We describe dissolved lithium content of water produced from the Haynesville Shale, a low-permeability, gasproducing formation of Jurassic age overlying the lithium-rich and stratigraphically close Smackover Formation. The 57 water samples and 5 rock samples were analyzed for major and minor elements and for stable water isotopes, and Li and Sr isotopes (selected samples). The samples show an average of 71.7 mg/L Li (16-125 mg/L range) for an average TDS of 126.7 g/L (55-206 g/L range). They also show that the Haynesville and Smackover dissolved lithium have distinct diagenetic pathways with a large gap in delta 7Li values (3.27 %o, n = 7 and 10.20 %o, n = 4, respectively). However, delta 7Li of Haynesville produced water and rock (-4.03 %o, n = 5) are consistent with fractionation of lithium partitioned into clay minerals relative to the resident water. The Haynesville Shale strontium isotope 87Sr/86Sr ratio (0.70830, n = 5) is higher than that of the time-period sea (0.7070), and consistent with a clastic 87Sr increase in a closed system, but lower than that of the published Arkansas carbonate-rich Upper Smackover values (0.70895, n = 72), which denotes a stronger clastic influence there. Water isotopes confirm that the samples represent formation water and have been little impacted by hydraulic fracturing or water condensation during sampling. However, low water production combined with marginal Li concentrations does not make the Haynesville Shale a primary target for lithium production.

Lithium-ion battery technology is a key component of energy storage and the demand for lithium (Li) will continue to increase in the next decades. Dissolved Li in subsurface brines represents a growing segment of the overall Li resource, and accurate estimates of Li concentrations in geological formations is crucial for its efficient management. This study employs advanced machine learning models to predict Li concentration in the Gulf Coast Basin using a comprehensive geochemical data set developed by the USGS. Six machine learning models-Gradient Boosting, Random Forest, K-Nearest Neighbor, Support Vector Regression, Neural Networks, and Extreme Gradient Boosting-were implemented and their accuracy compared and contrasted. In order to improve prediction accuracy, we also introduced a Multi-Target Sequential Chaining approach that imputes missing data from the data set sequentially. We tested three different data scenarios: (1) using the original data set (with outliers), (2) using the data set with outliers removed, and (3) introducing altered outliers to analyze model sensitivity. Results indicate that outliers impact model performance, with altered outliers helping assess robustness. The results of this work suggest spatial extensions of already known Li-rich areas (e.g., Smackover Formation) and confirm the low Li potential of most of the Gulf Coast.

The amount of lithium (Li) required to support the energy transition underscores the need to develop Li sources beyond the traditional hard-rock mining and surface continental brines. The objective of this study was to explore the potential for Li production from oilfield waters in the U.S. Gulf Coast region. Data on Li concentrations were obtained from similar to 2450 formation water samples taken in states along the greater Gulf Coast (NM, TX, OK, LA, AR, and MS). The dataset contains geochemical analyses that spans >70 years, including the USGS database (v2.3), compilation of recent data previously published by the authors, analyses from recently collected samples and from archived samples, and data from U.S. Bureau of Mines reports.Aqueous Li distribution is lognormal, typical of trace elements, with background levels in the 0-20 mg/L range. High concentrations (>= 80 mg/L) are restricted to some intervals at certain locations: Jurassic Smackover Formation, a well-known brine deposit in Arkansas (300+ mg/L Li), but also along the Gulf Coast in Texas, Louisiana, and Mississippi (several samples >= 100 mg/L Li); Permian and Pennsylvanian Granite Wash in the Anadarko Basin of Texas (several samples >100 mg/L Li) (not confirmed); and Cretaceous Edwards Formation of South Texas (several samples >= 100 mg/L Li). Genetic models for Li enrichment are lacking, but high Li seems to be related to the noted fact that Li concentrations increase with depth and with total dissolved solids. Other common characteristics of high-Li brines in the Gulf Coast are high K, Ca, and B, which reflect water-rock interactions, presence of major faults, and association with higher permeability intervals that can spread Li-bearing fluids. Host rocks for high-dissolved Li brines commonly are carbonates, possibly because of their limited sorption opportunities. Close connection to an igneous basement or to a depositional environment that supported accumulation of surface continental brines does not seem to be required.

The lower Permian of the Midland Basin of West Texas consists of mudrocks with some tight sandstones and carbonate-dominated intervals (Wolfcamp and Spraberry formations) averaging a thickness of 975 m (3200 ft). They have been the target of a large revitalization of oil and gas production from the basin using hydraulic fracturing stimulation and horizontal well technology. Little is known about the aqueous geochemistry of these intervals. We sampled produced formation water from 53 wells across the basin and put them in the context of the basin flow history. Analyzing for major and minor elements and stable water and strontium isotopes, we determined that: (1) TDS values vary from 72 to 135 g/L, decreasing with increasing depth and increasing stratigraphic age; (2) ionic composition is consistent with that of Na-Cl brines found elsewhere in the basin with no major impact of meteoric input or halite dissolution as sometimes seen in the Permian Basin; (3) stable water isotopes (delta 18O +5.0 +/- 2.0%o and delta 2H-28 +/- 5%o VSMOW) are enriched; slightly radiogenic 87Sr/86Sr ratio is consistent with local carbonates. Interpretative results suggest that the modern resident brines of the low-permeability Lower Permian originated as Permian evaporative brines coeval with Upper Permian thick halite deposition. They would have dolomitized Permian carbonates in their gravity-driven downward migration and been further modified through water-rock interactions, including smectite-illite conversion.

Perennial Balmorhea-area springs (including San Solomon spring), located at the southern boundary of the tectonic Delaware Basin, have been investigated for decades. Distantly-sourced brackish (∼2200 mg/L) baseflow and local low-TDS Ca–HCO3 stormflow (<500 mg/L) components mix in various proportions. Competing hypotheses have been offered to explain the slow spring flowrate decline (currently ∼20-30 cfs, 0.55–0.85 m3/s) during the past century. Despite many studies, aspects of spring behavior remain poorly defined, mostly due to geological, structural, and hydrogeological complexity of the capture area, including the ultimate source and multiple flow paths to the springs. Using a total of 84 samples (springs and neighboring wells), we clarified the spatial extent of the characteristic spring hydrofacies. We analyzed samples for major and trace elements, stable water isotopes, sulfur and oxygen isotopes of sulfate, and strontium isotope ratio. The geochemical data set was complemented by historical data (∼2250 samples) from the larger region and 134 ∼weekly measurements at San Solomon spring in Balmorhea State Park. In addition to refining earlier findings, we determined the combined evaporitic Guadalupian platform units and Castile Formation are the likely source of the mixed Na–Ca cation and mixed Cl–SO4 anion spring water type, and that brines from underlying formations do not contribute significantly, if any, to spring flow. The spring evaporitic Cl/Br signature is not compatible with the low value of the ratio in the brines. We also clarified the non-artesian nature of some of the Balmorhea-area springs. Located downslope and downgradient of the main springs, these gravity springs exhibit a higher TDS than the artesian springs but also lack an evaporative signal, which combined with higher nitrate (>3 mg/L N–NO3), and lower DIC 13C isotope, is consistent with irrigation return flow.

The Wolfcamp Shale within the Delaware Basin of West Texas and southeastern New Mexico is one of the largest unconventional oil reservoirs globally and also produces high volumes of relatively low salinity water. In contrast, the overlying Bone Spring Formation and, particularly, the shallower Delaware Mountain Group contain very saline brines. The objective of the study was to assess source and nature of waters produced from the Wolfcamp and to frame them in the larger geological and hydrogeological evolution of the Basin. A total of ~150 produced water samples were analyzed for major and trace element chemistry and isotope data and were complemented by public domain and confidential information. The formation shows a striking total dissolved solids (TDS) distribution with low TDS water samples (<25 g/L) restricted to the basin center that increases up to 100–140 g/L towards the basin margins. Water isotopes (heavy δ18O +5–+8‰ and δ2H -25–-10‰ VSMOW) and low Cl/Br mass ratios (100–200) support the hypothesis that low-TDS waters result from water released during smectite-to-illite conversion. Chemical and isotopic data preclude a meteoric water source and mixing hydraulic fracturing water. Anomalously high Water to Oil Ratios (WORs) in the Wolfcamp Shale (~4) are associated with higher clay content in the target horizon and with low-TDS produced water. These associations are considered causative with the high water cut and low salinity being intrinsically geologic and attributed to the thick (>2 km) mudstone package of Pennsylvanian to lower Permian age, combined with the effects of the Laramide uplift and Neogene tilting of the entire Basin.

Concern that hydraulic fracturing and natural gas production contaminates groundwater requires techniques to attribute and estimate methane flux. Although dissolved alkane and noble gas chemistry may distinguish thermogenic and microbial methane, low solubility and concentration of methane in atmosphere-equilibrated groundwater precludes the use of methane to differentiate locations affected by high and low flux of stray methane. We present a method to estimate stray gas infiltration into groundwater using dissolved nitrogen. Due to the high concentration of nitrogen in atmospheric-recharged groundwater and low concentration in natural gas, dissolved nitrogen in groundwater is much less sensitive to change than dissolved methane and may differentiate groundwater affected high and low flux of stray natural gas. We report alkane and nitrogen chemistry from shallow groundwater wells and eight natural gas production wells in the Barnett Shale footprint to attribute methane and estimate mixing ratios of thermogenic natural gas to groundwater. Most groundwater wells have trace to nondetect concentrations of methane. A cluster of groundwater wells have greater than 10 mg/L dissolved methane concentrations with alkane chemistries similar to natural gas from the Barnett Shale and/or shallower Strawn Group suggesting that localized migration of natural gas occurred. Two-component mixing models constructed with dissolved nitrogen concentrations and isotope values identify three wells that were likely affected by a large influx of natural gas with gas:water mixing ratios approaching 1:5. Most groundwater wells, even those with greater than 10-mg/L methane, have dissolved nitrogen chemistry typical of atmosphere-equilibrated groundwater suggesting natural gas:water mixing ratios smaller than 1:20. Plain Language Summary Hydraulic fracturing, horizontal drilling, and associated natural gas production have dramatically changed the energy landscape across America over the past 10 years. Along with this renaissance in the energy sector has come public concern that hydraulic fracturing may contaminate groundwater. In this study we measure the chemistry of dissolved gas from shallow groundwater wells located above the Barnett Shale natural gas play, a tight gas reservoir located west of the Dallas-Fort Worth Metroplex. We compare groundwater chemistry results to natural gas chemistry results from nearby production wells. Most groundwater wells have trace to nondetectible concentrations of methane, consistent with no measurable infiltration of natural gas into shallow groundwater. A cluster of groundwater wells have greater than 10 mg/L dissolved methane concentrations with alkane chemistries similar to natural gas. Using dissolved nitrogen and alkane concentrations and their stable isotope ratios in combination with chemical mixing models, we conclude that natural gas transported from the shallower Strawn Group affected these groundwater wells rather than natural gas from the deeper Barnett Shale, which is the target of hydraulic fracturing in this area. These results suggest that hydraulic fracturing has not affected shallow groundwater drinking sources in this area.

Effective, considerate shale play water management supports operations and protects the environment. A parameter often overlooked is total dissolved solids (TDS) of produced water from the formation. Knowledge of TDS is important to meet these dual goals. Subsurface TDS typically increases with depth. However, produced-water samples from the Eagle Ford Shale show a strong TDS decrease by a factor of similar to 10 with increasing well depth (similar to 200 000 ppm at similar to 2.5 km to 18 000 ppm at similar to 3.6 km). Water stable isotopes strongly suggest that the low TDS is not due to dilution by meteoric water. Rather, we attribute the change to smectite-to-illite conversion, in which the smectite interlayer water is released into the pore space. Depth, temperature, and other related indicators (source for K, excess silica) support such a mechanism. In addition, water-isotope patterns and Sr-87/Sr-86 ratios suggest a conversion operating with limited contributions external to the shale. Order-of-magnitude calculations show that the 8% of mixed-layer clay present on average in the Lower Eagle Ford Shale is sufficient to bring formation water TDS to observed levels when some of the resident water is expelled. Understanding that the low salinity is an intrinsic property of the formation water rather than due to short-term mixing allows stakeholders to have a more optimistic outlook on water recycling and on using produced water for other uses (irrigation, municipal).

Assessing natural vs. anthropogenic sources of methane in drinking water aquifers is a critical issue in areas of shale oil and gas production. The objective of this study was to determine controls on methane occurrences in aquifers in the Eagle Ford Shale play footprint. A total of 110 water wells were tested for dissolved light alkanes, isotopes of methane, and major ions, mostly in the eastern section of the play. Multiple aquifers were sampled with approximately 47 samples from the Carrizo-Wilcox Aquifer (250-1200 m depth range) and Queen City-Sparta Aquifer (150-900 m depth range) and 63 samples from other shallow aquifers but mostly from the Catahoula Formation (depth <150 m). Besides three shallow wells with unambiguously microbial methane, only deeper wells show significant dissolved methane (22 samples >1 mg/L, 10 samples >10 mg/L). No dissolved methane samples exhibit thermogenic characteristics that would link them unequivocally to oil and gas sourced from the Eagle Ford Shale. In particular, the well water samples contain very little or no ethane and propane (C1/C2+C3 molar ratio >453), unlike what would be expected in an oil province, but they also display relatively heavier δ13 Cmethane (>-55‰) and δDmethane (>-180‰). Samples from the deeper Carrizo and Queen City aquifers are consistent with microbial methane sourced from syndepositional organic matter mixed with thermogenic methane input, most likely originating from deeper oil reservoirs and migrating through fault zones. Active oxidation of methane pushes δ13 Cmethane and δDmethane toward heavier values, whereas the thermogenic gas component is enriched with methane owing to a long migration path resulting in a higher C1/C2+C3 ratio than in the local reservoirs.

This study investigates water-rock interactions induced by brine injection into the Cretaceous clastic Hosston Formation in South Central Texas, a potential host formation for disposal of produced water from the Eagle Ford Shale during oil and gas extraction. The primary objectives of this experimental study were to (1) predict the geochemical reactions between the injected brine and the host formation during disposal operation, and (2) assess potential impact on flow properties of the host rock. The results provide important information of potential impact to injectivity and available methods to alleviate the problems if they occur during brine disposal. Five batch autoclave experiments were conducted to react an argillaceous sandstone of the Hosston Formation with a series of synthetic brines and the Eagle Ford brine for 3 weeks at 193 bar and 80 degrees C. To better identify mineral dissolution and precipitation, rock samples were ion-milled and examined using scanning electron microscopy (SEM) before and after the experiments. For each experiment, approximately 15 aqueous samples were taken with increasing time intervals from hours to days. Dissolution of anhydrite, which results in a rapid increase of SO4 and Ca concentrations in the solution of similar to 500 ppm and similar to 200 ppm respectively, is the most significant reaction. Small amounts of dolomite and K-feldspar were also dissolved to release Mg, Ca, K, and Si. Small amounts of iron hydroxides and kaolinite precipitated. Geochemical modeling is able to reproduce the behavior of the major aqueous components and of the observed mineral dissolution and precipitation. Overall, change in flow properties from the water-rock interaction is insignificant: mercury intrusion capillary pressure (MICP) data show little difference in porosity, permeability, and pore-throat-size distribution between the pre-and post-experiment samples. This study suggests that injection of Eagle Ford brine into the Hosston Formation will pose minimum risks for the brine disposal operation. (C) 2017 Elsevier Ltd. All rights reserved.

Clusters of elevated methane concentrations in aquifers overlying the Barnett Shale play have been the focus of recent national attention as they relate to impacts of hydraulic fracturing. The objective of this study was to assess the spatial extent of high dissolved methane previously observed on the western edge of the play (Parker County) and to evaluate its most likely source. A total of 509 well water samples from 12 counties (14,500 km2) were analyzed for methane, major ions, and carbon isotopes. Most samples were collected from the regional Trinity Aquifer and show only low levels of dissolved methane (85% of 457 unique locations <0.1 mg/L). Methane, when present is primarily thermogenic (δ13C 10th and 90th percentiles of −57.54 and −39.00‰ and C1/C2+C3 ratio 10th, 50th, and 90th percentiles of 5, 15, and 42). High methane concentrations (>20 mg/L) are limited to a few spatial clusters. The Parker County cluster area includes historical vertical oil and gas wells producing from relatively shallow formations and recent horizontal wells producing from the Barnett Shale (depth of ∼1500 m). Lack of correlation with distance to Barnett Shale horizontal wells, with distance to conventional wells, and with well density suggests a natural origin of the dissolved methane. Known commercial very shallow gas accumulations (<200 m in places) and historical instances of water wells reaching gas pockets point to the underlying Strawn Group of Paleozoic age as the main natural source of the dissolved gas.

Geochemical interactions between shale and hydraulic fracturing fluid may affect produced-water chemistry and rock properties. It is important to investigate the rock–water reactions to understand the impacts. Eight autoclave experiments reacting Marcellus and Eagle Ford Shale samples with synthetic brines and a friction reducer were conducted for more than 21 days. To better determine mineral dissolution and precipitation at the rock–water interface, the shale samples were ion milled to create extremely smooth surfaces that were characterized before and after the autoclave experiments using scanning electron microscopy (SEM). This method provides an unprecedented level of detail and the ability to directly compare the same mineral particles before and after the reaction experiments. Dissolution area was quantified by tracing and measuring the geometry of newly formed pores. Changes in porosity and permeability were also measured by mercury intrusion capillary pressure (MICP) tests. Aqueous chemistry and SEM observations show that dissolution of calcite, dolomite, and feldspar and pyrite oxidation are the primary mineral reactions that control the concentrations of Ca, Mg, Sr, Mn, K, Si, and SO4 in aqueous solutions. Porosity measured by MICP also increased up to 95%, which would exert significant influence on fluid flow in the matrix along the fractures. Mineral dissolution was enhanced and precipitation was reduced in solutions with higher salinity. The addition of polyacrylamide (a friction reducer) to the reaction solutions had small and mixed effects on mineral reactions, probably by plugging small pores and restricting mineral precipitation. The results suggest that rock–water interactions during hydraulic fracturing likely improve porosity and permeability in the matrix along the fractures by mineral dissolution. The extent of the geochemical reactions is controlled by the salinity of the fluids, with higher salinity enhancing mineral dissolution.

There is concern about adverse impacts of natural gas (primarily methane) production on groundwater quality; however, data on trace element concentrations are limited. The objective of this study was to compare the distribution of trace elements in groundwater samples with and without dissolved methane in aquifers overlying the Barnett Shale (Hood and Parker counties, 207 samples) and the Haynesville Shale (Panola County, 42 samples). Both shales have been subjected to intensive hydraulic fracturing for gas production. Well clusters with high dissolved methane were previously found in these counties and are thought to be of natural origin. Overall, groundwater in these counties is of excellent quality with typically low elemental concentrations. Several statistical analyses strongly suggest that most trace element concentrations, generally at low background levels, are no higher and even reduced when dissolved methane is present. In addition, trace element concentrations are not correlated with distance to gas wells. The reduction in trace element concentrations is attributed to anaerobic microbial degradation of methane, is associated with a higher pH (>8.5), and, likely, with precipitation of carbonates and pyrite and formation of clays. Trace and other elements are likely incorporated within the precipitating mineral crystalline network or sorbed. High pH values are found throughout these high-methane clusters (e.g., Parker-Hood cluster), even in subregions where methane is not present, which is consistent with a pervasive natural origin of dissolved methane rather than a limited gas well source.

This study presents the complete set of stable noble gases for Barnett Shale and Strawn Group production gas together with stray flowing gas in the Trinity Aquifer, Texas. It places new constraints on the source of this stray gas and further shows that Barnett and Strawn gas have distinct crustal and atmospheric noble gas signatures, allowing clear identification of these two sources. Like stray gas, Strawn gas is significantly more enriched in crustal 4He*, 21Ne*, and 40Ar* than Barnett gas. The similarity of Strawn and stray gas crustal noble gas signatures suggests that the Strawn is the source of stray gas in the Trinity Aquifer. Atmospheric 22Ne/36Ar ratios of stray gas mimic also that of Strawn, further reinforcing the notion that the source of stray gas in this aquifer is the Strawn. While noble gas signatures of Strawn and stray gas are consistent with a single-stage water degassing model, a two-stage oil modified groundwater exsolution fractionation model is required to explain the light atmospheric noble gas signature of Barnett Shale production gas. These distinct Strawn and Barnett noble gas signatures are likely the reflection of distinct evolution histories with Strawn gas being possibly older than that of Barnett Shale.

The Intergovernmental Panel on Climate Change (IPCC) states that quantification of leakage emissions at CO2 geological storage sites will require direct monitoring techniques rather than standardized emissions factors. In addition, IPCC guidelines state that leakage emissions should be distinguished from natural baseline emissions. Current methods measure only bulk CO2 emissions and therefore do not have the capability to separate these components. We discuss that a process-based soil gas approach could potentially be used with an open chamber emissions quantification method to translate process-based concentration data into surface flux rates. In doing so, the various components of surface gas flux could be attributed, separated and individually quantified. We use soil gas data from the ZERT controlled release experimental site as the basis for our discussion and show that such a method may improve the accuracy and ease of leakage quantification. For accounting purposes, such capabilities could be used to more accurately assess the number of credits to be surrendered for leakage. For monitoring purposes, added benefits could include delineating the spatial extent of leakage emissions and providing real-time information on the effectiveness of remediation efforts.

Understanding the source of dissolved methane in drinking-water aquifers is critical for assessing potential contributions from hydraulic fracturing in shale plays. Shallow groundwater in the Texas portion of the Haynesville Shale area (13,000 km(2)) was sampled (70 samples) for methane and other dissolved light alkanes. Most samples were derived from the fresh water bearing Wilcox formations and show little methane except in a localized cluster of 12 water wells (17% of total) in a approximately 30x30 km(2) area in Southern Panola County with dissolved methane concentrations less than 10 mg/L. This zone of elevated methane is spatially associated with the termination of an active fault system affecting the entire sedimentary section, including the Haynesville Shale at a depth more than 3.5 km, and with shallow lignite seams of Lower Wilcox age at a depth of 100 to 230 m. The lignite spatial extension overlaps with the cluster. Gas wetness and methane isotope compositions suggest a mixed microbial and thermogenic origin with contribution from lignite beds and from deep thermogenic reservoirs that produce condensate in most of the cluster area. The pathway for methane from the lignite and deeper reservoirs is then provided by the fault system.

This study places constraints on the source and transport mechanisms of methane found in groundwater within the Barnett Shale footprint in Texas using dissolved noble gases, with particular emphasis on 84Kr and 132Xe. Dissolved methane concentrations are positively correlated with crustal 4He, 21Ne, and 40Ar and suggest that noble gases and methane originate from common sedimentary strata, likely the Strawn Group. In contrast to most samples, four water wells with the highest dissolved methane concentrations unequivocally show strong depletion of all atmospheric noble gases (20Ne, 36Ar, 84Kr, 132Xe) with respect to air-saturated water (ASW). This is consistent with predicted noble gas concentrations in a water phase in contact with a gas phase with initial ASW composition at 18 °C-25 °C and it suggests an in situ, highly localized gas source. All of these four water wells tap into the Strawn Group and it is likely that small gas accumulations known to be present in the shallow subsurface were reached. Additionally, lack of correlation of 84Kr/36Ar and 132Xe/36Ar fractionation levels along with 4He/20Ne with distance to the nearest gas production wells does not support the notion that methane present in these groundwaters migrated from nearby production wells either conventional or using hydraulic fracturing techniques.