Traditional petrophysical methods to evaluate organic richness and mineralogy using gamma-ray and resistivity log responses are not diagnostic in source rocks. We have developed a deterministic, nonproprietary method to quantify formation variability in total organic carbon (TOC) and three key mudrock mineralogical components of nonhydrocarbon-bearing source rock strata of the Eagle Ford Group by developing a set of log-derived multimineral models calibrated with Fourier transform infrared spectroscopy core data from the research borehole U.S. Geological Survey Gulf Coast 1 West Woodway. We determined that bulk density response is a reliable indicator of organic content in these thermally immature, water-bearing source rocks. Multimineral findings indicate that a high degree of laminae-scale mineralogical heterogeneity exists due to thinly interbedded carbonate cements amid clay-rich mudstone layers. The lower part of the Eagle Ford Group contains the highest average TOC content (4.7 wt%) and the highest average carbonate volume (64.1 vol%), making it the optimal target in thermally mature areas for source-rock potential and hydraulic-fracture placement. In contrast, the uppermost portion of the Eagle Ford Group contains the highest average volume of clay minerals (42.6 vol%), which increases the potential for wellbore stability issues. Petrophysical characterization reveals that porosity is approximately 30% in this relatively uncompacted formation. In this thermally immature source rock, water saturation is nearly 100% and no free hydrocarbons were observed on the resistivity logs. No evidence of borehole ellipticity was observed on the three-arm caliper log, and horizontal stresses are presumed to be directionally uniform in the vicinity of this near-surface wellbore. This shallow wellbore has a temperature gradient of 1.87°F/100 ft (16.3°C/km) and is likely influenced by earth surface heating.

Several crucial aspects of exploration and production of source-rock reservoirs are reliant on accurate characterization of mudstone mineralogical variability, including the evaluation of organic richness for resource potential and the determination of formation brittleness for hydraulic-fracture placement. Traditional methods for studying mudstone mineralogy rely on geochemical analyses of core samples taken at irregular intervals, often several feet apart, which may not capture complex lithologic and vertical heterogeneity. Geophysical logging data provide a continuous record of log responses to fluid and formation properties at depth increments on the order of inches; however, individual log responses do not uniquely correspond to specific mineral compositions (Crain, 2004). Calibrating log responses to core-derived mineralogy and total organic carbon (TOC) measurements enables the development of log-derived multimineral models, an often proprietary process embedded within expensive software, that retains nearly a 20-fold increase in sampling resolution over available geochemistry and mineralogy data collected on core.

To advance the understanding of source-rock mineralogy and rock properties in a shallow-burial environment where organic matter is thermally immature, where the mineral framework is not fully compacted, and where the groundmass of the mudstone is only partially or locally cemented, the U.S. Geological Survey (USGS) cored and logged a series of shallow research boreholes into Eagle Ford Group and overlying and underlying strata near Waco, Dallas, and Del Rio, Texas (Figure 1a). The Eagle Ford is currently the second most productive oil-producing continuous resource play in the United States, with 1.12 million barrels per day of liquids production in 2019 (Texas Railroad Commission, 2019), making the evaluation of the Eagle Ford Group a strategically important objective. From the USGS Gulf Coast 1 West Woodway borehole (Figure 1a, the red star; referred to as GC-1 hereafter), approximately 600 ft (183 m) of continuous core with a diameter of 2-3/8 in. (6.03 cm) and a comprehensive suite of wireline geophysical logs at 0.1 ft (0.03 m) depth increments were acquired.

The objectives of this study were to (1) develop a nonproprietary method to characterize organic richness and mudstone mineralogy by developing high-resolution models based on geophysical logs and calibrated with core mineralogical data from the GC-1, (2) classify mudstone lithology based on log-derived ternary diagrams, and (3) characterize petrophysical rock properties, saturating pore fluids, porosity, permeability, stress regime, mechanical integrity, and brittleness of these thermally immature, uncompacted source rocks. The Cenomanian-Turonian mudstones of the Eagle Ford Group and, to a lesser extent, the Cenomanian Pepper Shale Member of the Woodbine Formation and Del Rio Formation in the GC-1 are the focus of this study. Interdisciplinary data sets for this study included geophysical wireline logging curves, petrophysical multimineral models, core observation, and geochemical analyses.

The Eagle Ford Group in central Texas is an organic-rich calcareous mudstone regionally subdivided into upper and lower stratigraphic units (Denne et al., 2014, 2016a, 2016b; Fairbanks et al., 2016). Deposition of the Eagle Ford is associated with a second-order sea-level transgression during the Late Cretaceous (Robison, 1997; Dawson, 2000; Hammes et al., 2016). The lower stratigraphic unit contains transgressive deposits associated with lower energy and slightly anoxic marine paleoenvironments, whereas the upper unit contains regressive highstand deposits from oxygenated shallow-marine paleoenvironments (Liro et al., 1994; Dawson, 2000).

Strata studied in this paper include, from the base up (Figure 1b), the (1) lower Cenomanian Del Rio Formation, (2) lower to middle Cenomanian Pepper Shale Member of the Woodbine Formation, (3) middle to upper Cenomanian lower portion of the Eagle Ford Group, (4) lower to middle Turonian lower portion of the upper Eagle Ford Group, (5) middle to upper Turonian upper portion of the Eagle Ford Group, and (6) Coniacian-Santonian Austin Chalk (Figure 1b). Mudstone stratigraphic units from the Del Rio Formation to the uppermost part of the Eagle Ford Group were interpreted in this study based on core observations, log analysis, geochemistry data, and the literature (Dawson, 2000; Fairbanks et al., 2016; Birdwell et al., 2017; French et al., 2019). The mudstone stratigraphic nomenclature used herein is inferred by projection (into the borehole) of units defined by past mapping and stratigraphic studies conducted in the greater Waco area, and preliminary biostratigraphic work on the core (R. Denne, written communication, 2018).

Photographs of select intervals of the GC-1 core (Figure 2) show the five mudstone sedimentary facies of interest in this study. The Del Rio is observed to be a medium-gray calcareous mudstone with abundant fossils such as bivalve shells and thin laminations of foraminifera. This faunal assemblage is characteristic of oxygenated bottom-water conditions that were unfavorable for the preservation of organic material. Based on these core observations and previous study of microfauna within the Del Rio (Bullard, 1953), the depositional environment is presumed to be relatively warm neritic, such as the interface between lagoonal and shallow marine.

The Pepper is a noncalcareous mudrock with terrigenous quartz-siltstone laminations (Donovan et al., 2015). These sediments contain high clay content, including illite, kaolinite, and illite-smectite (Fairbanks et al., 2016). This unit may have been deposited in a shallow lagoon environment with brackish-water influence (Young, 1977). The contact between the Pepper and the underlying Del Rio is disconformable in this borehole.

Middle to upper Cenomanian mudstone of the lower part of the Eagle Ford Group (henceforth referred to as the lower Eagle Ford [LEF]) is a major source rock (Robison, 1997) and, in this borehole, exhibits the highest TOC of the Eagle Ford Group. In this core, the LEF is a laminated calcareous mudstone observed to contain foraminiferal microfossil layers and high-gamma-ray (GR) bentonitic clays likely originating from Laramide volcanic events to the northwest (Hammes et al., 2016).

The lower part of the upper Eagle Ford (upper Eagle Ford 1 [UEF1]) is an organic-rich calcareous mudstone. Based on GC-1 core observations, abundant globigerinid foraminifera are present and dispersed throughout the groundmass of the mudstone and as distinct laminations. The variable distribution of globigerinids in this interval may be indicative of fluctuations in volume of fine particles settling through the water column and/or fluctuations in the strength of bottom-water currents. This indicates deposition under oxygenated bottom-water conditions. Core and log findings indicate that the upper part of the upper Eagle Ford (upper Eagle Ford 2 [UEF2]) is a laminated mudstone containing the lowest average TOC and the highest average silica content of the upper Eagle Ford Group. The UEF2 is medium to dark gray mudstone with white globigerinid-rich laminations and occasional bivalve fossils. According to Dawson (2000), the mudstones of the upper part of the Eagle Ford in the Waco area contain highstand deposits of middle to upper Turonian age.

Geophysical log acquisition

Geophysical logging measurements (Figure 3a) were acquired at high vertical resolution 0.1 ft (0.03 m) depth increments instead of industry-standard 0.5 ft (0.15 m) depth increments. Nuclear and electrical geophysical measurements were obtained during the first logging run; acoustic and spectral GR measurements were acquired during the second logging run. Raw curve data, tools specifications, and acquisition parameters are provided in Burke et al. (2021). Adverse borehole conditions measured by the caliper log from 361 to 420 ft (110–128 m) over the Pepper resulted in unreliable wireline measurements for the density (RHOB), density porosity (DPHI), sonic transit time (DT), and sonic porosity logs (SPHI). Due to tool physics, geophysical logs for the GR, spontaneous potential (SP), 16N shallow and 64N deep resistivity (SN and LN), lateral induction and deep lateral induction (LAT and ILD), neutron porosity (NPHI), and temperature (TEMP) are unaffected. A water-filled limestone matrix density of 2.71 g/cm3 was used to calibrate density logging of all formations penetrated by this borehole.

Geochemical core analysis

Core mineralogy was determined using Fourier transform infrared spectroscopy (FTIR) on samples taken at approximately 2 ft intervals. Details on the collection of FTIR spectra are presented elsewhere (Washburn and Birdwell, 2013). Results from FTIR analysis include total clay minerals, total carbonate, and TOC content based on integration of different spectral regions; siliciclastic content was calculated by a difference of 100%. Bulk mineral categories and TOC contents determined by FTIR were independently verified using major element chemistry, Laboratory Equipment Corporation TOC, and X-ray diffraction (XRD) analyses (Birdwell et al., 2017; Birdwell and French, 2019). The FTIR results were used to calibrate the logs because FTIR data were available much sooner after sample collection than the TOC or XRD data sets. Because comparison of results between FTIR and XRD mineralogy was in agreement within the ±5 wt% error of the XRD measurements (Birdwell and Wilson, 2019), there was no reason to update the petrophysical multimineral models using mineralogy from XRD data.

Cluster analysis

The lithology log (Figure 1b) was created using Bayesian finite cluster analysis (Ellefsen and Smith, 2016) of multivariate geochemistry data involving infrared spectra collected on core samples. All spectra and absorbance wavenumbers (137 samples × 2542 wavenumber channels) were included in the analysis performed using a clustering method with squared Euclidean distance option. Four clusters were determined to be representative of the sample set, which sorted the samples in a way that is consistent with geochemical analyses (Birdwell et al., 2017; French et al., 2019). Spectra within each cluster were averaged and lithologic descriptions were determined based on the distribution of absorbance intensities for carbonate, clay minerals, and organic matter.

Core plug porosity analysis

Standard cylindrical core plugs oriented orthogonal to the wellbore axis were selected to study various lithofacies visible on the whole core. In total, 40 core plugs were taken at irregular intervals down the Eagle Ford section of the core. Core analysis included bulk density, bulk volume, and porosity and were determined by Core Labs in Denver, Colorado, using industry-standard Gas Research Institute (GRI) laboratory procedures (Guidry et al., 1995) for as-received and Dean Stark (Dean and Stark, 1920) methods. Each 300 g sample was crushed into a powder and sieved through 20- and 35-mesh screens. Bulk volume was measured by mercury immersion. Grain volume was measured using standard core analysis techniques (American Petroleum Institute, 1998) at ambient conditions using a double-chamber Boyle’s law porosimeter (model PORG-200) with helium as the expansion gas. Porosity was calculated from the difference between sample bulk volume and direct measurement of grain volume on crushed and dried samples. Using the GRI methods for estimating permeability in tight rocks was not feasible for these high-porosity shallow mudstones (Core Labs, personal communication, 2016).

Multimineral petrophysical models

Mineral components are widely used to describe sedimentary rocks in petroleum systems. Ternary diagrams displaying total carbonate, total clay, and total silicates are commonly used to differentiate mudrock types (Macquaker and Adams, 2003; Passey et al., 2010; Evenick and McClain, 2013; Chermak and Schrieber, 2014; Evenick, 2016; Donovan et al., 2017), and TOC is a critical property in petroleum systems for differentiating source rocks from other sediments (Schmoker, 1979, 1980; Schmoker and Hester, 1983; Fertl and Chilingar, 1988; Passey et al., 1990). Although well-log responses do not uniquely correspond to specific mineral compositions, multivariate modeling methods, which are often proprietary, can be used to determine mineral components (Figure 3b). Recent advancements in multimineral modeling for shale gas evaluation highlight the complexities of these unconventional resources (Quirein et al., 2010; Sondergeld et al., 2010; Ramírez et al., 2011; Salazar et al., 2017; Newsham et al., 2019).

The resultant four-components mineral models developed for this wellbore provide estimates of TOC, measured in weight percent (wt%) and converted to volume, volume of clay minerals (VCLAY), volume of silica including quartz, K-feldspars, and plagioclase (VQFP), and volume of carbonates (VCARB). The model for VCARB is a dependent parameter and is algebraically computed from a linear combination of the three other components, as volume fractions, subtracted from unity. Explicitly, this relation is given by
VCARB=1(VCLAY+TOC+VQFP).
(1)
A deterministic matrix inversion approach (Doveton, 1994) was used to independently solve a series of fitting equations for TOC, VCLAY, and VQFP. Equations are calibrated with mineral and TOC data obtained from FTIR analysis of core samples. Solutions are iterative until the sum of the squares between the computed and measured grain density is minimized. Fitting equations for the models are discussed in the next section.

TOC model

Various methods to calculate TOC from logs were investigated. These approaches included fitting equations using GR log response (Howell and Forsch, 1939; Asquith and Gibson, 1982; Fertl and Chilingar, 1988), logarithm of the resistivity response (Passey et al., 1990), formation density techniques (Schmoker, 1979, 1980; Schmoker and Hester, 1983), DT (Wyllie et al., 1956), and multivariable linear regression. The method that minimized the difference between core-derived and log-derived TOC values was found to be a linear relationship between the compensated density log (RHOB) and core-derived TOC (Figure 4a) with a standard error of 1.9689 wt%. This relation to estimate TOC from the logs is
TOC=26.8070*RHOB+62.2119.
(2)

Core-derived TOC is commonly measured in the laboratory in weight percent. To upscale to bulk volume as seen by logs, core-derived TOC was converted from weight percent to grain volume percent, and finally to bulk volume percent (Cluff et al., 2014). A density of 1.05 g/cm3 for thermally immature, relatively uncompacted organic material was used in this calculation, which falls within the anticipated range for hydrocarbon-rich or oil-prone type II organic matter (Tyson, 1995). This ensures that all four mineral components sum to 100%. However, the final model displays TOC in weight percent as this is the standard unit of measure for this parameter.

Petrophysically derived mineral fraction models

Bulk densities from as-received core plugs were compared to wireline log bulk density measurements (Figure 4b). Values that plot above the one-to-one parity line indicate the core imbibed fluid during the drilling and recovery process, and values below this line indicate core desiccation during handling (Cluff et al., 2014). Most data fall below the line, indicating that wireline bulk density curves include in situ pore fluids that subsequently evaporated from the recovered core.

Mineral fractions for clay and silica volumes were each determined independently. For this wellbore, the neutron-density separation curve was used to constrain the petrophysical model for the clay volume (VCLAY). This linear relationship (Figure 4c), which minimized the sum of the squares and is also representative of the data within one standard deviation, has a standard error of 0.1225 vol%, and is given by the relation:
VCLAY=0.0267*(NPHIDPHI)+0.2034,
(3)
where NPHI and DPHI are neutron porosity and density porosity in percent, respectively. In a similar fashion, the petrophysical model for the volume of quartz and feldspar (VQFP) was calculated from a multivariable linear regression, which minimized the sum of the squares of the errors, and is given by
VQFP=1.280+1.550×104*GR+1.530×102*NPHI+4.585×101*RHOB,
(4)
where GR is the gamma-ray curve, NPHI is neutron porosity in percent, and RHOB is density in units of g/cm3. Interestingly, the contribution from sonic velocity in this near-surface data set resulted in a negligible influence on regression modeling and was not used to obtain final modeling results for VQFP.

Crossplots showing correlations between core-derived and log-derived mineral fractions are presented in Figure 5. Trendlines show slopes generally near unity and R2 values between 0.55 and 0.75. Ideally, data should coincide with the parity line with a slope of unity; however, covariance among these variables results in a lower agreement than would be expected with fewer variables. According to log-based multivariate mineral modeling of a mudstone reservoir, positive and negative Pearson correlation coefficients range from −0.86 to +0.67, which provide sufficient precision for well-log interpretation (Table 1 and Figure 3; Huang et al., 2015).

The multimineral model reasonably predicts TOC (in wt%) as well as volumes of carbonate, clay mineral, and silica content for the Eagle Ford Group overall. However, the model shows slight discrepancy from core-derived calibration for TOC over the upper portion of the UEF2 chemofacies identified previously in this core (Birdwell et al., 2017; French et al., 2019). This may be due, in part, to the relatively higher densities and lower TOC values in this interval as compared to the other Eagle Ford chemofacies. Fitting equations (Figure 4a) for TOC were optimized for the lower density organic-rich mudstones of the Eagle Ford Group overall and not specifically optimized for the more siliceous UEF2.

These modeling results are consistent with geochemical and mineralogical characterization performed on core samples from this wellbore (Birdwell et al., 2017; French et al., 2019), and with a recent study of the LEF in the play area (French et al., 2020); XRD and major element chemistry data are also publicly available (Birdwell and French, 2019).

Comparisons of mineral components and TOC for mudstones of the Eagle Ford Group, Pepper Shale Member, and Del Rio Formation are summarized in Table 1. The LEF, which corresponds to the main organic-rich source rock in the thermally mature Eagle Ford play area, has the highest average TOC (4.7 wt%), highest average carbonate content (64.1 vol%), and lowest average clay volume (14.8 vol%). The UEF1 exhibits high volumes of carbonate (average of 60.2 vol%) and intermediate TOC values (average of 2.5 wt%), but the lowest silica content (average of 19.3 vol%). The UEF2 exhibits the largest VCLAY (42.6 vol%) and the lowest TOC (1.4 wt%) among the Eagle Ford mudstones. The Del Rio is an organic-lean (average 0.5 wt% TOC) clay-rich (average 42.2 vol%) mudstone. Ternary diagrams and mineral histograms for visualization and distribution of the data are shown in Figure 6 and discussed in the next two sections.

Log-derived ternary diagrams and mineral histograms

The resultant log-derived ternary diagrams (Figure 6a) and histograms of data distributions (Figure 6b) were developed using petrophysical multimineral model results to categorize lithology. Based on mudstone nomenclature designated in the literature (Evenick and McClain, 2013; Evenick, 2016; Donovan et al., 2017), the UEF2 interval is classified as a silica-rich argillaceous to mixed carbonate mudstone. The UEF1 and LEF intervals are classified as carbonate mudstones. The Del Rio is classified as an organic-lean argillaceous carbonate mudstone. The Pepper (not shown) is categorized as an argillaceous mudstone based on core-derived mineralogy only. Laminae-scale mineralogical heterogeneity within the mudstone intervals, as afforded by the high-resolution logging suite, is interpreted as colored polygons.

Petrophysical characterization of these thermally immature, relatively uncompacted Eagle Ford mudstones in central Texas from shallow depths provides information about lithology, porosity, pore-fluid identification, and fluid saturation. The multimineral models provide a detailed quantification of mineralogy at each depth increment, which greatly enhances basic lithology characterization afforded by the natural GR curve. Beyond the characterization of traditional petrophysical parameters, these mudstones were further described in terms of pore-fluid salinity, permeability, mechanical integrity, and stress regime. Earth surface heating is also identified in this near-surface borehole. This near-surface wellbore has reservoir temperature, pressure, and stress conditions that closely resemble standard temperature and pressure laboratory conditions; thus, static versus dynamic properties are essentially equivalent.

Comparison of wireline logging measurements and select rock properties is summarized in Table 2 for the five mudstone intervals penetrated by this study wellbore. Of note, the UEF2 argillaceous mudstones exhibit the highest Eagle Ford grain density at 2.269 g/cm3, and the multimineral analysis (Table 1) over this interval confirms smaller volumes of low-density, organic-rich materials and larger volumes of silica minerals. The UEF1 exhibits a high degree of heterogeneity due to thinly interbedded laminae of carbonate cement amid clay-rich mudstone layers. Relative to the UEF2 above and the LEF below, the UEF1 carbonate mudstone has intermediate mineralogical properties for clay minerals, carbonate, and TOC content. The LEF mudstone exhibits the highest GR response and lowest grain density, which is due to the higher amounts of organic matter (average TOC is 4.7 wt%; Table 1) in this interval. The average caliper diameter for the entire Pepper is elevated, based on the presence of washouts present in this mudstone interval. Core-derived mineralogy from bulk samples reveals high clay content in this mudstone. According to these analyses, the Del Rio is an organic-lean clay-rich carbonate mudstone with low resistivity and high porosity.

Porosity characterization

Four log-derived models and two core-derived models were used to quantify mudstone porosity (Figure 7) in the upper and lower parts of the Eagle Ford. Each logging tool is sensitive to different facets of porosity and rock matrix properties. According to the logs, the UEF2 has the highest average neutron porosity of 35.0% and ranges from 18.3% to 50.8% (Figure 7a). The UEF1 has an average porosity of 27.8% and ranges from 7.6% to 47.1%. The LEF has an average porosity of 30.7% and ranges from 12.8% to 49.0%.

Density logging measures atomic electron density within the probed volume of formation and, through a mathematical transform, is related to bulk density (Serra, 1984). Bulk density ρb is a function of matrix density ρma, fluid density ρfl, and porosity Φ as
ρb=ρma(1Φ)+ρflΦ.
(5)
The UEF2 has the lowest density porosity, averaging 26.0%, with a range from 16.6% to 35.3% (Figure 7b). The UEF1 has an intermediate density porosity of 26.7%, with a range of 14.2%–46.5%. The LEF has the highest density porosity, averaging 33.1%, with a range of 16.4%–50.0%.

Neutron-density porosity was calculated from an arithmetic average of these two log responses, essentially equalizing overestimation of porosity from the neutron response and underestimation of porosity from the density response (Dewan, 1983). Neutron-density porosity (Figure 7c) of the UEF2 is 30.5% with a minimum of 18.9% to a maximum of 40.6%. The UEF1 has an average neutron-density porosity of 27.2% with a range of 11.6%–39.8%. The LEF has an average neutron-density porosity of 31.9% and ranges from 17.5% to 44.0%.

Sonic porosity models (Figure 7d) were calculated from the Wyllie time averaged equation (Wyllie et al., 1956). That relation, in terms of compressional-wave DT, Δtc is given by
Δtc=(1Φ)Δtma+ΦΔtfl
(6)
for Δtma and Δtfl for matrix and fluid transit time, respectively. The UEF2 exhibits the highest sonic porosity with maxima at 29% and 33% (range of 28.1%–35.3%), whereas the UEF1 and LEF have similar sonic porosity characteristics with maxima at 26% and 29%, respectively. UEF1 sonic porosity ranges from 14.3% to 33.8%, and the LEF sonic porosity ranges from 21.9% to 36.7%. Based on log and core observations, the sonic porosity bimodal distribution is attributed to the presence of thin laminations of carbonate cement that are below the resolving power of the sonic logging tool.

Measured porosity of core plugs indicate that as-received porosity (Figure 7) ranged from 9.1 vol% to 25.3 vol%, with a peak frequency of 21.0 vol%. Dean Stark processed core samples (Figure 7f) exhibit porosities from 9.6% to 28.3 vol%, with a peak frequency of 22.5 vol%. Core plug sample locations were selected to characterize the different lithofacies visible on the whole core without benefit of core description or analytical results, which were available at a later date. Thus, core porosity values may not necessarily be representative of the porosity variation observed on the logs, over which continuous data were acquired at 0.1 ft (0.03 m) increments. In addition, the vertical resolution of core plug porosity is less than vertical resolution from log-derived porosity, resulting in a relatively small number of core plug samples in comparison to the large number of log-derived porosity values. Despite these differences, however, comparison of the distributions of core-derived and log-derived porosities (Figure 7g and 7h) shows reasonable agreement. As expected, the Dean Stark methods show a tighter spread than as-received core porosity.

Pore-fluid properties

Water saturation is nearly 100%, and no indications of free hydrocarbons are observed on the resistivity logs. Based on interpretation of the SP logs (Figure 3a), pore-fluid salinities slightly increase with depth and stabilize to a baseline in the lower part of the UEF2 and remain fairly constant to the bottom of the wellbore. Due to the shallow nature of this borehole, recent weathering may have enabled meteoric water to percolate down through the porous and fractured overlying Austin Chalk, thereby diluting the formation waters of the UEF2 and into the upper portion of the UEF1 (270 ft). The SP curve is stable and invariant down through the LEF, Pepper, and Del Rio units, indicating that formation waters of these deeper mudstone strata remain relatively less altered from surface water infiltration.

Permeability

Qualitative estimation of permeability and diameter of the invasion zone can be made from the magnitude of separation between the shallow and deep resistivity curves (Crain, 2004). The relative magnitude of this separation was used to qualitatively assess and rank permeability of each mudstone unit. All mudstones exhibit minimal separation of the shallow and deep resistivity curves, which indicates low permeability and a very shallow invasion zone over these intervals. Of the mudstones of the Eagle Ford, the UEF1 exhibits the highest permeability, followed by intermediate permeability in the LEF. The UEF2 exhibits the lowest permeability that is likely attributed to the higher clay content (42.6 vol%) relative to the other mudstone intervals in the Eagle Ford Group. Similarly, resistivity separation over the Pepper is minimal and interpreted as characteristic of low permeability in this clay-rich mudstone. In these carbonate mudstones, an increase in clay mineral volume is associated with a reduction in permeability.

Mechanical integrity and formation brittleness

Mechanical integrity of the mudstones penetrated by this borehole was qualitatively assessed using the caliper, density, and optical televiewer data sets in conjunction with mineralogy results and core corroboration. Findings indicate that mudstones of the Eagle Ford with higher carbonate and/or silica content exhibit the highest degree of mechanical integrity. Mudstones with higher silica content, from highest to lowest, are the UEF2, LEF, and UEF1. The mudstones with higher carbonate content, from highest to lowest, are the LEF, UEF1, and UEF2. Mudstones with higher combined silica and carbonate content, from highest to lowest, are the LEF, UEF1, and UEF2. Higher volumes of clay and volcanic ash layers correspond to reduced mechanical integrity; these findings are consistent with observations by other researchers (Tinnin et al., 2014; Tenorio, 2016; Alvarez and Vera de Newton, 2017). For the case of the organic-rich LEF interval, higher TOC does not appear to adversely affect mechanical integrity, rendering this interval the best candidate for hydraulic fracturing. The organic-lean clay-rich carbonate mudstones of the Del Rio exhibit intermediate mechanical integrity and a competent borehole wall. The lowest mechanical integrity is observed in the clay-rich noncalcareous Pepper.

Formation brittleness is inferred from the mechanical integrity findings. Findings indicate that mineralogy is the primary factor influencing brittleness of the mudstones in this wellbore. Higher volumes of carbonate or silica minerals correspond to an increase in brittleness, and higher volumes of clay correspond to an increase in ductility. Given this, the LEF and UEF1 are the optimal candidates for hydraulic-fracture placement in areas where these intervals are thermally mature and productive of hydrocarbons.

Geomechanics and principal stress regime

Evidence of principal stresses, as described by Zoback et al. (1985), was considered but not observed in either the three-arm caliper log or optical televiewer data. Based on these observations, horizontal minimum and maximum stresses are interpreted to be minimal or directionally uniform in this near-surface wellbore environment.

Earth surface heating

The average temperature of the near-surface section of the borehole down to 30 ft (9 m) in depth is 77.1°F (25.1°C), which is consistent with mean earth temperatures compiled for the United States and North America (NASA, 2020). Temperature gradients range from 1.32 to 1.93°F/100 ft (17.4°C/km) across the onshore Gulf Coast region (Burke et al., 2018, 2020). The temperature gradient in this near-surface wellbore is 1.87°F/100 ft (16.3°C/km), which is within the range of anticipated values for the Gulf Coast, but it is likely influenced predominantly by earth surface heating rather than burial.

The multimineral models enabled lithology characterization of the mudstone intervals. Based on the well-log analyses, the UEF2 is a silica-rich argillaceous to mixed carbonate mudstone with the highest silica and clay mineral content of the Eagle Ford Group in the Waco area. Increased silica content gives this unit the highest average density and the fastest average sonic traveltime compared to other mudstones of the Eagle Ford. Most of the carbonate content in the UEF2 is located at the base of the unit. The UEF1 is a highly laminated carbonate mudstone with discontinuous and discordant foraminiferal-rich layers. These carbonate laminations give this interval the lowest average porosity compared to other mudstones of the Eagle Ford. The LEF is an organic-rich carbonate mudstone with the highest carbonate content and lowest clay mineral volume of these mudstones. Bentonite clay layers are visible in the core, but they are below the resolution of wireline logging measurements. The LEF has the highest natural GR measurements and the highest average porosity. Density measurements in the LEF are a more reliable indicator of TOC compared to natural GR response. The Pepper is a finely laminated mudstone with the lowest average resistivity compared to all mudstones in this study. Adverse borehole conditions precluded several petrophysical characterizations of this mudstone and instead relied on core-derived mineralogy, which indicates high clay content, low TOC, and low carbonate content in the Pepper (Figure 3b, magenta circles). The Del Rio is a nonorganic rich argillaceous carbonate mudstone. Carbonate material is dispersed throughout the groundmass of the clay-rich mudstone, effectively reducing the porosity. The Del Rio exhibits the lowest porosity of mudstones investigated in this borehole. The relatively high GR and resistivity readings are in response to increased clay mineral volume and are not an indicator of organic richness in this wellbore as traditional formation evaluation methods suggest (Howell and Forsch, 1939; Schmoker, 1979, 1980; Schmoker and Hester, 1983; Fertl and Chilingar, 1988; Passey et al., 1990).

Portions of the Eagle Ford section with low porosity and high volumes of total carbonate are likely intervals with abundant carbonate cement. Typically, carbonate cement in limestone forms or develops under early and relatively shallow burial conditions (Schieber et al., 2016). Observation of Eagle Ford thin sections reveals that calcite cement is commonly associated with clusters or aggregates of foraminifera, coccoliths, and other carbonate skeletal debris concentrated along bedding laminations (Macquaker et al., 2014; McAllister et al., 2015; Schieber et al., 2016). The association of carbonate cement with carbonate grains indicates that carbonate grains serve as nucleation sites for the precipitation of calcite cement. Moreover, local dissolution of carbonate grains may serve as a calcite cement source, particularly if some of the biogenic grains were aragonite at the time of deposition. Mudstones with an abundance of biogenic carbonate and silica content at the time of deposition exhibit calcite and quartz cementation, which obstruct pore spaces and increase stiffness of the rock matrix (Milliken et al., 2016). In contrast to the Eagle Ford, portions of the Del Rio with low porosity, based on the log analysis, contain little carbonate, but they tend to have high volumes of clay minerals. Therefore, low porosity in the Del Rio, when carbonate is absent, indicates that low porosity is due to physical or mechanical compaction of clay in response to increasing overburden pressures during burial.

Petrophysical characterization of thermally immature source rocks in central Texas from shallow depths of less than 1000 ft is important for linking modern outcrop work to subsurface petroleum system studies that are reliant upon geophysical logging responses. High-resolution mineralogy models were developed using a deterministic workflow on a conventional suite of geophysical logs and calibrated with core analysis to quantify TOC and mudstone mineralogy in thermally immature, water-bearing, relatively uncompacted source rock strata in which traditional petrophysical workflows failed to accurately predict organic richness and mineralogy. This study determined that density log response, not GR, is a reliable indicator of organic content in these thermally immature, near-surface source rocks exhibiting high clay content. This workflow also provided a transparent nonproprietary method to build log-derived multimineral models applicable to any geophysical logging data set with core calibration.

According to model results, the LEF contains the highest average TOC content (4.7 wt%) and the highest average carbonate volume (64.1 vol%) of all Eagle Ford Group mudstone intervals, indicating that this zone has higher source-rock potential and hydraulic-fracture efficiency. The upper portion of the Eagle Ford as well as the Del Rio contain the highest VCLAY (averaging 42.6 vol% and 42.2 vol%, respectively), which increases the potential for wellbore stability issues during the drilling process and ductility issues during hydraulic fracturing.

In addition to mineralogical investigation, petrophysical characterization revealed mudstone porosity of approximately 30%, water saturation near 100%, and an absence of free hydrocarbons in the pore spaces. Based on the three-arm caliper log, no wellbore strain was observed, and horizontal principal stresses are presumed to be directionally uniform in the vicinity of this near-surface wellbore. The temperature gradient is 1.87°F/100 ft (16.3°C/km), which is within the range for the onshore Gulf Coast region, and it is more likely to be driven by earth’s surface heating than thermal effects of burial.

The authors wish to acknowledge the USGS Gulf Coast Petroleum Systems Team, the USGS Central Energy Resources Science Center, the USGS Petroleum Geochemistry Research Laboratory, and the USGS Core Research Center for access to and permission to show the USGS Gulf Coast 1 West Woodway core, geophysical well logs, core photographs, and geochemical data. We wish to acknowledge S. Cluff and R. Sharma for collaboration efforts on calculating the multimineral models. Authors are thankful for USGS reviews by K. Whidden, L. Ruppert, J. Herrick, K. Marra, O. Pearson, and J. Slate, as well as anonymous reviews which led to improvements to this paper. Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Research was funded by the USGS Energy Resources Program. All data associated with this paper are shown in the figures and tables herein or are available in the associated USGS data release.

Data associated with this research are available and can be accessed via the following DOI: https://doi.org/10.5066/P9KZVG7P.

Biographies and photographs of the authors are not available.

Freely available online through the SEG open-access option.