rempart. The rainy season was chosen as the acquisition window to be able to follow the density variations that occur in the massif during rainy events. It is composed of 3 parallel ~1m2 active detection planes recording the positions and the precise time of the particle’s hits. The detector readout has been developed on the early concept of connected “smart sensors”. It allows an optimized selection of the particles hits to perform their tracking. A post-processing analysis will translate the recorded tracking properties into a detailed image of the target within the acceptance of the detector, in our case the top part of the Maïdo Rempart.


Roy Bowling, Collier Geophysics, LLC
Phil Sirles, Collier Geophysics, LLC
Todd Schittenhart, Yeh and Associates, Inc.
Sam Holder, Yeh and Associates, Inc.
Leyla Safari, Yeh and Associates, Inc.

Scenic Loop road within the South Unit of Theodore Roosevelt National Park, ND circumnavigates thousands of acres of dramatic badlands, providing sweeping vistas for park visitors. The southern extent of Scenic Loop road winds near the top of steep hoodoos and buttes. These geologic features are comprised of poorly lithified claystones of the Bullion Creek and Sentinel Butte formations, and are highly erodible and residual clay soils are dispersive. As a result, roadway damage due to erosion and landsides is common along this portion of Scenic Loop road. Cold-patching and limited remediation efforts have been completed by the National Park Service during the lifetime of the roadway. During the winters of 2019-2020, a portion of the roadway was destroyed by a landslide making the Loop road impassible. This slide is one of many areas along the southern extent of Scenic Loop road where mass movement has been observed and has caused/is causing roadway damage, and prompted the need for engineering mitigation. Seismic velocity mapping, via refraction tomography and multichannel analysis of surface waves, was used to infer the depth of landslide shear-planes at multiple locations along the southern portion of Scenic Loop road, and aid in engineering design of landslide mitigation, drainage needs and roadway reconstruction. Given the low density and poor induration of the bedrock at this site, strong velocity contrasts between displaced and un-displaced landslide debris are not evident in the geophysical results. However, subtle velocity variations in the shallow near-surface are observed in both the p-wave refraction and s-wave surface wave analysis results. Strong correlations emerged when comparing the depths of these subtle velocity anomalies to shear-plane depths measured at sites with borehole inclinometers. These subtle correlations persisted at multiple other landslide locations and allowed for the interpretation of landslide shear-planes at sites with limited in-situ measurements of slip. The results of the geophysical investigation have been used to inform slope stability geotechnical modeling and the design of roadway repair and slope stabilization in order to re-open the Scenic Loop road.


Daniel Campos Halas, Geophysics GPR International, Montreal, QC Canada
Olivier Létourneau, Geophysics GPR International, Montreal, QC Canada
Janny Desgagné, Geophysics GPR International, Montreal, QC Canada

Standard MASW (Multi-channel Analysis of Surface Waves) processing workflows imply various interpreter operations such as trace editing, hand picking of the dispersion curve, and higher mode identification. These operations can be, at a certain level, subjective, and require a high level of expertise from the interpreter, especially for complicated data sets. In addition, standard unconstrained inversion routines are non-unique and provide infinite solutions. Finally, a proper interpretation of complex data sets is time consuming, even for a highly skilled geophysicist. The authors have developed a machine learning algorithm to automate the interpretation of standard MASW data sets used for the calculation of the “Vs30” parameter and the prediction of the sounding’s shear wave velocity profile.

The Vs30 results is obtained using Resnet34, a 34-layer convolutional neural network that can be utilized as a state-of-the-art image classification or regression model. In the case of the MASwAi network, Resnet34 is utilized as a regression against the combined stacked shots dispersion image (“phase shift” for MASW, and “cross-correlation” for SPAC) and other geophysical constraints such as the rock depth. The MASwAi neural network was trained using a large number of field data and augmented by complimentary synthetic data. The MASwAi network results accuracy was assessed with validation datasets, highly skilled geophysicists cross-interpretations, and direct in-situ Vs measurements conducted on test sites using the downhole seismic method and sCPT soundings. Vs30 results show that the network predictions fall within the accuracy rate of the MASW technique for shear wave velocity measurements.

Increasing demand for geophysical surveys, a competitive market, and the limited availability of skilled geophysicists are increasing challenges that need to be met to maintain a high level of quality necessary to provide accurate seismic site classifications required by national building codes. The development of a reliable machine learning tool to assist geophysicists in MASW interpretation will help to cope with these obstacles in such contexts.


J.R. Candlish, Weston Solutions, West Chester, PA, USA
Ryan Steigerwalt, Weston Solutions, West Chester, PA, USA

A remedial action is being conducted to cleanup munitions and explosives of concern over more than 1,000 acres of Croft State Park, South Carolina. Located near Spartanburg, former Camp Croft served as a World War II basic training center and is now a Formerly Used Defense Site (FUDS) ranked on the FUDS priority list. The rolling, densely forested terrain within Croft State Park is a coveted natural resource to the local community and a popular destination for recreational enthusiasts. In the not-so-distant past, analog geophysical methods would be the preferential approach for remediation to limit disruption to park operations and maintain a natural landscape. Park use and natural balance will always be contributing factors to remedial design at FUDS of this type. However, effectively executing the remedy is exponentially more complicated with increased emphasis of deploying advanced geophysical classification (AGC) technologies no matter the site conditions. Planning for Camp Croft required flexibility and the ability to adapt to challenges while incorporating the needs of land managers and contractually driven preferences for technology selection. Currently, dynamic AGC is being performed with six APEX units at Camp Croft. Extraordinary site preparation and startup activities were required leading into initial vegetation reduction around closely spaced mature growth trees. Because tree removal in the park is to be avoided, obstruction density was found to negate any effectiveness of robotic total station (RTS) positioning and impaired planned production rates of APEX. Because of plan flexibility and project delivery team willingness, work quickly pivoted to simultaneous localization and mapping (SLAM) demonstrations and integration with APEX. Daily operations have now completely transitioned from RTS to KAARTA Stencil for all APEX units. This presentation provides unique aspects of the project and an update on delivering dynamic AGC and SLAM at a challenging site on an unprecedented scale.


Anders vest Christiansen, Aarhus University, Denmark
Pradip Maurya, Aarhus University, Denmark
Jesper Pedersen, Aarhus University, Denmark
Frederik Ersted Christensen, Aarhus University, Denmark

In the 1950s and 1960s, a 12 hectares large area at Himmark Beach, the region of southern Denmark, was used to deposit around 53,000 cubic-meters chemical waste and 700,000-liter wastewater from industry. Himmark Beach is one of the ten most costly sites to remediate in Denmark. The contaminations are spreading through the subsurface structures in the coastal zones and migrating into the sea, posing a threat to wildlife, humans, and the groundwater. In order to fully characterize the site and make detailed remediation measures, it was decided to use initiate a detailed transient electromagnetic (TEM) mapping campaign to get a three-dimensional understanding of the geological structures, which in turn controls the contaminant pathways. Specifically, there is a shallow sand layer with high porosity right beneath the seafloor, which is underlain by an impermeable till layer. The main aim of the mapping was to detail the interface between sand and till to identify possible pathways for the pollution and pinpoint pockets in the till surface with increased contamination concentrations. The survey contained both ground-based towed measurements (tTEM) on land and waterborne measurements (FloaTEM) on the sea. The objective of the FloaTEM survey was to map the till layer starting from the shoreline and extending to the sea, whereas the onshore tTEM campaign aimed at identifying a local site with a thick impermeable clay sequence that could be used as a temporary host for the planned excavation of polluted sand without posing a threat to local groundwater resources. The tTEM survey was carried out using a single turn 3×3 m2 transmitter-coil and consists of low moment (LM) and high moment (HM) pulses of 3 A and 30 A respectively. For surveying on seawater, we used a 4-turn 4×4 m2 transmitter-coil with a 25 A HM pulse. The latest time gate was 1 ms for tTEM f and 3 ms for FloaTEM. Inversion results of the tTEM and FloaTEM surveys show a distinct three-layer sequence with increasing resistivity with depth. The layered sequence has been compared with the 154 boreholes drilled into the seabed. The comparison confirms that the layers correspond with the seawater column, saltwater saturated sand and in 1-9 m depth we find the till layer. In one part of the area, where there is limited borehole coverage, there is a distinct deepening in the till layer, which is filled with saturated sand, gravel and peat. This area is a potential highway for contamination transportation. Based on the results from the geophysics, the authorities have redefined their strategy for excavating the contaminated sand, gravel and peat. The geophysics was used to pin-point locations for new monitoring wells, and for designing the outline of the sheet pile wall, which will be constructed into the seabed around the contaminated zone during excavation.


Ned Clayton, NMR Services, Tucson, AZ, USA
Ryan Gee, NMR Services, Helsinki, Finland
Jim LoCoco, Mount Sopris Instruments, Denver, CO, USA
Martin Helmke, West Chester University, West Chester, PA, USA

Saprolite, residual soil derived from in-situ weathering of bedrock, is present across much of the Piedmont Plateau in the eastern U.S., as well as many other regions in humid temperate and tropical climates where crystalline bedrock is exposed at surface. Saprolites are an important part of the Critical Zone, connecting the atmosphere/climate and surface hydrology to groundwater hydrology, and providing water recharge and storage to fractured bedrock aquifer systems. High resolution and quantitative characterization of in-situ geologic/hydrogeologic properties of saprolites and the saprolite-bedrock transition zone, and their heterogeneity, is often limited or non-existent, resulting in a poor understanding of the hydrogeologic processes that link the vadose and saturated zones in saprolites to underlying bedrock aquifers.
A series of research boreholes on the West Chester University Campus in eastern Pennsylvania, U.S. intersect the surficial soils, saprolite and the saprolite transition zone before entering the underlying bedrock aquifer system. A suite of borehole geophysical logging measurements were conducted in one of the boreholes MW-8 in September 2021 as part of a Field Hydrogeophysics Workshop, including Borehole Magnetic Resonance (BMR) logging. BMR tools measure lithology-independent volumetric water content (total porosity in the saturated zone), as well as the pore scale movability of the water (pore size distribution in saturated conditions) using well-proven NMR measurement and processing methods. In turn, with additional information about the lithology, these downhole BMR measurements are used to derive moveable water volume (specific yield in the saturated zone), and capillary and clay bound water (specific retention in the saturated zone), as well as to estimate hydraulic conductivity that can be calibrated with well or core hydraulic tests. The BMR logs from MW-8, coupled with the EM conductivity log, provide high resolution delineation of hydraulically conductive portions of the saprolite transition zone, including a productive water zone that correlates with a screened interval in the offset MW-7 at similar depths. Above this zone the total porosity is high (25-40%), along with an increase in electrical conductivity, corresponding with the silt-rich, highly chemically weathered middle and upper saprolite zones. Upon closer review of gamma, induction and other logs, there appears to be several additional thinner and lower porosity hydraulically conductive zones below the main production interval that could potentially be hydraulically isolated, overall porosity decreasing with depth to the crystalline bedrock. Observations while drilling encountered several “false bedrock” zones before penetrating the true bedrock, also suggesting potential isolation from the main zone of interest.
This paper examines how the BMR is used to provide high resolution hydrogeologic characterization of saprolite, revealing that the saprolite transition zone at the West Chester University research site is more complex than first anticipated, with multiple hydraulically conductive zones present. The BMR, EM conductivity and other log results from MW-8 are compared with measurements from offset boreholes to determine if additional information and insights about the saprolite profile and lateral continuity can be gleaned.


Thomas I. Coleman, Carlos H. Maldaner, and Celia S. Kennedy, Silixa LLC, Missoula, MT, USA

Dams support two fundamental staples of socio-economic development, energy and water, in adequate supply and quality. They affect the natural magnitude and timing of streamflow, creating a system of managed waterways. With the benefits of managed river systems comes the responsibility of continuous monitoring to maintain the integrity of management structures. There are 11,500 large dams in North America (ICOLD 2017), 3,000 of which provide 60% of the electricity supplied to Canada and the northwestern United States. This talk will describe the challenges associated with risk analysis, andanalysis and explain the application of recent technological advances in fiber optic distributed sensing for continuous monitoring of structure integrity and advance warning systems for existing dam types.

Dam safety in North America has been managed using two methods of risk analysis. The Failure Modes and Effects Analysis (FMEA) and the Potential Failure Modes Analysis (PFMA), which follows the US Federal Energy & Regulatory Commission (FERC) Engineering Guidelines for the Evaluation of Hydropower Projects. Both methods use event trees, dam failure modes, calculated probabilities and a risk matrix, but struggle with the acquisition of adequate field data. In February 2017, the Oroville Incident sparked controversy associated with inconsistencies in risk analysis methods. Indeed, aIndeed, a new approach that integrates monitoring, data-tracking, and operational factors is needed, rather than relying on engineer judgment to estimate probabilities and consequences (ICOLD 2019). Advances made in fiber optic distributed temperature, acoustic and strain sensing (DTS, DAS, and DSS) have allowed autonomous and real-time data acquisition at high spatial and temporal resolutions. Unlike traditional sensing that relies on individual sensor measurements at predetermined points (eg. extensometer, pressure and temperature sensors, or geophones), distributed sensing utilizes optical fiber as the sensing element without any additional transducers along its path. Spatio-temporal coverage is continuous, with a single sensor system acting as tens of thousands of independent sensors, each with sensitivity similar to or better than point sensors. Because the optical fiber is the sensor, it includes no electronic or moving parts and requires no maintenance.

Embankments comprise 75% of the world’s dams. Concrete gravity dams rely on their material weight, while the thinner geotechnical design of arch dams utilizes their material strength. Buttress dams are hollow and rely on vertical projections spaced out along the downstream face. Variations of these fundamental designs, such as those exhibited in dykes, levees and weirs will not be discussed in the interest of conciseness. All dams can be monitored using similar cable installations, customized to their reservoir capacity and environment. Temporally-continuousTemporally continuous measurements can be collected over tens of kilometers of embankments and structures using robust, direct-bury fiber optic cable installed horizontally and vertically during construction, or via retrofitting processes using trenches and boreholes along concrete dam structures and spillways. Application of a single cable containing multiple optical fibers can provide the following data outputs. DTS interrogators provide independent temperature measurements every 0.25 m along the cable, with temperature resolution as fine as 0.01 °C. Ambient temperature fluctuations throughout the year cause reservoir temperature to change. DTS uses this signal as a tracer to estimate seepage flow through an earthen embankment and identify locations of potential failure. DSS absolute strain data provides a direct measurement of any deformation or settlement occurring along a dam. Strain measurements can be obtained every 0.10 m along the fiber optic cable with a strain resolution of 2 με. An increase in strain values at a specific location over time can indicate vertical or horizontal movement of a structure. DAS detects natural or induced microseismic events to evaluate seismic risk, or to be used for time-lapse seismic imaging based on passive and active methods. The results from these imaging methods are correlated to potential changes in physical properties of the structure material such as density, saturation related to water level and erosion, and shear strength which could also indicate locations with potential failure.

In summary, a single, cost-efficient fiber optical cable can be used to monitor large structures in detail, yielding complementary datasets for analysis in parallel to provide an independent means of evaluation using multiple sources of evidence to identify and locate potential failures. Achievability of large spatial coverage and autonomous operation that requires low maintenance and power makes this technology a cost-effective monitoring solution.


Yigrem A Dingo, Husky Engineering PLC*

Landslide represents one of the main constraints for the development of road infrastructures in many parts of Ethiopia. While rugged topography, intra-sedimentary clay horizons, thick talus deposits at the foot of hills and widespread tectonic fissures and faults account for the prominent inducing factors, heavy monsoon rainfall is the most common triggering factor for the majority of road failures caused by landslides. We present the results of geoelectrical investigation of a landslide that caused a severe damage to the Landslide Study at Felege-Selam-Amaya-Chida Road Project of Amaya segment of the recently built highway that transects the high relief and mountainous region of Southwestern Ethiopia. Vertical Electrical Sounding (VES) and Electrical Resistivity Tomography (ERT) measurements were taken at twelve sites, systematically distributed to cover the affected area. The resulting geoelectric sections revealed that the shallow subsurface is composed of four distinct geo-electric layers. The corresponding major lithological successions, from top to bottom, are unconsolidated conglomeratic layer, moist silty clay, moderately weathered and fractured basaltic rock and a possibly saturated portion of the basalt. Major discontinuities, indications of structural weak zones, have also been inferred based on abrupt vertical shift in geo-electrical layer boundaries between neighboring soundings and profiles. Based on the main findings, the road failure in the study area appears to be caused by a downslope movement of the subgrade composed of top unconsolidated sediment soaked with water from heavy rain and series of seepages. The location of the probable slip plane could be the inclined interface between the conductive clayey soil and the underlying resistive weathered basalt.

*Husky Engineering PLC, POBOX-046, Gelan Condominium, Akaki Kality Sub city Addis Ababa, Ethiopia


Charles L. Dorchester, Colorado School of Mines, Golden, CO, USA
Kamini Singha, Colorado School of Mines, Golden, CO, USA
Frederick D. Day-Lewis, Pacific Northwest National Laboratory, Richland, WA, USA

Dual-domain mass transfer (DDMT) is a conceptual model that interprets aquifer systems as being defined by a mobile porosity, an immobile porosity, and the mass-transfer rate coefficient that describes how easily solutes can move between these two domains. The mass-transfer rate coefficient describes the retention and release of solutes from immobile porosity, which has been shown to produce late-time concentration rebound and the slow release of solutes, otherwise known as anomalous tailing. Although DDMT techniques can describe contaminant storage and release more accurately than the advection-dispersion equation, they typically involve adjusting these model parameters until the model predictions fit observed data. To help constrain estimation of these parameters, electrical resistivity (ER) has been used in previous work; fluid electrical conductivity and bulk electrical conductivity show a hysteretic relationship in response to DDMT behavior because fluid sampling preferentially draws from the mobile domain, whereas ER methods are sensitive to the bulk composition. Here, we investigate the relationship between the mobile porosity, immobile porosity, and mass-transfer rate coefficient in a pore-scale numerical model by simulating solute concentration—converted to fluid electrical conductivity—and the subsequent bulk electrical conductivity. In particular, we study the application of ER to (1) constrain parameterization of a pore-scale DDMT model, and (2) assess spatial variability of effective parameters.
The pore-scale numerical model, developed in COMSOL Multiphysics, simulates fluid flow, transport of an electrically conductive solute, and direct-current electrical conduction in 2-D. We match an analytic solution of the 1-D DDMT equation to our simulated data using a nonlinear least-squares regression technique. We find that: (1) mobile domains and immobile domains are present, regardless of the system being explicitly defined with one versus two domains; and (2) the mobile porosity, immobile porosity, and mass-transfer rate can be estimated by analyzing the hysteretic relationship between fluid and bulk electrical conductivity, and perhaps meaningfully scaled to field systems. Overall, our findings support the hypothesis that hysteresis in ER data is a function of fast versus slow paths, even in the absence of pre-defined immobile porosity, so an open question is what “immobile” means in a macroscopic system with no explicit immobile pore space.


Tori S. Doven, Geophysicist, U.S. Army Corps of Engineers, Denver, CO
Nick Stolte, Environmental Engineer, U.S. Army Corps of Engineers, Louisville, KY
John Jackson, Geophysicist, U.S. Army Corps of Engineers, Denver, CO
Andrew Schwartz, Geophysicist, U.S. Army Corps of Engineers, Huntsville, AL

Advances in near surface electromagnetic induction (EMI) technology partnered with policy changes in the Military Munitions Response Program (MMRP) has resulted in a better understanding of limitations in anomaly density estimations. Recent case studies using advanced EMI sensors with higher resolution, 100% digital mapping, and intrusive investigation of grids has demonstrated that traditional methodologies have drastically underestimated the anomaly density as well as the total amount of work required to remediate Munitions Response Sites (MRSs). Additionally, these advances in technology and policy changes have had significant effects on estimated costs. Data from previous Remedial Investigations (RIs) has proven to be insufficient in quality and quantity to accurately scope a Remedial Action (RA) and develop a realistic cost model.
Underestimation of project costs have been noted in numerous Formerly Used Defense Sites (FUDs) Feasibility Studies (FSs). The costs developed in the FS phase are usually carried forward to the Decision Document (DD), Remedial Design (RD), and subsequently the Cost-to-Complete (CTC). Variables in site conditions including access, vegetation, terrain, and anomaly density are often misunderstood leading to the development of a poor cost estimates. This has resulted in schedule delays and postponement of contract awards. This becomes increasingly critical as the program progresses into the RA phase as the RA is where a significant portion of MMRP project’s environmental liabilities are encountered.
A case study will be presented in this talk which will highlight the uniqueness of site condition variables at each location. Data collection from the RI/FS will be discussed at length and new interpretations of the data will be presented to highlight coverage issues as well as any discrepancies in anomaly densities. Together, this information will be used to correct density estimations on data previously collected during the RI/FS phase as well as re-assess cost estimates.
New MMRP policies and advances in EMI technology have led to discrepancies in density assumptions and cost estimations in previously executed RI/FS phases which have been carried forward to the DD and RD. Subsequently, these estimated costs are also often used to calculate CTC estimates which make up a large portion of the FUDS environmental liabilities. This talk will focus on a case study which will highlight actions currently being taken to correct density estimations and solve cost discrepancies. Current improvements to data interpretation and understanding the impacts to cost models will be discussed; however, further analysis of multiple sites is required to develop an adequate, defensible data set to improve project understanding and costs.


Dave Duggins, Kaarta, Inc, Pittsburgh, PA, USA

Recently, many UneXploded Ordinance (UXO) mapping efforts in wooded areas have started using Simultaneous Localization and Mapping (SLAM) technology to replace/augment Robotic Total Stations for generating global position data. This process involves creating a 3D Geo-referenced map of wooded areas for generating global position data in the woods for Advanced Geophysical Classification and Digital Geophysical Mapping Sensors. This talk will lightly touch on the process used for the creation and geo-referencing of the maps, and evaluating their final accuracy. The presentation will cover some lessons learned in the initial mapping process and the actual localization process when the global position is transmitted over the serial interface. Slides will cover topics such as customizing the GPS NMEA Positional Data string to include roll, pitch, yaw, and localization confidence data. Additional slides will cover visualization of the point cloud data, and the creation of terrain maps. The presentation will also cover the creation of contour maps, terrain steepness maps, and obstacle maps. These obstacle maps can be created in .shp format suitable for comparing against coverage maps in order to validate that all areas have been covered. The presentation will also discuss generating tree Diameter at Breast Height for forestry applications. Finally, the presentation will give recommendations on how to create maps of large areas with limited computing resources.


Trey Dupont-Andrew, Susquehanna University, Selinsgrove, PA, USA
Ahmed Lachhab, Susquehanna University, Selinsgrove, PA, USA

Waterborne Ground Penetrating Radar (GPR) can be a powerful method to survey bathymetry, water, and sedimentation volumes within any water impoundments. Surveys are done without negative impact, and the potential for positive impact stemming from the results is significant. Faylor Lake, a small impoundment in Snyder County Pennsylvania, was the focus of this study. A 100 MHz transceiver was used to collect over 28,000 data points of depth locations spanning the entire lake area, using a 2-person crewed inflatable boat equipped with an electric trolling motor and the GPR apparatus. Collected data has generated contour maps and 3D models of the current bathymetry as well as the original topography of the basin prior to the construction of the dam in 1983. Previous studies have estimated the dielectric of sediments in lakes to calculate the velocity of the electromagnetic waves to estimate the depth of the sediment layer. In this project however, the dielectric of the lacustrine sediment was instead directly measured using a grab sample of sediment from Faylor Lake and analyzed in a lab using a 1600 MHz transceiver. The dielectric constant found from the bench model experiment has allowed for better and more accurate depth values for both bathymetry and sub-bathymetry of Faylor Lake. The sediment volume found using the improved dielectric constant represents 20% of the entire lake volume with 139201 m3 of sediment and 690967 m3 of water. The contour map of the bottom of the lake has also shown the main channel of the old Middle Creek stream. The channel was found at the same location that it had prior to the construction of the dam. The deepest point was found near the dam with a depth of 4.39 m and along the path of the old channel. The average lake depth was 1.63 m.


Mehrez Elwaseif, Jacobs, Houston, USA

Recent developments in multi-channel electrical resistivity systems allowed collecting Electrical Resistivity (ER) and Time-Domain Induced Polarization (TDIP) data simultaneously using multiple electrode arrays. The ER data is routinely used for mapping of near-surface geology and groundwater, characterization of contaminated sites, delineation of engineered structures, and locating voids and fracture zones. However, shallow groundwater and the presence of high fines content materials at survey sites result in ambiguity in identifying different soil deposits (e.g., sand vs. clay) using the ER results alone. On the other hand, the average chargeability of unmineralized sedimentary rocks derived from IP data increases with respect to clay content due to its ionic exchange characteristics, making the IP method a more powerful tool for mapping ore bodies and clay soil than the ER method. However, the signal to noise ratio (SNR) associated with the TDIP measurements is much smaller relative to the resistivity measurements, makes collecting, processing and interpreting the measured chargeability data challenging.
This work demonstrates few challenges associated with processing and interpreting TDIP and ER data through synthetic and field studies. The studies focus on dipole-dipole array configuration given its data acquisition speed advantage compared with other nested arrays. Additionally, this work presents strategies for editing TDIP data (e.g., dealing with negative chargeability measurements) and TDIP/ER processing approaches that offer potential improvements in the final ER and TDIP subsurface models. Despite the ER and TDIP limitations, the combined processing of both data types and other available data (e.g., from bore logs) improve the interpretation of soil heterogeneities and mapping clay soil.


Trever Ensele, Collier Geophysics, Denver, CO, USA
Roy Bowling,Collier Geophysics, Denver, CO, USA
Phil Sirles, Collier Geophysics, Denver, CO, USA
Linsey Chalfant, City of Fort Collins, Fort Collins, CO, USA

Halligan Dam is a 112-year-old 70 foot tall concrete thin-arch dam located 25 miles northwest of Fort Collins, Colorado. The Halligan Water Supply Project by the City of Fort Collins proposes to expand the dam height by 25 feet to expand the potential storage of Halligan Reservoir by 8,100 acre-feet. As part of the design process for the dam expansion, condition assessment of the existing dam was performed. In addition to traditional coring and sampling techniques, a method was developed that could provide insight into the internal condition of the dam that would bridge the gap between point sampling methods. Crosshole Seismic Tomography (CST) is a seismic geophysical method used to image the distribution of seismic compressional-wave velocity (Vp) of geologic material between two parallel or sub-parallel boreholes. This method was adapted to image the Vp distribution of Halligan Dam using the upstream and downstream faces of the dam itself as analogs to the boreholes in the CST method. Using rope-access techniques, sensor arrays (comprised of both hydrophones and geophones) were installed on the upstream face, and hammer-impact sourcing was performed along the downstream face. This “Through-Dam Seismic Tomography” (TDST) technique was performed at three regularly spaced intervals along the length of Halligan Dam. Tomographic sections through the dam structure were computed using the Vp travel times recorded for all active source-receiver pairs for each source location. Results successfully show Vp trends and variations within the imaged sections of the dam. These measurements, alongside traditional concrete coring, provide key information for dam assessment and expansion design.


Allan Foster, LRE Water, Denver, CO, USA
Jacob Bauer, LRE Water, Denver, CO, USA

In support of a municipal water supply evaluation, LRE Water subtracted a geophysical investigation to determine optimum well placement on an alluvial parcel (“Site”) along the Uncompahgre River in western Colorado, USA. The primary goal of this investigation was to evaluate groundwater well yields and evaluate whether the principle of Riverbank Filtration (RBF) will provide water quality benefits at the Site. The stepwise investigation consisted of 1) a geophysical investigation including seven seismic refraction tomography (SRT) surveys and seven electrical resistivity tomography (ERT) surveys, 2) monitoring and test-production well installation informed by the geophysical investigation, 3) aquifer testing and construction of a groundwater model utilizing the results of the ground- truthed geophysical investigation. This case study provides insights into the costs and benefits of geophysical investigations to aid in the design of alluvial well fields.


Richard Funk, Tetra Tech Inc, Bothell, WA, USA
T. Jeffrey Gamey, Tetra Tech Inc, Oak Ridge, TN, USA
Stephen Billings, Black Tusk Geophysics, Vancouver, BC, Canada.

Many former and active Department of Defense ranges and installations have Munitions and Explosives of Concern (MEC) in the underwater environment posing a potential, current, or future hazard. The Army Corps of Engineers has evaluated formerly used defense sites and found that there are more than 10 million acres potentially containing MEC in underwater environments, at approximately 400 sites. Remediation of these sites in a cost-effective manner requires as few deployments and as few false positive targets as possible. This necessitates a highly reliable, single-pass dynamic classification system. Design objectives are to detect and classify medium-size ordnance (e.g. 60-mm mortars and larger) with a single-pass towed platform at minimal standoff distances in water depths between 5 and 150 feet.
There have been a number of technologies developed and tested for underwater MEC wide-area detection, including some 15 projects under the SERDP and ESTCP programs. The objective of the project “UltraTEM Marine towed system for detection and characterization of buried ordnance” is to design and demonstrate a vessel-towed single-pass marine dynamic classification system for wide-area assessment and full coverage surveys. This will be achieved by modifying and integrating Gap Explosive Ordnance Detection’s and Black Tusk Geophysics’ existing UltraTEM® package and associated software into Tetra Tech’s towed electromagnetic array (TEMA) platform, and then demonstrate its capabilities over a series of blind targets at controlled site. If approved, this would be the first marine system approved for advanced geophysical classification in the United States.
Previous versions of this sensor technology have been optimized for European dredging sties, where interest is focused on larger and deeper targets. For DoD munitions response sites, the coil configuration and deployment platform both need to be optimized for smaller, shallower (<1m burial depth) targets of interest (TOI). Dynamic classification of smaller TOI, as well as consistent differentiation between TOI and clutter, requires data with high signal-to-noise ratio (SNR) and multiple transmitter excitation directions.
Initial tests were conducted in spring 2021 in both fresh and salt water in the Seattle area to refine the weight and balance distribution and demonstrate the reliable flight characteristics required for precision surveying. This presentation details the results of the October 2021 shakedown test of the system at Sequim Bay over a variety of targets including ISOs, MEC simulants and clutter in a blind test area. The UltraTEMA-4 successfully passed these initial field tests and modifications based on the lessons learned from this effort are being implemented. The final ESTCP demonstration is planned for 2022 at Sequim Bay.


D. R. Glaser, US Army ERDC Cold Regions Research & Engineering Laboratory, Hanover, NH USA
T. Sullivan, US Army ERDC Cold Regions Research & Engineering Laboratory, Fairbanks, AK USA
A. M. Wagner, US Army ERDC Cold Regions Research & Engineering Laboratory, Fairbanks, AK USA

The energy balance of permafrost depends on omnidirectional thermodynamic and hydrologic gradients. Frequently, studies focus on top-down permafrost thaw in the form of maximum seasonal surficial thaw depth, i.e. active layer thickness—which is used to approximate permafrost vulnerability. However, bottom-up warming has the potential to degrade permafrost more quickly than the top-down energy transfer depending on subsurface conditions. Heat fluxes from geothermal gradients and seasonal groundwater flow can contribute to net permafrost inventory reduction by causing melt within and beneath permafrost. Unlike the active layer, there is no impermeable ice layer to restrict groundwater infiltration, as such the melt water simply rejoins the groundwater. Electrical resistivity tomography methods consistently delimit the lateral extent of shallow permafrost and provide active layer insight; however, limitations in the method restrict the ability to accurately detect and image the bottom of permafrost. When a semi-continuous layer of highly resistive permafrost is present, the measurement geometry of a surface based ERT array is such that signal cannot efficiently penetrate the formation and thus provides an inaccurate depiction of the bottom of permafrost. This phenomenon is present in many layered ERT datasets. High-contrasts in subsurface resistivity layers result in signal attenuation, creating a shadow zone. Further, this erroneous lower layer estimate will likely skew overall permafrost inventory estimates where ERT is used as the primary mapping technology. Here, we demonstrate forward ERT modeling results and a field dataset incorporating depth electrodes to improve the measurement geometry and better resolve the bottom of permafrost. Future work will seek to add additional depth electrodes and monitor the seasonality of the base of permafrost.


D. R. Glaser, US Army ERDC Cold Regions Research & Engineering Laboratory, Hanover, NH USA
B. E. Barrowes, US Army ERDC Cold Regions Research & Engineering Laboratory, Hanover, NH USA
F. Shubitidze, Thayer School of Engineering, Dartmouth College, Hanover, NH USA
L. D. Slater, Rutgers University, Earth & Environmental Sciences Department, Newark, NJ USA

Standoff electromagnetic induction (EMI) sensing techniques offer efficient, in-situ characterization of the electrical properties of soils. Galvanic, or direct, electrical measurements of soils offer the ability to further correlate relaxation responses with soil physical properties through pedophysical modeling. Our efforts seek to leverage the benefits of both standoff EMI and direct electrical measurements. As such, we studied and compared electrical (spectral induced polarization (SIP)) and EMI phenomena over the frequency range of 3 Hz through 10kHz. A low-frequency EMI instrument was developed (LFEMI), modeled after our high-frequency EMI (HFEMI) system (nominally 1kHz – 20MHz), looking to overcome HFEMI’s limitations when measuring physical parameters such as permittivity of low conductivity materials including soils. The HFEMI system was initially designed for the detection of improvised explosive devices and unexploded ordinance which have a high conductivity when compared to soils, resulting in a much higher signal-to-noise ratio making the response easier to observe. We investigated low frequency measurements of low conductivity samples, and where the HFEMI system was less effective at extracting relaxation curves of the inphase and quadrature components of the secondary magnetic field from soil or other media. In addition, HFEMI measurements often contained significant 1/frequency (1/f) instrument noise at low parts of the frequency band, also known as pink noise, that obscured the signal of interest from the soil. The LFEMI employs larger diameter coils with a greater number of turns to improve the signal-to-noise ratio (SNR) at low frequencies allowing measurements as low as 1 Hz. Further, the desired effect of reduced noise and removal of the 1/f noise was substantially achieved. EMI measurements in this broadband, low-frequency measurement range can be correlated with direct SIP complex electrical conductivity measurements of soils. We seek to observe relaxation type responses within the LFEMI signal indicative of mechanistic polarization type phenomena, i.e. ionic charge relaxation. We present LFEMI results for engineered soils consisting of sand-pyrite mixtures with comparisons to direct electrical measurements of the same.


Larisa Golovko, PhD, Landviser LLC, Houston, TX
Yuri Manstein, PhD, Emmanuil Kant Baltic State University and KB Electrometry Ltd, RUSSIA

The multifrequency domain electromagnetic profiling is very promising instrument for fast and detail soil mapping in precise agriculture. Mounting the EM device on a sled or a cart for towing by an ATV can speed up the data gathering process before and after the crop growing season. Mounting the EM device on an UAV provides the unique opportunity to obtain detailed soil maps during crop growth and monitor nutrients uptake by plants (especially nitrogen and potassium). This setup could also be used to access properties of flooded soils under rice production, where access to the field by any other means is not possible.
Several acres of the soils were mapped by authors using AEMP-14 – multifrequency (up to 14 frequencies) EM profiler with 2.5 – 250 kHz frequency range. The device signal was processed to convert it into soil resistivity. In conductive soils, where resistivity is less than 100 Ohm m, the penetration depth depends on the frequency, and the signal can be converted to the apparent resistivity with good accuracy. The EM data were compared to the laboratory tests of soil samples collected in key locations of soil electrical resistivity map, the strong correlations were found between EM data and soil pH and available water at different depths.
However, the EM signal strongly depends on soil conductivity and height of the device above the ground. The height can be adjusted through the construction of the device carrier. In some cases, in the resistive soils the signal can be too weak for a reasonable data processing into resistivity.
We propose a solution through a combined methodology, where EM data are calibrated with DC resistivity mapping at some depths and with vertical electrical sounding (VES) measurements at key locations rather than just by collecting and analyzing soil samples. All DC resistivity measurements can be carried by compact LandMapper instrument with various probes and cable sets, easily adjustable for different soil profiles. This combined methodology of EM and DC resistivity measurements shows a potential to be streamlined to quickly and non-destructively map agricultural soils under various crops and farm practices.


Katherine Grote, Missouri University of Science and Technology
Yunyi Guan, Missouri University of Science and Technology

Multispectral data acquired with unmanned aerial vehicles (UAV) are frequently used in agricultural applications to assess crop health or predict yields. When compared to most ground-coupled geophysical technics, UAV data are easy to collect, inexpensive, easy to process, and have high resolution. However, the penetration depth of UAV data is usually small, and multispectral data are not currently used to monitor soil properties. In this research, multispectral data were acquired using UAVs, while ground-coupled geophysical data were acquired to provide information on soil properties. The geophysical techniques used were ground penetrating radar (GPR), which provided information on soil volumetric water content (VWC), and electromagnetic induction, which provided electrical conductivity (EC) measurements. While multispectral data are not inherently sensitive to VWC or EC, vegetation vigor can be influenced by these properties. Thus, measurements of vegetation vigor from multispectral data have potential to correlate to soil properties. In this research, machine learning (random forest method) was used to predict VWC and EC based upon multispectral data. Results showed that multispectral data can be used to improve prediction of these parameters.
This research was performed at an agricultural research station in an 8.9 ha field with claypan soil. The field was divided into 56 plots that varied by crop type (corn, soybeans, or alfalfa) and by drainage (no tile drainage, moderate drainage, intensive drainage). Multispectral data were acquired over the entire field when the crops were mature, while geophysical data were acquired somewhat earlier. Geophysical data were acquired twice, once when the soil was fairly dry and once when the soil was saturated. Geophysical data were acquired in 18 traverses across the site, and the high-resolution multispectral data were averaged within the footprint of each geophysical technique. Machine learning was performed for both wet and dry conditions and assuming different levels of knowledge about the field; analysis was done assuming the crop type was unknown (all crops combined), assuming the crop type was known (each crop independently), assuming the drainage configuration is unknown, and assuming the drainage configuration was known. Results showed that multispectral data were most accurate in predicting VWC and EC when both crop type and drainage configuration was known, and knowledge of the drainage configuration had more impact on prediction than did crop type. Prediction was more accurate in dry soil than in wet, but this may be partially due to the lower overall variability observed in drier soils. An analysis of the predictive strength of different types of multispectral data (vegetative indices (VI) and raw multispectral data) showed that VI which were most important for prediction of VWC and EC were NDRE, VARI, and blue band data.


Mats Lundh Gulbrandsen, I•GIS, Aarhus, Denmark
Tom Martlev Pallesen, I•GIS, Aarhus, Denmark
Thomas Bager Rasmussen, I•GIS, Aarhus, Denmark
Thomas Mejer Hansen, Aarhus University, Aarhus, Denmark
Ulrika Sabel, VA Syd, Malmö, Sweden
Nafyad Serre Kawo, University of Nebraska-Lincoln, Lincoln, NE, USA

The importance of handling uncertainty of data and being able to present the ambiguity of geo models of any kind have got more and more attention the last couple of years. In the same way as it is important for a geo modeler to understand the uncertainty and limitations of data to make an adequate geo model, it is important for a decision maker to understand the uncertainty and limitations of a geo model to perform adequate and qualified decisions. Whether the geo model is made to manage infrastructure projects or nature resources like oil, gas, or groundwater, being able to understand the uncertainty and limitations of a geo model potentially have great economic value.
One way of presenting the uncertainty and ambiguity of a geo model is to present a suite of geo models instead of just one. Due to the underdetermined nature of the inverse problem to be solved (making a geo model is an inverse problem), several solutions will all fit the available data and information about the problem. Multiple Point Statistics (MPS) is an overall methodology representing a series of simulation algorithms all utilizing the available information (geophysics, boreholes, background knowledge, etc.) to produce a series of geo model realizations all fitting the available information. Generating many of these realizations allow for any kind of statistical computations to be made: “What is the probability of having sand at a specific location and depth?”, “What is the probability that location A and B are placed in a fully connected clay layer?”, etc.
In this presentation we will present the results from 3 different case studies where MPS modelling is utilized in a hydrogeological context. The 3 case studies are from the Indian Wells Valley in California, the Shell Creek watershed in eastern Nebraska, and the Alnarp Valley in the southern part of Sweden.


Samie Hamad, Missouri University of Science and Technology, Rolla, MO, USA
Salah Shaniba, Mellitah Oil & Gas B.V, Tripoli, LIBYA
Wajdi Ammar, University of Colorado, Denver, CO, USA

Non-destructive testing tools, such as ground penetrating radar (GPR) and the portable seismic property analyzer (PSPA), extensively used in the past two decades for monitoring, quantifying, and mapping the deterioration of bridge decks. Using PSPA and GPR ensures regular monitoring of bridge conditions, leads to the early detection of deterioration, and plays a major role in bridge serviceability. This is important, as not knowing the integrity of bridge decks increases maintenance costs and presents public safety hazards.
The goal of this paper were to address the condition of August A. Busch bridge deck owned by the Missouri Department of Conservation (MDC), also generated plan-view maps, showing the thickness of the bridge deck and details of the pattern, placement, and density of the deck’s reinforcement steel bars. This is significant because the MDC no longer has design details or as-built drawings. This research also assessed the capability and compatibility of the three different assessment approaches (visual inspection, PSPA, and GPR) when used together. If this technology proves to be cost-effective, the MDC can acquire these data for each of Missouri’s bridge decks.
Visual inspection, GPR, and PSPA data were acquired on the bridge deck located at August A. Busch Conservation Area. Over 90% of the August A. Busch Bridge’s deck was in good condition with an average compressive strength of over 2500 psi. Ground penetrating radar data showed no indication of significant deterioration. the overall bridge deck was determined to be in fair to good condition, and it was recommended that the August A. Busch Bridge deck be inspected every 24 months.


Sherif M. Hanafy, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia
Nawaf A. Al-Omari, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia
Ahmed AlShaibani, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia

High-resolution near-surface seismic imaging is required for many applications such as solving shallow geology, engineering, or environmental studies. It, also, has become important in calculating accurate static corrections for deep reflection processing. Near-surface seismic applications require the investigation of the subsurface velocity model for depths less than 300 m from the ground surface. In order to achieve the required high-resolution subsurface investigation, the seismic data must be recorded using high-fold and densely spaced receivers. The recording of high-fold and densely spaced receivers’ seismic data with conventional acquisition techniques is highly time consuming and requires labor for receiver planting and cable moving, which is a major obstacle in shallow applications of the seismic method.
Contrary to the conventional acquisition, the land-streamer acquisition saves a lot of field-time since all receivers and cables are pre-connected and the time required for the field setup is almost 20% of the time required to setup the conventional survey. Time-saving factor is even more distinct in case of recording a very long profile where in conventional surveys, the profile must be divided into small sections and the receivers/cables must be collected and re-installed after recording each segment, which is very time consuming, however, in land-streamer these breaks are not required.
In this work we recorded one seismic profile using both conventional and land-streamer setups. The total profile length is 955 m. In case of conventional setup, we divided the profile into three segments, each one is 475 m with 50% overlap between each two segments. Here, we recorded 192 shot gathers, each one has 96 or 144 receivers, depending on the shot position. The shot and receiver intervals are 5m. It took 2 field days (10-working hours/day) to complete the conventional survey. Another 265 shot gathers were recorded on the same profile using a 48-receiver land-streamer with 1m receiver intervals. Shot gathers 1 to 75 has shot intervals equal to 5m, while shot gathers 75 to 265 has shot intervals of 2.5m. The land-streamer survey was completed in about four working hours.
The first arrival traveltimes of both profiles are manually picked and then inverted to generate the two corresponding traveltime tomograms. The conventional traveltime tomogram shows a depth of penetration of 150m from the ground surface, where four subsurface layers and a possible fault at offset 480m are shown on the subsurface tomogram. The land-streamer tomogram shows a maximum depth of penetration of 30m from the ground surface. It reveals more information about the lateral and vertical velocity variations and detailed shape of the shallow reflectors relative to the conventional tomogram. It also shows the subsurface fault, however, at shallower depth relative to the conventional tomogram.
To confirm the field result, one synthetic numerical test is conducted. The first arrival traveltimes are calculated using ray-tracing algorithm. In this synthetic test, we mimic both the conventional and the land-streamer field surveys. The synthetic tomograms show similar results to the field tomograms.
In summary, the land-streamer survey is much faster in the field and shows more details of the near-surface layers, however, the conventional survey can reach larger depths relative to the land-streamer.


Thomas Mejer Hansen, Aarhus University, Denmark
Tom Martlev Pallesen, I•GIS, Aarhus, Denmark
Thomas Bager Rasmussen, I•GIS, Aarhus, Denmark
Mats Lundh Gulbrandsen, I•GIS, Aarhus, Denmark
Ulrika Sabel, VA Syd, Malmö, Sweden

Collection of electromagnetic (EM) data (from airborne systems, boats, towed instruments on the surface) has found widespread use, as it allows efficient imaging of large areas of subsurface variations in resistivity, that has a variety of uses as for example for ground water management, contamination analysis, permafrost imaging, and mineral exploration. Traditionally the analysis of such EM data has been done in a sequential order: First EM data have been inverted to resistivity volumes, which have then been analyzed by geologists, hydrologists or petro-physicists, depending on the use case. This leads to a workflow in which it is practically impossible to account for uncertainty, not least because analysis is typically performed on a single optimal model (both the output from the geophysical analysis, and the subsequent geological analysis). Further, most all available geophysical inversion methods used to convert EM data into resistivity volumes, make implicit assumptions about the subsurface variability (such as simplicity/smoothing). In practice this means that the produced resistivity volumes may by construction be inconsistent with the actual subsurface geology.
Recently probabilistic inversion methods have become available that allow taking into account geological information, directly in the inversion, as prior information. This means that the outcome of the inversion will by construction be consistent with known assumed geological variability. And, the outcome can be direct estimates of geological features, such as: what is the probability of locating a potential ground water reservoir? To fully utilize such a probabilistic framework, the main challenge is to quantity geological information, by construction of an algorithm that can generate realizations of an assumed statistical model representing geological structures and associated uncertainty.
Using airborne EM data from the southwestern Sweden, we will demonstrate how to implement such a probabilistic approach to directly quantify the probability of locating a specific gravel layer of interest. We will do this in a two-step process: The first involves quantifying prior information to describe a probabilistic conceptual geological model, that represent expected layers, and their potential thicknesses. Then we describe a probabilistic link between geological properties and the associated resistivity distribution. This represents a joint prior conceptual geological and resistivity model. Once this has been established, we apply a probabilistic inversion that allow quantifying the posterior probability of any feature that is described in the prior model. Here the main focus is the existence of a relatively thin gravel layer, that has potential to contain easy to produce drinking water.
We present a novel framework, and the methods needed to implement the framework, that allow quantifying jointly, and unbiased, geological and geophysical information into one model, from which decision makers can directly get answers, consistent with the available information.


David Harro, G3 Group USA
Henok Kilfu G3 Group, USA
Glen Anderson, Wood, USA

The use of geophysical assessment of dams in deep karst environments can yield valuable information for such complex geology. One of the most used methods in the assessment of dams over karst is electrical resistivity (ER). ER has progressed from 2D resistivity curves to inversion of complex array patterns that produce image. There are however limitations to the current data collection and inversion methodology that prevents higher resolution and higher quality images. These limitations are based on the geometry of array and data collection that results in trapezoidal-shaped inversion data set with decreasing data at depth.
As electrode spacing i.e. 1a, 2a, 3a, etc. increases, the number of measurements decreases for each progressively deeper data level. This is a fundamental limitation of the technique and this results with an inversion of progressively less data the deeper the survey, thus lower resolution images. The bottom third of the surface ER represents only 10.5% of the data for the inversion. The reduced data with depth results in a trapezoidal image that has approximately 40% blank space of the total rectangular area dir3ectly below the array. In practice this results in the inability of the ER image to collect information on the abutments of embankment dams and has low resolution at depth.
A unique technique called the Multi-Electrode Resistivity Implant Technique (MERIT) overcomes some of the fundamental limitations of surface ER. MERIT utilizes a combined surface array and array of small permanent implants driven to depths of up to 50 feet. The tomographic configuration of the surface and deep buried array results in overlapping field density that significant increases data acquisition of up to five times and can increase the depth by third to twice that typical surface ER array. At depth the overlapping fields of MERIT optimized data collection significantly increases the data points in the lower third of the inversion that results in higher resolution than surface ER. The MERIT image is a full rectangular image and as such can be used to evaluate the ends of the array such as at image abutments of an embankment dam.
A case study is presented an embankment of a dam in Florida where deep buried sinkhole features were identified over 250 feet deep. Due to the higher resolution of the MERIT images, the location and measurements of the sinkhole throats were possible and were encountered at depths of 100 feet. The permanent implants are being utilized to perform 3D surveys of the sinkhole features and are being incorporated into the long-term monitoring plan for the site.


Koichi Hayashi, Geometrics, San Jose, CA, USA
Stefan Burns, San Jose, CA, USA
Cari Roughley, Napa Valley College, Napa, CA, USA

Napa Valley is located in North San Francisco Bay Area about 50 km from San Francisco, California. The Valley is well-known for its wine as the surrounding mountain ranges protect it from cold, damp onshore breeze from the Pacific Ocean, maintaining the valley’s warm and dry climate most suitable for growing grapes. The mountain ranges around the valley are rising by several active faults parallel to the famous San Andreas Fault and the valley is under threat of earthquakes. The 2014 South Napa earthquake occurred in South Napa County on August 24 at 3:20 a.m. Pacific Daylight Time, measuring 6.0 on the moment magnitude scale. The epicenter was located southwest of downtown Napa, approximately 6.0 km northwest of American Canyon near the West Napa Fault. The earthquake was the largest in the San Francisco Bay Area since the 1989 Loma Prieta earthquake. Significant damage and several fires were reported in the southern Napa Valley area, including damage in the nearby city of Vallejo within Solano County. The earthquake caused one death and injured approximately 200 people. Clear surface rupture appeared along West Napa Fault for at least 5 km. The damage from the earthquake was concentrated between downtown Napa and St Helena Hwy (CA 29). Near-surface geology possibly contributed to the concentration of the damage. To reveal the cause of the damage and evaluate future earthquake risk, we estimated a three-dimensional (3D) S-wave velocity (Vs) model of the Napa, California, U.S. using microtremor array measurements (MAM) and horizontal to vertical spectral ratio (H/V) at approximately 100 sites. The investigation area is approximately 12 km by 16 km at Napa, CA including the valley floor and surrounding hills, Mayacamas Mountains on the west and Vaca Mountains on the east, bounded by the West Napa Fault and the Soda Creek Fault. MAM was collected with eight to twenty 2 Hz geophones, and the maximum receiver spacing ranged from 30 to 1500 m. Ambient noise for MAM and H/V were collected for 20-120 minutes. A spatial auto-correlation (SPAC) method calculated phase velocities from the vertical component of ambient noise. Minimum frequency of dispersion curves ranged from 1 to 10 Hz. H/V was calculated from three-component (3C) seismic ambient noise using a single 3C 2Hz geophone. The peak frequency of H/V ranged from 0.25 Hz to 10Hz. There were clear differences between valley floor and surrounding hills both dispersion curves and H/V. In the H/V spectra, there is a clear peak of H/V at a frequency of 0.3 Hz in the valley floor sites whereas there is no clear H/V peak below 1 Hz in the hill sites. Joint inversion of a dispersion curve and H/V spectrum estimated Vs profiles to 30 m to 1000 m depth. There is a large difference in resultant Vs profiles. Depth to a shallow engineering bedrock with Vs of 760 m/s is 300 m and 30 m at typical valley floor sites and hill sites respectively. It indicates that the velocity model changed considerably along the Soda Creek Fault. The result of inversion and geological model implies that the low frequency peak of 0.3 Hz at the valley floor site is mainly due to a deep bedrock with Vs more than 2500 m/s at approximately 1000 m depth. We compiled all Vs profiles together with the 3D Vs model based on geological information and estimated a preliminary 3D Vs model to a depth of 1000 meters. The VS30 obtained from the MAM ranged between 200 m/s and 970 m/s. Clear H/V peak frequencies of 0.25 to 0.4 Hz were consistent in the valley floor. The depth to the bedrock with Vs of 760 m/s ranged between almost surface to greater than 300 m.


Scott Ikard, U.S. Geological Survey, Austin, TX, USA

In the Gulf Coastal Plain of south-central Texas, the rocks that contain the Carrizo-Wilcox aquifer crop out at the land surface along a relatively narrow band that strikes from southwest to northeast. The lower Guadalupe River is incised into the outcrop and is forming a depositional alluvial terrace on top through valley filling and entrenchment into the fill. Surface-water (SW) and groundwater (GW) exchanges are difficult to map in the river and quantify by streamflow gaging because streamflow is regulated by dam releases. A water-borne self-potential (WaSP) logging survey was therefore completed in 2016 to map gaining and losing sub-reaches in the lower Guadalupe River by measuring the electrical streaming-potential in the river along a 15-kilometer long profile across the outcrop. The WaSP geophysical inverse problem is formulated herein and solved to confirm qualitative interpretations of apparent gaining and losing sub-reaches inferred from the 2016 WaSP survey data and demonstrate that the electrical potential measured in the lower Guadalupe River was a streaming-potential attributed to superimposed SW-GW exchanges occurring simultaneously over variable spatial-scales. Qualitative interpretations of SW gain and loss from WaSP data are supported by the regularized inverse models of the geospatial distributions of streaming-current sources and sinks on and beneath the riverbed, which reproduce the streaming-potential measured in the SW and characterize gaining and losing reaches of the river by the opposing dipolar electrical polarities of streaming-current and GW sources and sinks on the riverbed. The inverse modeling results indicate that WaSP surveying is a viable method to map SW-GW exchange processes in rivers and may enable quantification of streamflow gains and losses over variable spatial scales if a petrophysical relation between electrical resistivity and riverbed permeability can be established.


Ahmed Ismail, Oklahoma State University, Stillwater, OK, USA
Patrick Meese, Oklahoma State University, Stillwater, OK, USA
Oluseun Sanuade, Oklahoma State University, Stillwater, OK, USA

We acquired five seismic reflections lines at the Tri-State Mining District in northeast Oklahoma using the seismic land streamer to detect unmapped mining voids. Detecting subsurface voids is always required to reduce risks to both property and human life. This study was conducted at the Tar Creek Superfund Site characterized by hazardous toxic levels of zinc, cadmium, and lead in mine waste, soil, air, and water. Despite recent improvements in subsurface void detection using geophysics, hazardous conditions can limit the applications of many geophysical methods in void detection. This study tested the effectiveness of the seismic reflection data acquired by land streamer in void detection. The acquired seismic data were analyzed as P-wave reflection, refraction, and multichannel analysis of surface wave to generate multiple images of the subsurface and maximize the chance of detecting voids. The results showed that the seismic land streamer data have successfully detected multiple mining voids along the acquired profiles. This study demonstrated that a seismic land streamer is an effective tool for acquiring suitable data to detect subsurface voids.


John Jansen, Sr. Geophysicist, Collier Geophysics, West Bend, WI

The hydrogeologic parameters and structure of an aquifer are the most critical elements that determine the success of a ground water study such as siting a production well, developing a groundwater management plan for a basin, or building a reliable groundwater model. Unfortunately, these parameters are largely unknown prior to committing to a site and drilling a well. Drilling to deep aquifer is expensive and the amount of available well data is often very limited. It is also difficult to predict how far the conditions encountered in a well can be extrapolated within the basin and many significant hydraulic features are often missed or poorly understood. As a result, many groundwater projects are hampered by limited by sparse subsurface information and many well projects have failed due to unexpected or poorly understood stratigraphic or structural complexities. Methods that can map aquifer thickness and porosity, identify faults, fracture zones, and discontinuities in confining units at lower cost and with greater data density are critically needed in many groundwater basins.
Seismic reflection technology has been developed by the oil and gas industry to map subsurface reservoirs in detail. This technology is now economically feasible for use in the groundwater industry. Seismic reflection surveys produce high resolution images of the subsurface and can be used to identify favorable aquifers, map the presence and continuity of confining units, and identify faults or fracture zones that may create hydraulic boundaries. Seismic reflection surveys are able to map fine scale stratigraphic details to depths of several thousand feet. Modern processing and interpretation techniques can identify permeable sand zones, faults, and other stratigraphic and structural features that control well yield. New high-resolution seismic reflection surveys can be designed to image the zone of interest and provide far superior visualization of the subsurface than can be obtained by other methods. Unfortunately, the cost to acquire high resolution seismic reflection data is relatively high which has limited the application of the method for water supply applications.
Fortunately, in many areas seismic reflection data has been collected for oil and gas exploration or other objectives. This data is often available for purchase at a nominal price, though the quality and acquisition geometry of older data is generally inferior to contemporary high-resolution surveys. The quality of vintage data can often be improved through reprocessing using contemporary techniques and by optimizing the choice of parameters to focus on the intervals of interest, which tend to be shallower than the original survey objectives.
The use of pre-existing vintage seismic reflection data will be demonstrated by a case history in Central California where seismic data collected for geothermal studies in the 1990s has been reprocessed to map the structure and stratigraphy of aquifer zones and confining units at depths of 400 to 2,000 feet. Attribute processing was used to map the net and gross sand of the major aquifer intervals, map the thickness and continuity of clay confining units, and map faults and fracture zones. The data is being used to calibrate the data from a regional airborne EM survey and improve the conceptual hydrogeologic framework, which will assist selecting potential future ASR and production well sites.
A second case history will document the use of a contemporary 3D petroleum survey to map aquifer units at depths of 400 to 2,000 feet. A ERT survey was added to explore the upper 400 feet of the site. The survey located several wells sites and led to the development of an approved water supply of 400 acre feet with a non-tributary designation.


M. Andy Kass, HydroGeophysics Group, Aarhus University, Aarhus, Denmark
Denys Grombacher, HydroGeophysics Group, Aarhus University, Aarhus, Denmark
Matthew Griffiths, Dept. of Electrical and Computer Engineering, Aarhus University, Aarhus, Denmark
Mathias Østbjerg Vang, HydroGeophysics Group, Aarhus University, Aarhus, Denmark
Lichao Liu, HydroGeophysics Group, Aarhus University, Aarhus, Denmark
Jakob Juul Larsen, Dept. of Electrical and Computer Engineering, Aarhus University, Aarhus, Denmark

The signal-to-noise ratio remains a major challenge in surface nuclear magnetic resonance (NMR), where low-amplitude NMR signals are measured in areas with high ambient noise such as near urban areas or infrastructure. While significant advances have been made in signal processing, transmit schemes, and technology, noise remains a critical issue, often requiring large numbers of stacks to extract useful information. This approach is time-consuming and can limit the number of sites measured per day, reducing the regional mapping capabilities of surface NMR.
Recent advances in transmitter and modelling capabilities have allowed for a greater variety of transmit schemes such as a steady-state protocol. This transmitter sequence provides an orders-of-magnitude improvement compared to standard free-induction decay measurements (FID), providing high-fidelity groundwater measurements at high-speeds in areas previously inaccessible to surface NMR due to noise.
The steady-state approach uses a pulse train of identical pulses separated by a constant repetition time. In contrast to FID sequences, which require several seconds per stack and potentially hours to acquire a full sounding at one site, the close separation between transmit pulses in the steady-state scheme eliminates the wait time between subsequent observations. By varying the current between sequences—similar to an FID—as well as the transmit pulse length and repetition time, a sounding can be performed at a fraction of the time of FID measurements with a similar or improved signal-to-noise ratio.
We have acquired well over 100 sites with the steady-state sequence, acquiring up to 16 sites in a single day using the Apsu instrument developed by the HydroGeophysics Group (HGG). We have developed an associated processing workflow to produce inverted models of water content, T2*, and T2 with depth at each. Results are consistent with FID measurements and nearby water table depths/aquifer tests where available.
Here we present the background of the steady-state approach and discuss the modular processing workflow. We then show a selection of examples of data and associated models from a variety of noise conditions in Denmark.


M. Andy Kass, HydroGeophysics Group, Aarhus University, Aarhus, Denmark
Pradip Maurya, HydroGeophysics Group, Aarhus University, Aarhus, Denmark
Thomas Højland Lorentzen, Dept. of Civil Engineering, Technical University of Denmark, Lyngby, Denmark
Jeppe Frost Schjerning, Dept. of Geoscience, Aarhus University, Aarhus, Denmark
Jesper Pedersen, HydroGeophysics Group, Aarhus University, Aarhus, Denmark
Anders Vest Christiansen, HydroGeophysics Group, Aarhus University, Aarhus, Denmark

Permafrost degradation is increasingly becoming one of the world’s most pressing climate problems in polar and sub-polar regions, having huge impacts on infrastructure and the environment. It is therefore vital to understand subsurface dynamics and structures in these regions. Recently developed and advanced geophysical sensors such as the towed Transient ElectroMagnetic system (tTEM) can be utilized to gain a detailed understanding of subsurface conditions in permafrost regions through rapid mapping and integration with multiple disparate datasets. However, interpretation of electrical and electromagnetic data is complicated in these areas by strong induced polarization (IP) effects often associated with permafrost and underlying geology.
A survey utilizing an adapted tTEM system (SnowTEM) in Ilulissat, Greenland provides a large-scale example of data and associated preliminary inversion models incorporating IP effects. The purpose of the survey was two-fold: to understand 3D distribution of (specifically saline) clays, and to delineate the subsurface geometry within areas planned for development of road systems and a new airstrip. In order to provide information for planning decisions in terms of engineering stability and a greater understanding of how a warming climate and melting permafrost will affect local infrastructure and water supply, the effects of IP and their consequences to interpretation must be better understood.
The SnowTEM survey made a 3D coverage of an area of approximately 640 hectares with a nominal 25m line spacing. In the presence of marine clays, permafrost, and possible schists, the recorded voltage decays show a wide range of IP effects, manifesting as steeply decaying curves, sign-shifts, non-monotonic decays, and completely negative data. Therefore, inversion taking polarization effects into account was critical. We present a catalogue of varying IP effects and their associated inverted models using a hybrid stochastic approach to understand their mapping from the data space to the model space. We show the dependency of IP inversion on the starting model, discuss optimal filtering parameters in the context of IP, and present a pathway toward inversion of a complete dataset.


Kevin Kingdon, Black Tusk Geophysics, Inc.
Len Pasion, Black Tusk Geophysics, Inc.
Laurens Beran, Black Tusk Geophysics, Inc.
David Sinex, Black Tusk Geophysics, Inc.
Heesoo Chung, HydroGeologic, Inc.
Charles Nycum, HydroGeologic, Inc.
Meghan West, HydroGeologic, Inc.
Stephen Billings, GapEOD
William Rowlands, GapEOD

The UltraTEM sensors are dynamic advanced geophysical classification (AGC) sensors which acquire high resolution dynamic AGC data at production rates comparable with traditional DGM arrays while simultaneously providing the ability to screen or classify targets. The UltraTEM Classifier is a 5 transmitter and 11 receiver cube EMI array designed for “one-pass” classification performance, while the person portable UltraTEM Screener is a 1 transmitter, 6 receiver cube EMI array designed for detection and informed source selection (ISS). At MSFC-003, the UltraTEM Classifier and UltraTEM Screener were deployed as part of the Interim Measures implementation. The MSFC-003 site features a range of munitions types with the potential for chemically configurable UXO, areas of very high target densities, infrastructure and subsurface utilities, and a variable magnetic geology across the site. These factors combine to create a challenging AGC problem. The UltraTEM Classifier was used to collect approximately 25 acres of full-coverage grid data. In areas that were inaccessible to the Classifier towed array, the person-portable UltraTEM Screener was used to survey an additional 3 acres for which cued target locations were selected via ISS for subsequent cued surveying with a MetalMapper 2×2 sensor. In areas where RTK GPS was not available, the UltraTEM screener used the Kaarta Stencil for positioning. These data were then used for anomaly detection, inversion, and screening or classification. In this presentation, we provide details of the data collection, processing steps, and classification results.


Ahmed Lachhab, Susquehanna University, Selinsgrove, PA, USA
Benjamin Nicolson, Susquehanna University, Selinsgrove, PA, USA
Skylar A. Brion, Susquehanna University, Selinsgrove, PA, USA

Seismic Refraction, Electrical resistivity surveys in addition to four drilling logs were combined to investigate the hydrogeology of an aquifer at the Center for Environmental Education and Research (CEER) in Susquehanna University. The equipment setup used in this study encompassed two Geode Exploration Seismograph (by Geometrics) with 48 geophones and a SuperSting R8, (by AGI) with 56 electrodes. The surveys conducted used different arrays of geophones and electrodes with different spacing both as a gride and as individuals transects in specific places. This was done to reach different depths going from shallower to deeper levels into the subsurface and explain in-depth certain questionable areas. SRT showed better results of the nature of the geology to the bedrock formation beneath CEER while ERT provided better information of the hydrology of the site and clearly identified the preferential flow present in this area instead. The stratigraphy at CEER as was eventually confirmed by the drilling logs and was found to represent an alluvial deposit with sand, gravel, silt loam, and cobbles at the bottom of this formation. These layers were not uniformed, throughout the site as the graded aspect was not respected in certain areas. This variability of the geology was also found to be responsible in generating multiple preferential flow and this was visible in combining all three methods used in this study.


Mike Law, DMT Geosciences Ltd, Calgary, AB, Canada
Jane Dawson, DMT Geosciences Ltd, Vancouver, BC, Canada

Geophysical investigations can be a powerful tool to explore and delineate physical properties of the subsurface. In the planning stage of remediation and reclamation, geophysics provides valuable guidance for borehole location, soil sampling, and delineating. Electromagnetic (EM) surveys are the most used preliminary geophysical methods to explore the terrain conductivity at a site of interest, but other methods, such as pseudo-3D imaging can provide high-value horizontal and vertical delineation at much greater depths, without many of the drawbacks and limitations of EM surveys.
This study focuses on a collaboration project between DMT Geosciences Ltd and two other consulting firms focused on historical oil and gas sites located in southern Alberta, where geophysics and previous soil sampling information was combined to inform a complex Phase II Environmental Site Assessment. Two sites are highlighted: one site with complex topography that relied heavily on geophysics for preliminary delineation of salt impact; a second site illustrates how pseudo-3D resistivity imaging captured salt impact that extended to depth, well beyond where EM surveys failed to delineate the true extents of impact.
This project demonstrates the effectiveness of strong collaboration between multiple consulting firms. Phase II borehole sampling from the environmental firm is compared with the most recent geophysical results to identify contours of resistivity / conductivity that suggest environmental significance for remediation planning. Both geophysical and borehole data previously acquired by other firms were incorporated into this study in order to best inform the most recent field-work and ensure the best possible assessment and remediation strategy is performed.


Jeffrey Leberfinger, PGp, PG, PIKA International, Harrisburg, PA
Craig Murray, Parsons, Denver, CO

Availability of Advanced Electromagnetic Induction (EMI) Sensors remains a significant factor in the success of Advanced Geophysical Classification (AGC) implementation on Munition Response Projects. The presentation will provide an update on the current status and availability of existing advanced EMI sensors; the Geometrics MetalMapper 2×2 (MM 2×2), AcornSI’s Man Portable Vector (MPV), NovaTEM, LLC ‘s TEMSENSE, Gap EOD UltraTEM system, and the White River Technologies, Inc. Dynamic APEX system. While all the systems have been approved by the Department of Defense AGC Accreditation Program (DAGCAP) and have been used on AGC projects in 2021 and 2022, the current manufacturers of the sensors continue to make improvements on their functionality, durability, and data quality based on feedback from their use on AGC projects from both industry and government.


Xinglin Lu, School of Electrical Engineering, Chongqing University, Chongqing 400044, China
Xuquan Hu, School of Electrical Engineering, Chongqing University, Chongqing 400044, China
Zhengyu Xu, School of Electrical Engineering, Chongqing University, Chongqing 400044, China
Xian Liao, Chongqing Triloop Prospecting Technology Co., Ltd., Chongqing, 402660, China
Longhuan Liu, Chongqing Triloop Prospecting Technology Co., Ltd., Chongqing, 402660, China
Zhihong Fu School of Electrical Engineering, Chongqing University, Chongqing 400044, China

The water-rich giant karst cave is the risk source of tunnel construction in the karst area. To reduce the risk of tunnel construction, it is necessary to accurately explore the spatial distribution range of the karst cave in the direction of advance and both sides of the tunnel. Reflection seismic and transient electromagnetic (TEM) are the primary geophysical tools for tunnel advanced prediction, they have strengths and weaknesses. The tunnel seismic can only describe the boundary interface between cave bodies and surrounding rock in the direction of advance. Although the TEM method can detect the spatial distribution range of cave bodies, the inversion results depend on the initial resistivity model. The single method is insufficient to describe the spatial distribution of cave bodies. To overcome the shortcoming of a single method, we developed tunnel seismic and TEM joint detection of karst cavern and established a complete data processing flow. The tunnel seismic migration profiles can describe the interface between the karst cave and surrounding rock in the direction of advance. We put forward a reasonable initial model: 1) The layer’s thicknesses are determined by the impedance interface from the tunnel migration data; 2) The initial resistivity values are determined from the pilot holes and prior geological data. Comparison of the inversion results of different initial models of tunnel face horizontal line data proves that the proposed initial model method can reduce multi-solution, save calculation time, and improve inversion accuracy. The multi-directional TEM inversion profiles can describe the spatial distribution of the giant cave. Combined with the results of tunnel seismic and TEM, the interpretation errors can be reduced, and the extension of the karst cave can be delineated. Later excavation results also verify the accuracy of the prediction results.

SAGEEP 2022 Denver, Colorado USA


Kristian Macias, University of Mississippi, Oxford, MS, USA Dr. Ron Counts, University of Mississippi, Oxford, MS, USA

The Adams Mill fault is a high-angle reverse fault that thrusts Piedmont bedrock ~2.4 meters over an unconsolidated gravel deposit at the original entrance of the Smithsonian National Zoological Park in Washington, D.C. The fault was exposed more than 100 years ago in a road cut near the intersection of Adams Mill Road NW and Clydesdale Place NW, but the region is aseismic, and the gravel was presumed to be millions of years old, therefore, the fault was never considered to be a serious seismic hazard. However, in 2017 the USGS obtained a luminescence age of 451 ± 34 ka for the faulted gravel, but the fault could be significantly younger. Early 20th century construction projects have revealed several other faults in the city, including the 18th and California Street NW, the Calvert Street faults. Furthermore, United States Geological Survey (USGS) drilling in the late 20th century identified near-surface sediments with increased displacement with depth, interpreted as evidence of multiple episodes of faulting. These faults are aligned along the same strike, and a USGS map published in 2017 shows them as one continuous fault connecting the Rock Creek Shear Zone to the Stafford Fault System, traversing under several important National Monuments and buildings that include the White House and the Washington Monument.
Dozens of geophysical surveys using electrical resistivity tomography (ERT), ground penetrating radar (GPR), and refraction microtremor seismic (ReMi) were acquired in downtown Washington D.C. to map the subsurface, locate the Adams Mill fault, if possible, and test how these methods worked in a highly urbanized environment. The GPR and ERT surveys conducted at the National Zoo and Washington Monument clearly show the fault in multiple east-west profiles. Near the Adams Mill fault exposure at the Zoo, multiple GPR profiles show what appear to be additional faults with less displacement, suggesting the Adams Mill fault is part of a larger fault zone. If fault slip was distributed across multiple faults, the magnitude of the resulting earthquake could be larger than expected for the observed 2.4 m of displacement.
The 2011 Mw 5.8 Mineral, Virginia earthquake had an epicenter 125 km away, yet caused approximately $300 million in damage to Washington, D.C, demonstrating how well seismic energy is transmitted and amplified through coastal Plain sediments. Consequently, the Adams Mill fault could pose a considerable seismic hazard to the city that was previously unknown.


Md Lal Mamud1, 2, Parsa Bakhtiari Rad1, Leti T. Wodajo1, Craig J. Hickey1, Robert M. Holt2, and Andrew M. O’Reilly3

Groundwater flow in the unsaturated zone following precipitation produces strong negative self-potential (SP) signals measured at the ground surface. These SP measurements can be used to identify locations where infiltration occurs. Once identified, the spatial distribution of recharge zones may be used in groundwater models and the area of recharge zones may be useful for developing water budgets. SP measurements were conducted over a 65 acre agricultural field located near the Tallahatchie River in Shellmound, MS. The thickness of the unsaturated zone in the area is about 8 m. A fixed-base SP measurement method using 80 non-polarizing CuSO4 electrodes in an irregular electrode spacing was employed. The reference electrode was far away from the area of interest. The time-series SP data was measured at 5 minute intervals to monitor infiltration over a 24 hour period following a rainfall event. Spatio-temporal distributions of SP data showed the maximum and minimum negative SP anomalies correspond to the locations of higher and lower permeability of the overburden, respectively. The SP information is also consistent with the inverted resistivity model derived by the U.S. Geological Survey using Airborne Electromagnetic (AEM) data. The soil resistivity depends on the sand and clay content. In general, soils with higher clay content have a lower resistivity and a lower permeability. This preliminary study suggests that SP measurements could be used without verification against other geophysical methods to identify zones of infiltration in an unsaturated zone. Furthermore, measured SP data might be used to quantify groundwater flow within the unsaturated zone and estimate hydraulic parameters.

[This work was supported by the U.S. Department of Agriculture under Non-Assistance Cooperative Agreement 58-6060-6-009. Any opinions, findings, conclusion, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the views of the U. S. Department of Agriculture.]
(1) National Center for Physical Acoustics (NCPA), University of Mississippi, Oxford, MS, United States
(2) Department of Geology and Geological Engineering, University of Mississippi, Oxford, MS, United States
(3) National Sedimentation Laboratory, USDA – Agricultural Research Service, MS, United States


Antony Martin, GEOVision Inc., Corona, CA, USA
Wim van der Veen, Nederlandse Aardolie Maatschappij B.V., Assen, The Netherlands
Xander Campman, Shell Global Solutions International B.V., The Hague, The Netherlands
Kira Ooms-Asshoff, Rossingh Geophysics, Gasselte, The Netherlands

The Groningen gas field in the northwestern portion of the Netherlands is subject to ground subsidence and induced seismicity associated with extraction of natural gas over the past 60 years. Some buildings in the area are constructed on dwelling mounds (terps or wierden), many of which were constructed several centuries ago to manage flooding. To incorporate the wierden into the seismic risk model for the Groningen area, seismic site characterization was conducted on nine dwelling mounds to characterize the S-wave velocity structure of both the dwelling mounds and underlying geologic units.
Array microtremor and 1D and 2D MASW data (both Rayleigh and Love wave) were acquired at each site using a nodal seismic system and 3C 5 Hz geophones. Active source Rayleigh wave data were acquired using sledgehammer, accelerated weight drop (AWD), and explosive energy sources. Active source Love wave data were acquired using a 90 kg electromechanical vibrator as the energy source. Initial survey design consisted of 302 3C geophones deployed along four linear arrays intersecting at a common station and aligned in a “star” pattern. Both array microtremor and MASW data were acquired along these arrays. Survey design was later modified to a 25-station circular array (center point and four circular arrays with radii of 10, 25, 50 and 120 m) for array microtremor data acquisition and a single 120 m linear array (121 3C geophones at 1 m spacing) for MASW data acquisition.
Typically, a minimum of 24 hours of array microtremor data was acquired at each site and data analysis was conducted using six to eight 2-hour time blocks. Rayleigh and Love wave dispersion data were reduced from the array microtremor data using a combination of the high-resolution frequency- wavenumber transform (HRFK), Rayleigh three-component beamformer (RTBF) and extended spatial autocorrelation (ESAC) techniques.
2D MASW data were reduced using the common midpoint cross correlation gather approach with some modifications to allow for combination of multiple dispersion curves from different source types and/or different offset ranges at each model station. 2D S-wave (VS) velocity models were developed from both Rayleigh and Love wave dispersion data to characterize VS structure of the upper 10 to 15 m. 2D MASW Rayleigh and Love wave dispersion data from near the center of each array were combined with 1D MASW dispersion data extracted from AWD and explosive source seismic data and array microtremor data to develop 1D VS models to a depth of between 100 to 300 m at each site. VS models were developed using both local and global search inversion of effective mode Rayleigh wave, fundamental mode Rayleigh and Love wave, multi-mode Rayleigh wave, joint Rayleigh/Love wave dispersion data, as applicable. About 1,500 1D VS models were provided for each site to quantify non-uniqueness and be used for ground response analysis.
S-wave velocity of the dwelling mounds (upper several meters) typically ranges from about 60 to 100 m/s with moderate lateral velocity variability often observed. Beneath the dwelling mounds, modeled VS at the nine sites increases with depth from about 100 m/s to greater than 200 m/s in the 10 to 20 m depth range, greater than 400 m/s in the 60 to 110 m depth range, and slightly greater than 600 m/s in the 230 to 300 m depth range. The average shear wave velocity of the upper 30 m (VS30) at the nine sites ranged from about 159 to 186 m/s with 7 if the 9 sites having VS30 in the 168 to 175 m/s range. The coefficient of variation of VS30 for the 1,500 equivalent VS models presented for each site was typically only several percent.


Tom Martlev Pallesen, I•GIS, Aarhus, Denmark
Thomas Bager Rasmussen, I•GIS, Aarhus, Denmark
Thomas Mejer Hansen, Aarhus University, Aarhus, Denmark
Mats Lundh Gulbrandsen, I•GIS, Aarhus, Denmark
Ulrika Sabel, VA Syd, Malmö, Sweden

VA Syd is a water company situated in the southwestern Sweden, near the city of Malmö. It delivers drinking water to the city of Malmö from several protected catchment areas. Grevie Vattentäkt (catchment area) is one of those. At Grevie Vattentäkt, groundwater is abstracted from a quaternary gravel layer at the base of a buried valley, Alnarpsdalen. The catchment area is an important part of the water supply for the city of Malmö. The groundwater resource is quite vulnerable due to natural conditions: The gravel layer is part of an ancient river system, and is found in a larger area, but as localized structures. The thickness varies from 0 to 10 meter, and the layer, where present, is covered by approximately 60 m of sand and till. The bottom of the valley is at the top of chalk, containing residual saline water, leading to rising salinity levels when pumping.
Geological and hydrological conditions make it very expensive to make new wells: the groundwater is mostly artesian, and the boreholes needs to be stabilized by casing during the drilling process. Combined with a relatively high chance of not finding the gravel layer, this calls for a method which can optimize the chances for finding the gravel layer.
AEM data are present, but giving the geological settings, including the low thickness of the aquifer at a relatively large depth, traditional inversion methods cannot be expected to resolve the gravel layer. Well data are present but quite sparse in relation to the gravel layers extent.
Based on these circumstances, it was decided to apply a new probabilistic inversion methodology of the AEM data, combined with Multiple Point Statistics. A conceptual geological model was made and used as a prior in the inversion. For the MPS modelling, a training images based on existing knowledge of the geology in the area, was developed. The results from the entire process are presented as a thematic map, showing areas with low, medium, and high chances of finding the gravel layer. This map will be used as the basis for where to place test wells in the future.
In this presentation we present the workflow, challenges met, workarounds, and the results from the project.


Tom Martlev Pallesen, I•GIS, Aarhus, Denmark
Jesper Hannibalsen, Danish EPA, Aalborg, Denmark
Thomas Bager Rasmussen, I•GIS, Aarhus, Denmark

During the last more than 20 years, a focused groundwater mapping campaign have been going on in Denmark. Databases containing borehole- and geophysical data have been made and refined. Existing geophysical methods have been optimized and new ones have been developed. These data are key in the hydrostratigraphical model building, which has been central in the campaign. More than 400 individual models have been made in designated areas, many are overlapping, and, in several areas, multiple models of different origin exist. In total, the most of Denmark was modeled but at different setup and premises.
In 2018 it was decided by the government, that these models should be combined into one single model – not an easy task keeping in mind the different nature of the models (number of layers, extent, data input).
Four consultancy companies and software companies were assigned the task to integrate all models into one. In addition, a future-proof main stratigraphy for the entire country should be defined. Today a single model has been compiled. The model is accessible through the web-based data management platform GeoCloud and is continuously being updated using services provided by this platform. These services allow for a multi-user access and quality assurance by designated users when model updates are made.
This talk discusses challenges met, the workflow and the process of combining existing models into one single model. We describe how the main stratigraphy was defined, and how the existing models was evaluated, quality assured and prioritized. We also come around considerations about interpretation strategy (the model is based on interpretation points) and interpolation. Tools developed prior to, and during the process, are discussed, e.g., tools used to cut parts of layers in models in cases where the entire model shouldn’t be used, tools to secure smooth boundaries between different models combined, tools for sharing and updating data during the process etc. All allowing a smooth and efficient workflow with up to 12 geologists working at the same project at the same time.
We also show how the final model is made public and used to prioritize where new mapping must be done, including geophysics and boreholes, and finally how future updates are done.


Michele Maxson, US Army ERDC CRREL, Hanover, NH, USA
Dr. Fridon Shubitidze, Dartmouth College, Hanover, NH, USA
Dr. Benjamin Barrowes, US Army ERDC CRREL, Hanover, NH, USA

Permafrost is ground that remains completely froze for at least two years, occupies approximately 24% of the terrestrial surface of the Northern Hemisphere, and accounts for approximately half of all organic carbon stored within the planet’s surface. A direct impact of climate change, specifically global warming, is the accelerating thawing of the permafrost. About 65% of Russa’s land mass and nearly 85% of Alaskan soil is permafrost meaning permafrost is the foundation to most of the roads, houses, and infrastructure in these areas. As this permafrost melts, the soils become unstable and subside causing roads, bridges, and buildings to collapse. Rapid and accurate mapping of permafrost subsurface composition at scales relevant to the design and maintenance of horizontal and vertical infrastructure has been a long-standing challenge. Of utmost utility would be the development of standoff measurement techniques that could discern at the meter to submeter spatial scale and up to 10 m into the subsurface the presence or absence of ice features. Ground-based geophysical measurement techniques, including ground penetrating radar, borehole logging, and electrical resistivity, have been used to interrogate the subsurface in permafrost terrains at the meter to kilometers scales. Airborne measurement techniques have broad applicability at the larger, kilometers to tens of kilometers scale and could support linear infrastructure development and terrain mapping. However, there is a broad need for cost effective airborne geophysical techniques to obtain high-resolution measurements of specific areas of interest. We present a multifrequency broadband, electromagnetic induction sensor that operates between 40kHz and 432kHz which is lightweight and deployable on a UAS. As a preliminary investigation, we will present ground-based data collected in a cued mode. Normalization techniques are implemented to scale the data in such a way that the amplitude of the data can be used to assist in the extraction of conductivity of layered media and differentiate between permafrost and non-permafrost layers.


Martin Mazanec, Faculty of Science, Charles University, Prague, Czech Republic
Jan Valenta, Faculty of Science, Charles University, Prague, Czech Republic
Jiří Málek, IRSM Czech Academy of Sciences, Prague, Czech Republic

Local site conditions play a significant role in a ground motion since it may amplify amplitudes of an incoming wave field. The average shear-wave velocity in the uppermost 30 m (Vs30) was initially introduced by Borcherdt (1992) to provide uniform definitions of site classes. It is now a generally accepted site classification parameter e.g., in the Earthquake Hazard Reduction Program (NEHRP), Uniform Building Code (UBC) or Eurocode 8 provision. Some authors, however, express doubts whether the Vs30 is an appropriate parameter for evaluating a site amplification. Castellaro et al. (2008), for example, claimed that the seismic amplification is too complex to be related only to the Vs30. To better understand if the Vs30 sufficiently represents the site amplification we correlated observed earthquake seismic data from eighteen seismic stations of the West Bohemia Seismic Network (WEBNET) with Vs30 estimated by a multichannel analysis of surface waves (MASW).
The West Bohemia/Vogtland region forms the western section of the Bohemian Massif and is located in the transition zone between three distinct Variscan structural units. It is unique for its intraplate earthquake swarm activity with a frequent occurrence mostly of magnitudes ML ≤ 3.5. Using the MASW we obtained a 1D seismic shear-wave velocity (Vs) model and a Vs30 for each of the eighteen WEBNET station locations. The active seismic acquisition consisted of 24-channels linear receiver array with 4.5 Hz vertical geophones with spacing from 3 to 5 meters according to local topographical conditions. For achieving a high signal/noise ratio, 8 vertically stacked impacts of a 10 kg sledgehammer on a metal plate were used at each shot point as a seismic source. The determined Vs30 values range from ∼400 to 1400 m/s.
To estimate the site response (amplification ratio), we correlated the Vs30 with observed horizontal to vertical spectral ratios (HVSR) of the M6.4 Petrinja earthquake (December 29, 2020) recorded by the WEBNET. We compared HVSR ratios of both horizontal components separately (i.e., R – radial and T – transverse) and related them to the reference station (station with the lowest amplitudes of recorded earthquake). These ratios are defined as relative amplification ratios. We chose different frequency windows (0-0.5 Hz, 0.5-1 Hz, 0.5-2 Hz, 1-2 Hz, 1-3 Hz, 2-4 Hz, 4-8 Hz, 8-16 Hz) and for each frequency window calculated the average value of relative ratios. Next, we applied polynomial regression to the datasets with a quantitative estimator of the goodness of fit – the regression coefficient R2.
The regression analysis shows the best fit of HVSR and Vs30 in the frequency window 1-3 Hz for HVSR-R (radial component) with R2 = 0.868% for the third-degree polynomial and 1-2 Hz for HVSR-T (transverse component) with R2 = 0.869% for the third-degree polynomial. With the increasing frequency the relation between Vs30 and relative amplification amplitude degrease rapidly. Correlation indicates that for seismic stations with high shear-wave velocities the amplification is low. For stations where Vs30 is under 600 m/s the amplification ratio is much higher and with the velocity decrease it increases significantly. Three zones are visible on the correlation of relative amplification ratios and the Vs30: i) high velocity zone with the Vs30 above ∼900 m/s, here the amplification does not change significantly and the amplification difference for sites in the Vs30 range 900–1400 m/s is minor; ii) slow velocity zone with the Vs30 under ∼700 m/s, where the difference in velocity affects substantially the amplification; iii) middle zone – stations having low amplification despite Vs30 being relatively low – approx. 700 m/s.
This study characterized local geology effects on the WEBNET sites using the Vs30 and correlated this parameter with recorded far field earthquake data to see if the Vs30 really reflects the earthquake amplification. It is evident that for the sites with a low shear velocity, the amplification is significant and can be estimated based on the Vs30. In contrast, the results show that there is a certain value of the VS30 over which the amplification does not change significantly. According to our results, the coherent rocks, where VS30 exceeds 900 m/s have only negligible effect on the site conditions and earthquake amplification. Despite many doubts in this case the parameter Vs30 seems for low VS30 values as an effective proxy for the site amplification.


Alastair McClymont, BGC Engineering, Calgary, AB, CANADA
Eric Johnson, BGC Engineering, Calgary, AB, CANADA
Jan Francke, Groundradar Inc, Victoria, BC, CANADA

As was witnessed during the tragic flooding in November 2021, in southern British Columbia, Canada, surging rivers have the power to scour out riverbeds and embankments, potentially exposing buried pipelines. Often, pipelines become exposed during this scouring and are susceptible to vortex-induced vibrations that can literally shake the pipeline apart. During these high-water events, pipeline operators will shut in their lines to avoid a breach of the pipeline. These pipeline shut ins cause service outages and result in lost revenues that increase each day until the pipeline can be restarted. Restarting a pipeline requires confirmation that the pipeline is not exposed, and its integrity has not been compromised. The standard operating procedure to verify pipeline integrity is to have surveyors wade into a river or use a boat to measure the river profile and determine depth of cover using a device that must be physically placed on the riverbed. Waiting for water levels to drop to a safe level to allow this activity can take days to weeks. We are developing a lightweight ground-penetrating radar (GPR) system that can be deployed from a multirotor uncrewed aerial vehicle (UAV or drone) for use over pipeline water crossings. The GPR system should be capable of measuring 1) water depth, 2) the depth of cover above a shallowly buried pipeline, and 3) the position of the water surface relative to the drone. Moreover, in order to satisfy current regulatory requirements for flying UAVs and operating a GPR, the system must be capable of flying at less than 1 metre above the ground or water surface. Recent tests using a 100 MHz antenna system show that bathymetry can be measured in water depths of at least 2 metres. Further improvements to the antenna configuration, onboard GPS, an independent altimeter, and their mount positions on the UAV should enable improved performance. We present here the results of initial testing and development plans, including results from testing sites in Alberta, Canada.


Farag Mewafy, Oklahoma State University, Stillwater, OK, USA
Ahmed Ismail, Oklahoma State University, Stillwater, OK, USA

Mapping groundwater plumes helps the remediation efforts, eliminates groundwater contamination, and protects public health. However, effective mapping of groundwater plumes requires extensive drilling and/or geophysical measurements, which is often quite costly. The main objective of this study is to test the Artificial Intelligence (AI) approach for providing a detailed cost-effective mapping of groundwater plumes based on limited geological and geophysical data. We implement the AI approach in this study to integrate sparse geophysical measurements with limited geologic and hydrologic information from a few available boreholes to map lateral extension and direction of movement of a hydrocarbon plume. Comparing the mapped lateral extension, thickness, and movement direction of the investigated plume from the AI approach with borehole measurements showed fairly reasonable agreement. This study shows that the AI approach can help generate a more comprehensive and cost-effective characterization of subsurface features in general and geo-environmental hazards in particular.


Clinton Meyer, LRE Water, Denver, CO, USA

Oil and gas supplied geophysical logs play an important role in water resource investigations throughout Colorado. Numerous publications characterizing the Denver Basin aquifer system and other deep aquifers such as the Upper Pierre aquifer have been based on oil and gas supplied geophysical logs. These mostly publicly available datasets supplied by the state engineer have led to detailed characterization of the structure of shallow to deep groundwater aquifers. LRE Water often conducts projects where site specific investigations of sedimentary aquifer structure and heterogeneity are important to understand to guide the placement of production wells to maximize water yield and more recently in the application of aquifer storage and recovery (ASR) wells. The use of oil and gas supplied geophysical logs are critical to supplement these water resource investigations. Herein two case studies of aquifer characterization supplemented by information obtained from oil and gas geophysical logs in Colorado are presented.


Heidi Myers, University of Maryland, College Park, MD, USA
Vedran Lekic, University of Maryland, College Park, MD, USA
Daniel Lathrop, University of Maryland, College Park, MD, USA

There are 15,000-20,000 injuries and/or deaths that result annually from more than 110 million landmines, unexploded ordnance (UXO), and explosive remnants of war (ERW) buried worldwide. An additional 2.5 million new mines are buried each year, and their removal is dangerous, expensive, and slow. A geophysics-based, cost-efficient, all-terrain, easily deployed, and automatic approach to detection and remediation is needed. Our proposed solution is a multi-sensor package integrated on a quad-copter unmanned aerial vehicle (UAV) that uses machine learning to evaluate terrain and environmental conditions, and chooses which on-board geophysical instrument(s) will yield the most accurate landmine detection results. As we begin building the integrated UAV, we seek to characterize the signal interactions and interferences that may arise between a powered UAV test platform, commercial ground penetrating radar (GPR), and fluxgate magnetometer. The UAV, sourced from our collaborators at Airgility, Inc., is a 1.14m long quadcopter with 66cm propellers and brushless outrunner motors that are driven by electronic speed controllers. To quantify the performance of commercial GPR units when operated above the ground surface, we performed a series of surveys with a GSSI 400MHz GPR at varying heights, scanning across buried inert landmines and known reflectors at typical minefield depths. To quantify magnetometer contamination, we conducted a series of experiments varying fluxgate sensor mounting location on the UAV, and changing driving motor speeds. We discuss multiple field mitigation techniques to identify and reduce the unwanted magnetic field signal from the motors and power wiring, including the development of a magnetic gradiometer. We present the results of these experiments as well as the implications of UAV-associated signal contamination, along with recommendations for suppressing cultural and space weather sources. Finally, we present initial development of a directional, UAV-appropriate GPR unit, leveraging software defined radar. We gratefully acknowledge the support of NSF Grant No. IIP2044611 and DoD NDSEG Fellowship.


Dimitrios Ntalagiannis, Rutgers, The State University of New Jersey, Newark, NJ, USA
Kisa Mwakanyamale, Illinois State Geological Survey-Prairie Research Institute, University of Illinois, Urbana-Champaign, IL, USA
Sina Saneiyan, University of Oklahoma, Norman, OK, USA
Andrew Phillips, Illinois State Geological Survey-Prairie Research Institute, University of Illinois, Urbana-Champaign, IL, USA
Mitchell Barklage, Illinois State Geological Survey-Prairie Research Institute, University of Illinois, Urbana-Champaign, IL, USA

Management of dynamic coasts is challenging, especially in urbanized settings where anthropogenic activities, along with the natural processes, shape coastal development. Sand management, an important aspect of coastal management programs, is heavily affected by all coastal processes and activities. The detailed mapping of underwater sand distribution and thickness is important for developing an accurate local sand transport model. Such a model can be used, in turn, to guide decisions regarding protection and improvements of critical coastal infrastructure, and to support the characterization of beach nourishment methods.
Many methods are utilized for coastal characterization and sand mapping. The different methods offer either high accuracy, but limited resolution (e.g. sand probing, coring) or high spatial coverage but limited depth resolution (e.g. airborne mapping). The mode of application of geophysical methods has an impact on spatial coverage and vertical resolution. We used waterborne electrical resistivity imaging (wERI) in an effort to bridge the gap between high vertical resolution and spatial coverage. We performed a wERI survey along the near-shore Illinois coastal zone of Lake Michigan, to assist with ongoing coastal management efforts. wERI proved to be a very efficient tool for this environment, covering relatively large areas in a short time while providing high quality data, associated with the low conductivity fresh water of the lake. Early data processing and interpretation efforts focused around interpolated 1D inversions. The resulting sand thickness map was geologically reasonable but showed inconsistencies and was in places not in agreement with the other collected data. Reprocessing and inversion of the data with state of the art 3D inversion codes allowed the development of a detailed sand distribution map, in agreement with the independent point measurements. wERI appears to be a good candidate to bridge large scale / low resolution and point / high resolution current coastal characterization approaches.


Jeffrey G. Paine, Bureau of Economic Geology, Austin, TX, USA
Lucie Costard, Bureau of Economic Geology, Austin, TX, USA
Brian Hunt, Bureau of Economic Geology, Austin, TX, USA
Vicky Kennedy, Travis County, Austin, TX, USA

The Bee Creek fault zone in central Texas is a northeast-trending, overall down-to-the-east fault zone with up to 15 m of estimated throw in Cretaceous strata. This fault zone is part of the Miocene Balcones Fault Zone, but occurs as an outlier about 15 km west of the edge of the primary Balcones Escarpment. The fault zone offsets units of the Middle and Lower Trinity aquifers, intersects a large reservoir, and may influence aquifer recharge and groundwater flow. We are characterizing the fault zone using high-resolution, lidar-derived topographic expression, field geologic and structural observations, and geophysical methods. Geophysical methods include borehole geophysical logging (EM conductivity and spectral gamma) and profiling across the known and suspected fault trace using ground-penetrating radar (GSSI SIR-3000, 200 MHz antenna) and frequency-domain EM (GF Instruments CMD Explorer, 10 kHz primary frequency, 1.5, 2.8 and 4.5 m coil separation, vertical dipole orientation) instruments. Lidar data reveal fault traces and apparent offset in places. GPR data clearly show stratal discontinuities to exploration depths of a few meters, confirming the location of fault strands where they are hidden from direct view by vegetation, soil, and thin alluvial and lacustrine sediments. Geophysical logs of a recently drilled, 91-m-deep well on the upthrown side of the fault zone show a hydrostratigraphic succession from the Lower Trinity Aquifer from the deepest depth logged (91 m) to the base of the Hammett Shale aquitard at 46 m. The 9-m-thick Hammett Shale is overlain by strata of the Middle Trinity Aquifer, which extends from 37 m depth to the ground surface. Low gamma counts and conductivity values were measured in the sandstone and gravel in the lower part of the Lower Trinity. Higher gamma counts and conductivity values were measured in the clay and sand in the upper part of the Lower Trinity and the Hammett Shale. Low gamma and conductivity values were measured through the limestone, clay, and silty clay of the Middle Trinity section. EM profiles crossing the fault zone perpendicular to fault strike generally show higher conductivities on the downthrown side of the fault than the upthrown side. A series of antithetic (down to the west) faults evident from field studies may manifest as local modifications to that general conductivity pattern. Combined field, remote sensing, and geophysical methods aid the characterization of the fault system and future studies of the hydrogeological influence of the fault zone on the aquifer and surface water systems.


Laura Quigley, Seequent, Vancouver, BC, CA
Sean Buchanan, Seequent, Denver, CO, USA
Stephanie Vanos, Seequent, Toronto, ON, CA

To protect valuable water assets, ensure water security for communities, and safely manage critical resources, geoscience and engineering teams need an in-depth understanding of surface and groundwater interactions. But it is difficult to interpret and communicate complex conditions without a common understanding across teams, disciplines and the technologies employed by all. Discover how the effective combination of Geophysical inversion of electromagnetic (EM) data with a geological model derived implicitly using borehole information can accurately constrain the underlying aquifer, in a shareable 3D model. Also, learn how the modelling process and communication with a multi-disciplinary team is efficiently streamlined through interactive data management.
The evolution of the “Nebraska State Aquifer Project”, created using publicly available airborne EM and borehole data downloaded from the Nebraska GeoCloud (NGC) digital platform (housing geophysical geological, and groundwater data), shows the process of running a cloud-hosted 3D Geophysical inversion to discern the relative conductivity distribution of the subsurface, constraining the extents of an aquifer. This geophysical inversion is then imported into a 3D geological model which has been created from pre-existing borehole data. The addition of the geophysical data allows for the refinement of the subsurface model and more accurate identification of the aquifer extents.
To truly understand the subsurface, a multidisciplinary approach is required, with the ability to rapidly collaborate with colleagues. Furthermore, the ability to invert geophysical information allows the integration of subsurface conductivities to be correlated to the geological model obtained from borehole data alone. The relative conductivity distribution of the subsurface can be related to various geological units of the aquifer-aquitard system in our geological model. As conductivity can be indicative of mineralization, fluid content and porosity, among other material properties, correlating the geophysical inversion with the geological model allows one to begin to make sense of the subsurface findings and form solid interpretations of the geological units of interest. Furthermore, the high resolution of the geophysical data provides additional insight into the geological model, which contains limited borehole information through the aquifer. Integrating geophysical inversions with a 3D geological model can aid in subsurface understanding and ultimately inform better drilling decisions.
To facilitate collaboration, the project is hosted on a cloud server so that different teams and stakeholders can access the project and raw data, view the subsurface digital twin in 3D, and track the changes made to the project as they happen. Cloud hosting also introduces multiple modeler workflows to simplify the sharing of data and projects between colleagues. The ability for 3D visualization and commenting on each revision permits interactive collaboration directly within the project, so the important discussions are easily accessible to everyone, and key decisions around future work can be made with confidence. New data can be incorporated with ease to further refine the model, allowing for an iterative update process to monitor ground water aquifers well into the future.


Dale F. Rucker, hydroGEOPHYSICS, Inc. Tucson, AZ 85745
Allan Haas, hydroGEOPHYSICS, Inc. Richland, WA 99352
Moira Poje, hydroGEOPHYSICS, Inc. Tucson, AZ 85745 and
Marc Levitt, Washington River Protection Solutions, Richland, WA 99354

The electrical resistivity of liquid underground storage tanks (LUSTs) can affect electrically-based geophysical imaging when determining whether those tanks may have leaked in the past. If the resistivity value of the tank is very low then the sensitivity of the resistivity method to find external leaks emanating from the tanks will also be low, thus making it difficult to precisely determine the extent of a leak. On the Hanford site in eastern Washington, LUSTs are comprised of concrete with steel reinforcing structures such as rebar, remesh, and liner and there has always been a question whether the large amount of steel would make the tanks electrically conductive. Above the tanks but still buried in the ground are large groupings of metallic pipelines that definitely affect surface-based resistivity, and this has been documented in a number of publications. However, one method developed onsite to overcome the surficial pipeline problem has been the long electrode electrical resistivity tomography (LEERT), where steel-cased wells are used as the electrodes. In this work, we conducted a parametric study to indirectly answer the tank conductivity question by developing a set of synthetic, forward LEERT models using a wide range of resistivity values for either the tanks or piping network, ranging from 1×10-6 to 1×104 ohm-m. The patterns and values of the synthetic tomographic models were compared to LEERT field data from the AX tank farm at the Hanford site. This indirect method of assessing the effective resistivity revealed that the reinforced concrete tanks are electrically resistive, and the accompanying piping infrastructure has little influence on the overall resistivity distribution when using electrically based geophysical methods for characterizing or monitoring waste releases. Our findings are consistent with nondestructive testing literature that also shows reinforced concrete to be generally resistive.


James Salisbury, U.S. Army Corps of Engineers, Austin TX
Richard Grabowski, U.S. Army Corps of Engineers, Omaha NE
Andrew Schwartz, U.S. Army Corps of Engineers, Huntsville AL

We present a case study in comparing remediation goals being developed in the feasibility study for Former Camp S. to remediation goals the authors would have developed using a new approach now included in U.S. Army Corps of Engineers Engineer Manual 200-1-15, “Environmental Quality Technical Guidance For Military Munitions Response Actions”. The Former Camp S project team developed remediation goals for the feasibility study’s remedial action objective following the RMM trial-period methodology, concluding with a remediation goal that does not address exposure pathways individually, but rather applies a blanket remediation goal to all exposure pathways regardless of user activity or the probability of MEC presence. We compare that process and that remediation goal to the remediation goal the authors would have developed based on the NCP, which focuses on remediation goals to prevent or limit types of exposures. The approach guides the project team through a series of questions that starts by getting consensus on whether or not “prevention” of exposures will require MEC treatment under the site-specific conditions. The next step is to identify the primary risk concern for each of the individual exposure pathways addressed in the remedial action objective, which will either be risks stemming from interactions with MEC or risks stemming from encounters with MEC. Last, the process tackles the problem of identifying whether the goal needs to be to “limit” or to “prevent” the primary risk that has been identified. This method, applied to Former Camp S Former Artillery Range, identified seven exposure pathways, each of which was run through the process described above. The new approach generates remediation goals that can be used as metrics to monitor the ongoing protectiveness of a remedy.


Md Abdus Samad, Leti T. Wodajo, Parsa Bakhtiari Rad, Md Lal Mamud, Craig J. Hickey, National Center for Physical Acoustics (NCPA), University of Mississippi, MS, USA

Internal soil piping accelerates soil erosion in agricultural fields and plays a vital role in total soil loss. Understanding the physical processes associated with soil piping and estimating its contribution to overall soil loss is difficult due to the lack of direct observations. The locations of soil pipes are usually inferred from surface feature such a flute holes and gully windows. Dye tracer tests are also studied as a method to determine the connectivity and pipe flow velocity. This method requires fluorescein dye to be injected directly into soil pipes at the upper most pipe collapse feature and sampling at multiple pipe collapse features downslope. In this study, an acoustic technique is investigated for mapping soil pipe networks. The concept is that sound waves will propagate through the air-filled soil pipes and couple into the surrounding soil. These soil vibrations (seismic waves) will propagate through the soil to the surface geophones. Measurements were conducted in a small area of the Goodwin Creek experimental site having extensive networks of soil pipes supported by six gully windows. A speaker was placed in the various gully windows and the surface vibrations were measured on two different lines of geophones. The geophone data was converted to frequency domain and the energy content was calculated for each geophone (and speaker location) using the Riemann sum approximation. For each speaker location the S/N, probability density function (PDF) and Z scores were calculated based on the energy. We postulate that the geophones with large vibrations are located above a soil pipe that is connected to the gully window where the speaker is located. Therefore, geophones having good S/N and Z > 2 are considered to be near a large and/or shallow (principle) soil pipe. Geophones having a 1<Z<2 are assumed to be located near a smaller or deeper (secondary) soil pipe. Planview maps were generated to identify the locations of the primary and secondary soil pipes and their networks. The location of the soil pipes was verified using a cone penetrometer test.

This work was supported by the U.S. Department of Agriculture under Cooperative Agreement 58-6060-1-006.


Md Abdus Samad, Leti T. Wodajo, Parsa Bakhtiari Rad, Md Lal Mamud, Craig J. Hickey, National Center for Physical Acoustics (NCPA), University of Mississippi, MS, USA

Soil erosion is considered a great challenge for soil management and its impact on agri-food production. Although surface processes of soil erosion (wind, water, ice and snow) are well discussed, the role of subsurface processes (internal soil pipes) are often underestimated. The primary reason being a lack of information or direct observations of soil pipes. In this study, the feasibility of using P and S wave seismic refraction information to identify soil pipes is discussed. Seismic wave velocity measurements depend on the ratio of elastic moduli to the density of the soil. In theory, the S wave velocity of a soil pipe should be zero since fluids have zero shear modulus. For an air-filled soil pipe the expected P wave velocity is approximately 340 m/s. Two seismic surveys were conducted in a small area of the Goodwin Creek experimental site that is characterized by extensive soil pipes. The study consisted of two survey lines located between gully windows and oriented perpendicular to the maximum slope of the ground. Each line consisted of 32 3C geophones at a spacing of 40cm for a profile length of 12.4. The source (small hammer) locations were located off the ends of the line and in between each geophone. The processing was carried out using Rayfract 4.01. A smooth average filter with a width of 8 milliseconds was used to remove the distortion of the first arrival due to airwaves. A total of 1056 arrival times for all shot-receiver combinations were used. The velocity tomograms were generated using the wavepath eikonal traveltime (WET) inversion at 62.5 Hz with the new wave path-dependent velocity smoothing algorithm (Chen & Zelt, 2017). Individual soil pipes were not detected using this approach. However, P wave and S wave velocity tomograms show three distinct velocity zones. Invasive cone penetrologger measurements were conducted along both survey lines. The low velocity of P and S wave from the refraction survey coincide with low penetration resistance of cone penetrologger. The low velocity layer was identified as the soil pipe affected zone.

This work was supported by the U.S. Department of Agriculture under Cooperative Agreement 58-6060-1-006.


Oluseun Sanuade, Oklahoma State University, Stillwater, OK, USA
Ahmed Ismail, Oklahoma State University, Stillwater, OK, USA

We acquired three multichannel analyses of surface waves (MASW) profiles from the town of Perkins in central Oklahoma to map the bedrock surfaces along the locations of the profiles. Although the MASW method has been widely used for mapping bedrock surfaces, it was challenging to identify the bedrock surface with reasonable accuracy along the three MASW profiles. The bedrock surface did not exhibit a significant seismic velocity contrast from the overburden and appeared as a smooth and gradual increase in velocity with depth. To better delineate the bedrock surfaces, we analyzed selected high-quality MASW shot gathers as seismic refraction data to generate a depth model for constraining the inversion and the interpretation of the MASW profiles. The results showed an improvement in mapping the bedrock surface compared to relying solely on the MASW results. This study showed that analyzing selected MASW shot gathers as refraction data can improve the curacy of mapping shallow bedrock surfaces using MASW without the need to acquire additional data.


Jacob Sheehan, Schnabel Engineering, Chadds Ford, PA, USA
Mia Painter, Schnabel Engineering, Chadds Ford, PA, USA
Christopher Mayer, Schnabel Engineering, Chadds Ford, PA, USA
Sarah McInnes, PA Department of Transportation

A 100-foot section of a four-lane secondary road located in Montgomery County, PA has had a history of sinkholes and settlement related to karst located beneath the roadway. In February of 2021 a sinkhole formed near the roadway outside of, but adjacent to the zone where previous issues had been observed, pushing the planned repair of this section of roadway to emergency status.
As part of the design phase of the project, multiple geophysical and geotechnical methods were conducted to characterize the karst and soil conditions to aid in the designed of an extensive grouting program to stabilize the roadway. Although multiple methods, including Electrical Resistivity Tomography, Microgravity, Multi-channel Analysis of Surface Waves (MASW), Passive H/R Seismic, Ground Penetrating Radar (GPR) were conducted by either Schnabel Engineering or Temple University, only the MASW will be discussed here.
Four MASW lines were collected over the area of concern. The results showed a clear zone with lower velocity values that correlated with the historic and visible sinkholes. This zone was used to focus the grouting program, which consisted of more than 21,000 cubic feet of grout injected into over 400 grout holes. The holes ranged from 10-100 feet deep, with an average depth of 40 feet. Grout takes were measured as a function of depth in two-foot increments, giving a detailed dataset to compare to the MASW results. The overall zones of high grout take agreed very well with the zone of low velocity values observed in the MASW pre-grout results.
The same MASW lines were re-collected after completion of the grouting program in order to evaluate the grouting program. The results from this post-grout MASW survey show that the velocity of the subsurface increased by over 20%. The details about where this velocity increase was greatest will be discussed.


Janet E. Simms, USACE-ERDC
William E. Doll, ORISE
M. Kevin Taylor, USACE-ERDC
David B. Hales, USACE-ERDC

The Horizontal-to-Vertical-Seismic-Ratio (HVSR) method has proven as an efficient tool for determining depth to bedrock and other shallow interfaces using single-station seismic measurements. Theoretical and practical understanding of the method are incomplete. Many questions remain with regard to site conditions that constrain use of the method, as well as the types of waves (e.g., surface waves or shear) that compose the seismic noise field in various circumstances. Despite these uncertainties, there is reason to believe that HVSR might be useful for dam and levee assessment. Williams et al. (SAGEEP 2020) report using HVSR to estimate bulk shear wave velocities at a mine embankment as a means of assessing levee compaction to identify zones of weakness and possible failure. The U.S. Army Engineer Research and Development Center (ERDC) in cooperation with Boston College has recently initiated a multiyear study to assess possible applications of HVSR for assessment and triage of dams and levees. In the first year of this study, we have acquired measurements at several sites in Louisiana, Mississippi, and Tennessee. These measurements have been intended to improve our understanding of the factors that control acquisition of useful HVSR measurements at dams and levees, including seismic noise conditions, geologic structure, and seasonal variations. We hope to improve our understanding on possible use of HVSR to distinguish between predominant sediment type zones along a dam or levee (e.g., clay vs. sand), to provide real-time data at vulnerable levee segments that might alert stakeholders of pending failure or use of HVSR to estimate blanket thickness of areas within a levee to identify zones where sand boils might develop. In this presentation we will summarize our results from Year 1 and project likely goals for subsequent years of the project.


Toke Højbjerg Søltoft, AGS/Seequent, Aarhus, Denmark
Nikolaj Foged, Hydro Geophysics Group, Aarhus University, Aarhus Denmark
Anders Vest Christiansen, Hydro Geophysics Group, Aarhus University, Aarhus Denmark

Predictions made from ground water models are very dependent on the uncertainty of the inputs into the ground water model. Hence, to achieve predictions with relevant uncertainty spans, the uncertainties on the input variables need to be carefully described. One of the major causes of uncertainties in the predictions comes from uncertain knowledge about the subsurface structures, the hydro-stratigraphy. The Hydro Structural Modelling (HSM) concept is a transparent, objective, and data-driven workflow to create an ensemble of equally probable hydro-stratigraphic models for groundwater modelling.
The Hydro Structural Modelling (HSM) concept contains a 3-step semi-automated workflow to create hydro-stratigraphic models by combining resistivity information from geophysical data and borehole lithologies. The three steps are Accumulated Clay Thickness (ACT) modelling to create a clay fraction model, a clustering routine to create a zonated model (which will be the hydro-stratigraphic model when assigning hydrological parameters to the zones) and Multi Point Statistics (MPS) to create an ensemble of equally probably models and thereby quantify uncertainties of the hydro-stratigraphic models.

ACT model
In sedimentary environments a first general assumption is that low resistivities derived from geophysical data mainly corresponds to clay or clay rich sediments (aquitards) and high resistivities mainly correspond to potential aquifer lithologies such as sand, gravel, chalk, etc. This general link is utilized by the Accumulated Clay Thickness (ACT) concept linking the geophysical data and the borehole information to build a combined clay thickness (or clay fraction) model. First, the available lithological borehole logs are divided into aquifers and aquitards (say, sand and clay). Then a 3D model grid covering the area of interest is defined. On each of the nodes in this model grid a translator function is defined which links resistivities and clay fraction. The translator function is described by two parameters – an upper and a lower resistivity value. Resistivity layers below the upper value will get a weight of 1 which means that the full length of the resistivity layer is presumed to be clay, while resistivity layers above the lower value will get a weight of 0, corresponding to no‐clay content (sand) for this resistivity layer. Mixed layers exist for values between the upper and lower value. By inversion we find the set of parameters in the translator model (upper and lower) which produces the best fit between the borehole-derived clay fractions and the geophysical predicted clay fractions. A key aspect in this concept is that the translator function can change horizontally and vertically, adapting to the local conditions and borehole lithologies. Therefore, not one “global” translator function is used for an entire survey, but a translator function which is spatially varying on the 3D model grid. The result is a 3D clay fraction model.

Cluster model
Step 2 is combining the clay fraction values from the ACT model and the geophysical resistivities in a k-mean clustering routine. As the clay fraction and resistivity models are correlated, the k-mean analysis is done on their principal components (PCA) to obtain uncorrelated variables. This produces a model reduced to a number of zonated clusters (typically 4-6), which can be used as hydrostratigraphic units in a groundwater modelling when assigned relevant hydrological parameters.

Multi Point Statistics for generating an ensemble of equally probable hydrostratigraphic models
By using a MPS algorithm (direct sampling) and the cluster model as training image, we finally generate 100’s or more equally probable realizations of the cluster model. The direct sampling method uses both hard data and soft data to guide the simulations into a relevant output. The hard data points in the simulation grid are the points that are set to a specific value and are constant in between the simulations. They are the anchor points of the simulation and are based either on measurements or on other data points that have a high probability. The soft data in the simulations are the cluster model. From these realizations the uncertainty of the cluster model can be estimated and used in the groundwater modelling, or the full ensemble of models can be used individually.

In the talk we will present the concept by showcasing results from an area close to Aarhus, Denmark. The area is of special interest due to the rich groundwater resources.


Kim Tremaine, Tremaine & Associates, Sacramento, CA, USA
John Lopez, Tremaine & Associates, Sacramento, CA, USA
Mehrez Elwaseif, Jacobs, Houston, TX, USA

Non-invasive and minimally invasive field methods were employed to reconstruct the former landscape within a heavily modified urban setting in downtown Sacramento, California. This was achieved using several datasets, including coarse-grained near-surface measurements of apparent conductivity and magnetic susceptibility, vertical profiles of fine-grained electrical conductivity, CPT soundings, and direct-push bore samples. Interpretations relied upon sequence stratigraphic analysis, grouping together related strata into Landform Sediment Assemblages providing environmental context, delineated primarily on stratal stacking patterns. Correlations between stratal sequences were made using observations of lateral continuities and key markers within specific sequences. In this manner, a three-dimensional working model of the project area was conceptualized. Cross sections suggest a shallow channel ran through the project area bordered by two pronounced natural levee deposits believed to represent consecutive back-to-back flood events in the 1860s. This work was conducted to assess the suitability of remnant landforms to host archaeological deposits possibly impacted by planned development.


Brett R. Trottier, U.S. Geological Survey, Storrs, CT, USA
Roelof J. Versteeg, Subsurface Insights, Hanover, NH, USA
Dale Werkema, U.S. Environmental Protection Agency, Newport, OR, USA
Alireza T. Meyal, Subsurface Insights, Hanover, NH, USA
Gregory F. Partridge, Subsurface Insights, Hanover, NH, USA
Carole D. Johnson, U.S. Geological Survey, Storrs, CT, USA

As flooding events increase in frequency and intensity in response to climate change, permeable pavement offers a method of flood mitigation that can reduce surface runoff in developed areas as well as promote direct aquifer recharge. The reduction of surface runoff helps to limit erosion, pollution, demands on waste-water treatment, and potential hazards. In 2021, an infiltration test station was developed in Storrs, Connecticut to evaluate the relative efficacy of permeable pavement relative to natural soil during simulated rainfall events, and to demonstrate remote and autonomous geophysical methods to continuously monitor long-term infiltration. The test site consisted of four plots: two with permeable pavement pads, one natural soil, and one impermeable control plot. Each plot was approximately 2.5 meters squared and had electrical resistivity probes installed to depths of approximately 1 meter. To replicate precipitation, three watering wands were connected to a constant water source, secured to a fixed structure, and centered above each plot at a height of approximately 0.75 meter. Three of the plots were watered for a three-hour period while the covered control plot was not. Before each test, desired flow rates were calibrated by collecting a targeted volume of water over a set time interval from each wand. Predetermined flow rates were calculated to approximately span the range of local rainfall rates. A 0.5-inch/hour flow rate was used to emulate light rainfall, a 2-inch/hour flow rate for moderate rainfall, and a 4-inch/hour flow rate for torrential downpour. Relatively minimal ponding was observed on the porous pads during testing, but as simulated precipitation rates increased, so did surface runoff. Modifications were made to accommodate the higher test rates, reduce surface runoff, and ensure vertical flow through the pads. A challenge of assessing permeable pavement efficacy is understanding the recharge impact within the subsurface. To overcome this challenge, time-lapse electrical resistivity probes were installed below each plot. The data collected from these probes were automatically uploaded to Subsurface Insights for processing and visualization. During the test, the probes recorded sudden drops in apparent resistivity, which were interpreted as infiltration in direct response to the simulated rainfall. As these probes measured resistivity along the length of the probe itself, a 1-dimensional profile of infiltration over time was observed. As contact resistance varies between probes, data were plotted relative to a temporal datum to compare infiltration responses between pavement plots. The data exhibited a sharp decrease in resistivity within a few minutes after the artificial recharge events began, followed by a sudden rebound in resistivity at the conclusion of each test, which lasted several hours, and then a gradual attenuation to pre-test values within approximately 1-1.5 days. The rapid response times indicate that both concrete pads have high permeability and similar attenuation rates to the natural soil, which demonstrates the permeable pavements have a comparable rate of infiltration and do not retain water for a longer period than the natural soil. These results suggest that infiltration through permeable pavements reasonably simulates infiltration through natural soils and may have promising applications as an alternative to impermeable pavements and structures as the demand for flood mitigation technology increases with the increase in flooding events. Additionally, these results demonstrate the successful application of remote autonomous resistivity probes to monitor recharge beneath permeable pavements. Such monitoring technology is important for municipalities and stakeholders to evaluate permeable pavement effectiveness and inform maintenance cycles.


Ao Wang, Université Rouen Normandie, UMR 6143 M2C, Rouen, France
Fayçal Rejiba, Université Rouen Normandie, UMR 6143 M2C, Rouen, France
Cécile Finco, Université Rouen Normandie, UMR 6143 M2C, Rouen, France
Luis Cavalcante Fraga, Université Rouen Normandie, UMR 6143 M2C, Rouen, France
Ayoub Saydy, Université Rouen Normandie, UMR 6143 M2C, Rouen, France, and Envisol, Rouen, France
Ludovic Bodet, Sorbonne Université, UMR 7619 METIS, Paris, France
Cyrille Fauchard, Cerema Normandie-Centre, Research Team ENDSUM, Grand-Quevilly, France

The bearing capacity of near-surface soils is an essential component in describing potential unimproved landing zones. The most straightforward measurement of soil bearing capacity can be done by a geotechnical method, the Dynamic Cone Penetrometer (DCP), at one or multiple locations on the area of interest. However, this method can only provide the soil bearing capacity at the measurement location. A bearing capacity map of a large area requires a lot of manpower and time to perform intensive measurements. Therefore, the use of geophysical methods to generate soil bearing capacity maps is a worthwhile research topic because they estimate physical properties of the soil through non-destructive and efficient measurements.
The shear-wave velocity (Vs), as a mechanical property of the soil, is most directly related to the soil bearing capacity, but the correlation between Vs and bearing capacity cannot be expressed by a single mathematical equation, since it is material or site dependent. Therefore, we try to establish the relation between them using a geostatistical framework.
A test site (30 long, 5 m width and 0.75 m depth) has been built in Cerema Normandie-Centre (Center for Studies on Risks, the Environment, Mobility and Urban Planning) with three areas with different ranges of standardized California Bearing Ratio (CBR). Surface wave method MASW (Multiple-channel Analysis of Surface Waves) and DCP measurements were performed on the centerline of the test site. 2D surface wave analysis and inversion are applied on the seismic data to extract the cross-section of Vs in the three areas.
At first stage, the raw data of DCP measurements, i.e. the penetration depth of the cone after each blow, were compared with the inverted Vs profile. A strong correlation between DCP data and Vs is observed for each area. In the second stage, a cokriging method is applied using Vs as the secondary variable and a small number of DCP data as the primary variable, to build the DCP data map of the three areas. Thanks to the important density of DCP data on the test site, this calculated DCP map can be compared with actual measurements to verify the feasibility and robustness of the use of MASW method to build soil bearing capacity maps.
Tests and verifications of this statistic-based method are required on field measurements under less controlled or even unknown conditions. Further studies are in progress where the relationships between DCP and other geophysical methods, such as EMI (Electromagnetic Induction), ERT (Electrical Resistivity Tomography) and TDR (Time Domain Reflectrometry), are currently being assessed.


Eric A. White, U.S. Geological Survey, Storrs-Mansfield, Connecticut
Carole D. Johnson, U.S. Geological Survey, Storrs-Mansfield, Connecticut
Catherine A. Fargen, U.S. Geological Survey, Louisville, Kentucky
Moriah L. Will, U.S. Geological Survey, Louisville, Kentucky
Robert A. Darner, U.S. Geological Survey, Columbus, Ohio
David C. Lampe, U.S. Geological Survey, Indianapolis, Indiana
E. Randall Bayless, U.S. Geological Survey, Indianapolis, Indiana
John W. Lane, Jr., U.S. Geological Survey, Storrs-Mansfield, Connecticut

Improperly abandoned oil and gas wells can contaminate groundwater and contribute to climate change through leakage of brine and methane gas, presenting a significant challenge to water-resources management and contaminant mitigation. During the summer of 2021, the U.S. Geological survey (USGS) carried out a towed transient electromagnetic (tTEM) survey at a municipal well field near Fort Knox, Kentucky impacted by brine contamination from an abandoned gas well.
The tTEM surveys were conducted over three days along two grids in a field upgradient and adjacent to several groundwater pumping wells. First, a coarse-scale reconnaissance survey was conducted over the entire area of interest. Based on the results of the reconnaissance survey and pre-existing site information, a smaller, focused gridded survey consisting of north-south and east-west profiles was conducted using real-time inversion software results to guide the survey and ensure delineation of the areal extent of the brine contamination.
The tTEM survey results delineated the areal extent and depth of the brine contamination and indicate accumulations of brine at or near the bedrock surface at a depth of about 30 meters. Additionally, the tTEM surveys results provided information on the underlying hydrostratigraphy and other characteristics of the brine plume useful for informing the site conceptual model and an improved understanding of the hydrostratigraphic framework. The tTEM results are consistent with those of nearby observation wells, multiple electrical resistivity surveys conducted between 2007 and 2010, and, more recently, bedrock surface mapping using the passive seismic HVSR method. Additionally, the results verified the efficacy of the diversionary pumping and hydraulic containment system in redirecting the saline water away from the groundwater discharge wells used for public water supply and confirmed the approximate location of the principal contaminant source. The tTEM surveys provided a more rapid and efficient means to delineate and characterize the subsurface as compared to conventional geophysical methods which would likely have required months of fieldwork to investigate the same area.

Copyright 2022 Environmental and Engineering Geophysical Society