Field evaluation of semi-automated moisture estimation from geophysics using machine learning.

Geophysical methods can provide three-dimensional (3D), spatially continuous estimates of soil moisture. However, point-to-point comparisons of geophysical properties to measure soil moisture data are frequently unsatisfactory, resulting in geophysics being used for qualitative purposes only. This is because (1) geophysics requires models that relate geophysical signals to soil moisture, (2) geophysical methods have potential uncertainties resulting from smoothing and artifacts introduced from processing and inversion, and (3) results from multiple geophysical methods are not easily combined within a single soil moisture estimation framework. To investigate these potential limitations, an irrigation experiment was performed wherein soil moisture was monitored through time, and several surface geophysical datasets indirectly sensitive to soil moisture were collected before and after irrigation: ground penetrating radar, electrical resistivity tomography (ERT), and frequency domain electromagnetics (FDEM). Data were exported in both raw and processed form, and then snapped to a common 3D grid to facilitate moisture prediction by standard calibration techniques, multivariate regression, and machine learning. A combination of inverted ERT data, raw FDEM, and inverted FDEM data was most informative for predicting soil moisture using a random regression forest model (one-thousand 60/40 training/test cross-validation folds produced root mean squared errors ranging from 0.025-0.046 cm3/cm3). This cross-validated model was further supported by a separate evaluation using a test set from a physically separate portion of the study area. Machine learning was conducive to a semi-automated model-selection process that could be used for other sites and datasets to locally improve accuracy.

Evaluation of Models for Estimating Hydraulic Conductivity in Glacial Aquifers with NMR Logging.

Nuclear magnetic resonance (NMR) logging is a promising method for estimating hydraulic conductivity (K). During the past ∼60 years, NMR logging has been used for petroleum applications, and different models have been developed for deriving estimates of permeability. These models involve calibration parameters whose values were determined through decades of research on sandstones and carbonates. We assessed the use of five models to derive estimates of K in glacial aquifers from NMR logging data acquired in two wells at each of two field sites in central Wisconsin, USA. Measurements of K, obtained with a direct push permeameter (DPP), KDPP , were used to obtain the calibration parameters in the Schlumberger-Doll Research, Seevers, Timur-Coates, Kozeny-Godefroy, and sum-of-echoes (SOE) models so as to predict K from the NMR data; and were also used to assess the ability of the models to predict KDPP . We obtained four well-scale calibration parameter values for each model using the NMR and DPP measurements in each well; and one study-scale parameter value for each model by using all data. The SOE model achieved an agreement with KDPP that matched or exceeded that of the other models. The Timur-Coates estimates of K were found to be substantially different from KDPP . Although the well-scale parameter values for the Schlumberger-Doll, Seevers, and SOE models were found to vary by less than a factor of 2, more research is needed to confirm their general applicability so that site-specific calibration is not required to obtain accurate estimates of K from NMR logging data.

GW/SW-MST: A Groundwater/Surface-Water Method Selection Tool.

Groundwater/surface-water (GW/SW) exchange and hyporheic processes are topics receiving increasing attention from the hydrologic community. Hydraulic, chemical, temperature, geophysical, and remote sensing methods are used to achieve various goals (e.g., inference of GW/SW exchange, mapping of bed materials, etc.), but the application of these methods is constrained by site conditions such as water depth, specific conductance, bed material, and other factors. Researchers and environmental professionals working on GW/SW problems come from diverse fields and rarely have expertise in all available field methods; hence there is a need for guidance to design field campaigns and select methods that both contribute to study goals and are likely to work under site-specific conditions. Here, we present the spreadsheet-based GW/SW-Method Selection Tool (GW/SW-MST) to help practitioners identify methods for use in GW/SW and hyporheic studies. The GW/SW-MST is a Microsoft Excel-based decision support tool in which the user selects answers to questions about GW/SW-related study goals and site parameters and characteristics. Based on user input, the tool indicates which methods from a toolbox of 32 methods could potentially contribute to achieving the specified goals at the site described.

Assessment of NMR Logging for Estimating Hydraulic Conductivity in Glacial Aquifers.

Glacial aquifers are an important source of groundwater in the United States and require accurate characterization to make informed management decisions. One parameter that is crucial for understanding the movement of groundwater is hydraulic conductivity, K. Nuclear magnetic resonance (NMR) logging measures the NMR response associated with the water in geological materials. By utilizing an external magnetic field to manipulate the nuclear spins associated with 1 H, the time-varying decay of the nuclear magnetization is measured. This logging method could provide an effective way to estimate K at submeter vertical resolution, but the models that relate NMR measurements to K require calibration. At two field sites in a glacial aquifer in central Wisconsin, we collected a total of four NMR logs and obtained measurements of K in their immediate vicinity with a direct-push permeameter (DPP). Using a bootstrap algorithm to calibrate the Schlumberger-Doll Research (SDR) NMR-K model, we estimated K to within a factor of 5 of the DPP measurements. The lowest levels of accuracy occurred in the lower-K (K < 10-4  m/s) intervals. We also evaluated the applicability of prior SDR model calibrations. We found the NMR calibration parameters varied with K, suggesting the SDR model does not incorporate all the properties of the pore space that control K. Thus, the expected range of K in an aquifer may need to be considered during calibration of NMR-K models. This study is the first step toward establishing NMR logging as an effective method for estimating K in glacial aquifers.

Geophysical Assessment of a Proposed Landfill Site in Fredericktown, Missouri.

In cooperation with the U.S. Environmental Protection Agency (EPA), the U.S. Geological Survey (USGS) collected numerous types of geophysical data at a site in Fredericktown, Missouri, in June 2018. Various geophysical surveys were collectively used to help evaluate the overall suitability of the site for use as a mine waste-soil and sediment repository, and to evaluate the suite of geophysical methods for rapid collection and preliminary assessment of sites with shallow sediments. Land-based geophysical methods, which included frequency-domain electromagnetic induction (FDEM), electrical resistivity tomography (ERT), horizontal-to-vertical spectral ratio passive seismic (HVSR), and shear-wave refraction, were used to determine the depths to crystalline bedrock and characterize the overlying unconsolidated sediments (or regolith). Water-borne FDEM profiles and forward-looking infrared (FLIR) thermal image surveys were conducted along the Fredericktown City Lake shoreline to identify locations of potential interactions between groundwater and surface water. Sediment temperature profilers were installed at two locations along the shoreline to characterize shallow unconsolidated sediment thermal properties and support the interpretation of the other geophysical surveys. Geophysical reconnaissance methods including the FDEM and HVSR methods, were used to rapidly evaluate the vertical and lateral extent of overburden, or unconsolidated sediments, overlying the bedrock at the site. The results of these methods were compared to reference geophysical methods of ERT and shear-wave refraction surveys that have greater accuracy and are more labor intensive and time-consuming. A goal of the project was the evaluation of the validity and reliability of this suite of reconnaissance geophysical methods as a means by which shallow (less than 3 meters (m)) sediments can be rapidly assessed. Two orthogonal ERT survey profiles, which used 28 electrodes spaced 1 m apart in dipole-dipole and combined Wenner-Schlumberger configurations, were collected to determine the subsurface resistivity. The results were inverted to produce electrical resistivity profiles that were compared to the FDEM and HVSR survey results. The FDEM data were collected along cleared paths through the proposed disposal cell locations. The data were inverted to generate depth-dependent estimates of electrical conductivity along the transects. An analysis of the depth of investigation (DOI) indicated the FDEM imaged to depths of about 3 m below land surface. The ERT, FDEM, and HVSR indicated the depth to crystalline bedrock was approximately 1.5 m below land surface with shallower and deeper areas. Results from this investigation indicate this suite of methods will likely perform well at sites with shallow depths to bedrock and strong conductivity and acoustic impedance contrasts, where the FDEM and HVSR methods can provide estimates of the depth to bedrock, and ERT and shear-wave refraction surveys might not be worth the added time and expense.

Characterizing the diverse hydrogeology underlying rivers and estuaries using new floating transient electromagnetic methodology.

The hydrogeology below large surface water features such as rivers and estuaries is universally under-informed at the long reach to basin scales (tens of km+). This challenge inhibits the accurate modeling of fresh/saline groundwater interfaces and groundwater/surface water exchange patterns at management-relevant spatial extents. Here we introduce a towed, floating transient electromagnetic (TEM) system (i.e. FloaTEM) for rapid (up to 15 km/h) high resolution electrical mapping of the subsurface below large water bodies to depths often a factor of 10 greater than other towed instruments. The novel FloaTEM system is demonstrated at a range of diverse 4th through 6th-order riverine settings across the United States including 1) the Farmington River, near Hartford, Connecticut; 2) the Upper Delaware River near Barryville, New York; 3) the Tallahatchie River near Shellmound, Mississippi; and, 4) the Eel River estuary, on Cape Cod, near Falmouth, Massachusetts. Airborne frequency-domain electromagnetic and land-based towed TEM data are also compared at the Tallahatchie River site, and streambed geologic scenarios are explored with forward modeling. A range of geologic structures and pore water salinity interfaces were identified. Process-based interpretation of the case study data indicated FloaTEM can resolve varied sediment-water interface materials, such as the accumulation of fines at the bottom of a reservoir and permeable sand/gravel riverbed sediments that focus groundwater discharge. Bedrock layers were mapped at several sites, and aquifer confining units were defined at comparable resolution to airborne methods. Terrestrial fresh groundwater discharge with flowpaths extending hundreds of meters from shore was also imaged below the Eel River estuary, improving on previous hydrogeological characterizations of that nutrient-rich coastal exchange zone. In summary, the novel FloaTEM system fills a critical gap in our ability to characterize the hydrogeology below surface water features and will support more accurate prediction of groundwater/surface water exchange dynamics and fresh-saline groundwater interfaces.

Evaluating long-term patterns of decreasing groundwater discharge through a lake-bottom permeable reactive barrier.

Identifying and quantifying groundwater exchange is critical when considering contaminant fate and transport at the groundwater/surface-water interface. In this paper, areally distributed temperature and point seepage measurements are used to efficiently assess spatial and temporal groundwater discharge patterns through a glacial-kettle lakebed area containing a zero-valent iron permeable reactive barrier (PRB). Concern was that the PRB was becoming less permeable with time owing to biogeochemical processes within the PRB. Patterns of groundwater discharge over an 8-year period were examined using fiber-optic distributed temperature sensing (FO-DTS) and snapshot-in-time point measurements of temperature. The resulting thermal maps show complex and uneven distributions of temperatures across the lakebed and highlight zones of rapid seepage near the shoreline and along the outer boundaries of the PRB. Repeated thermal mapping indicates an increase in lakebed temperatures over time at periods of similar stage and surface-water temperature. Flux rates in six seepage meters permanently installed on the lakebed in the PRB area decreased on average by 0.021 md-1 (or about 4.5 percent) annually between 2004 and 2015. Modeling of diurnal temperature signals from shallow vertical profiles yielded mean flux values ranging from 0.39 to 1.15 md-1, with stronger fluxes generally related to colder lakebed temperatures. The combination of an increase in lakebed temperatures, declines in direct seepage, and observations of increased cementation of the lakebed surface provide in situ evidence that the permeability of the PRB is declining. The presence of temporally persistent rapid seepage zones is also discussed.

Bioremediation in Fractured Rock: 1. Modeling to Inform Design, Monitoring, and Expectations.

Field characterization of a trichloroethene (TCE) source area in fractured mudstones produced a detailed understanding of the geology, contaminant distribution in fractures and the rock matrix, and hydraulic and transport properties. Groundwater flow and chemical transport modeling that synthesized the field characterization information proved critical for designing bioremediation of the source area. The planned bioremediation involved injecting emulsified vegetable oil and bacteria to enhance the naturally occurring biodegradation of TCE. The flow and transport modeling showed that injection will spread amendments widely over a zone of lower-permeability fractures, with long residence times expected because of small velocities after injection and sorption of emulsified vegetable oil onto solids. Amendments transported out of this zone will be diluted by groundwater flux from other areas, limiting bioremediation effectiveness downgradient. At nearby pumping wells, further dilution is expected to make bioremediation effects undetectable in the pumped water. The results emphasize that in fracture-dominated flow regimes, the extent of injected amendments cannot be conceptualized using simple homogeneous models of groundwater flow commonly adopted to design injections in unconsolidated porous media (e.g., radial diverging or dipole flow regimes). Instead, it is important to synthesize site characterization information using a groundwater flow model that includes discrete features representing high- and low-permeability fractures. This type of model accounts for the highly heterogeneous hydraulic conductivity and groundwater fluxes in fractured-rock aquifers, and facilitates designing injection strategies that target specific volumes of the aquifer and maximize the distribution of amendments over these volumes.

An overview of geophysical technologies appropriate for characterization and monitoring at fractured-rock sites.

Geophysical methods are used increasingly for characterization and monitoring at remediation sites in fractured-rock aquifers. The complex heterogeneity of fractured rock poses enormous challenges to groundwater remediation professionals, and new methods are needed to cost-effectively infer fracture and fracture-zone locations, orientations and properties, and to develop conceptual site models for flow and transport. Despite the potential of geophysical methods to "see" between boreholes, two issues have impeded the adoption of geophysical methods by remediation professionals. First, geophysical results are commonly only indirectly related to the properties of interest (e.g., permeability) to remediation professionals, and qualitative or quantitative interpretation is required to convert geophysical results to hydrogeologic information. Additional demonstration/evaluation projects are needed in the site remediation literature to fully transfer geophysical methods from research to practice. Second, geophysical methods are commonly viewed as inherently risky by remediation professionals. Although it is widely understood that a given method may or may not work at a particular site, the reasons are not always clear to end users of geophysical products. Synthetic modeling tools are used in research to assess the potential of a particular method to successfully image a target, but these tools are not widely used in industry. Here, we seek to advance the application of geophysical methods to solve problems facing remediation professionals with respect to fractured-rock aquifers. To this end, we (1) provide an overview of geophysical methods applied to characterization and monitoring of fractured-rock aquifers; (2) review case studies showcasing different geophysical methods; and (3) discuss best practices for method selection and rejection based on synthetic modeling and decision support tools.

Flow and sorption controls of groundwater arsenic in individual boreholes from bedrock aquifers in central Maine, USA.

To understand the hydrogeochemical processes regulating well water arsenic (As) evolution in fractured bedrock aquifers, three domestic wells with [As] up to 478 µg/L are investigated in central Maine. Geophysical logging reveals that fractures near the borehole bottom contribute 70-100% of flow. Borehole and fracture water samples from various depths show significant proportions of As (up to 69%) and Fe (93-99%) in particulates (>0.45 µm). These particulates and those settled after a 16-day batch experiment contain 560-13,000 mg/kg of As and 14-35% weight/weight of Fe. As/Fe ratios (2.5-20 mmol/mol) and As partitioning ratios (adsorbed/dissolved [As], 20,000-100,000 L/kg) suggest that As is sorbed onto amorphous hydrous ferric oxides. Newly drilled cores also show enrichment of As (up to 1300 mg/kg) sorbed onto secondary iron minerals on the fracture surfaces. Pumping at high flow rates induces large decreases in particulate As and Fe, a moderate increase in dissolved [As] and As(III)/As ratio, while little change in major ion chemistry. The δD and δ(18)O are similar for the borehole and fracture waters, suggesting a same source of recharge from atmospheric precipitation. Results support a conceptual model invoking flow and sorption controls on groundwater [As] in fractured bedrock aquifers whereby oxygen infiltration promotes the oxidation of As-bearing sulfides at shallower depths in the oxic portion of the flow path releasing As and Fe; followed by Fe oxidation to form Fe oxyhydroxide particulates, which are transported in fractures and sorb As along the flow path until intercepted by boreholes. In the anoxic portions of the flow path, reductive dissolution of As-sorbed iron particulates could re-mobilize As. For exposure assessment, we recommend sampling of groundwater without filtration to obtain total As concentration in groundwater.

A small-diameter NMR logging tool for groundwater investigations.

A small-diameter nuclear magnetic resonance (NMR) logging tool has been developed and field tested at various sites in the United States and Australia. A novel design approach has produced relatively inexpensive, small-diameter probes that can be run in open or PVC-cased boreholes as small as 2 inches in diameter. The complete system, including surface electronics and various downhole probes, has been successfully tested in small-diameter monitoring wells in a range of hydrogeological settings. A variant of the probe that can be deployed by a direct-push machine has also been developed and tested in the field. The new NMR logging tool provides reliable, direct, and high-resolution information that is of importance for groundwater studies. Specifically, the technology provides direct measurement of total water content (total porosity in the saturated zone or moisture content in the unsaturated zone), and estimates of relative pore-size distribution (bound vs. mobile water content) and hydraulic conductivity. The NMR measurements show good agreement with ancillary data from lithologic logs, geophysical logs, and hydrogeologic measurements, and provide valuable information for groundwater investigations.

A computer program for flow-log analysis of single holes (FLASH).

A new computer program, FLASH (Flow-Log Analysis of Single Holes), is presented for the analysis of borehole vertical flow logs. The code is based on an analytical solution for steady-state multilayer radial flow to a borehole. The code includes options for (1) discrete fractures and (2) multilayer aquifers. Given vertical flow profiles collected under both ambient and stressed (pumping or injection) conditions, the user can estimate fracture (or layer) transmissivities and far-field hydraulic heads. FLASH is coded in Microsoft Excel(5) with Visual Basic for Applications routines. The code supports manual and automated model calibration.