Deep resistivity “turnover” effect at oil generation “peak” in the Woodford Shale, Anadarko Basin, USA

The Devonian Woodford Shale in the Anadarko Basin is a highly organic, hydrocarbon source rock. Accurate values of vitrinite reflectance (Ro) present in the Woodford Shale penetrated by 52 control wells were measured directly. These vitrinite reflectance values, when plotted against borehole resistivity for the middle member of the Woodford Shale in the wells, display a rarely reported finding that deep resistivity readings decrease as Ro increases when Ro is greater than 0.90%. This phenomenon may be attributed to that aromatic and resin compounds containing conjugated pi bonds generated within source rocks are more electrically conductive than aliphatic compounds. And aromatic and resin fractions were generated more than aliphatic fraction when source rock maturity further increases beyond oil peak. The finding of the relationship between deep resistivity and Ro may re-investigate the previously found linear relationship between source rock formation and aid to unconventional play exploration.


Introduction
In a lot of early-stage unconventional shale play explorations, several key geochemical parameters, including organic richness (TOC) and thermal maturity, are need to be known to evaluate potential unconventional shale play (Comer 2005;Jarvie 2011). This study focuses on hydrocarbon source rock thermal maturity, which is scaled by vitrinite reflectance (R o ) in practice. It is particularly significant to map those potential shale play targets' organic matter thermal maturity in 3-D geological framework if to target at shale oil, because the most appropriate maturity range of many successful shale oil plays is centered on 0.90% R o based on the successful cases of shale oil plays in North America (Comer 2005;Slatt et al. 2009a;Jarvie 2011). The reason behind these successful experiences would be source beds' maturity control hydrocarbon phase and mobility. In other words, organic maturity drives the success of shale oil play to some extent. But in most cases, cost of the analysis and insufficient well cuttings or cores generally restrict the interpretation of a basin's maturity to the contouring of a relatively small number of vitrinite reflectance values over a large area. Another successful experience from the shale oil "boom" in North America is that the economically profitable shale oil "target" is usually the "old" source rock beds of the conventional petroleum systems (Loucks and Ruppel 2007;Singh 2008;Comer 2008;Slatt et al. 2009a, b). The Woodford Shale of Late Devonian to Early Mississippian age is a typical resource play target of these types of shale plays. It has not only proven to be an excellent source rock, charging the conventional reservoirs in Kansas and Oklahoma (Comer and Hinch 1987;Burruss and Hatch 1989;Philp et al. 1989;Jones and Philp 1990;Comer 1991b;Wang 1993;Wang and Philp 1997), but also become a frontier for unconventional Edited  resource play exploration and production (Cardott 2014;Kvale and Bynum 2014). The unconventional Woodford resource play in Oklahoma encompasses four regions, namely Anadarko-Woodford, Arkoma-Woodford, Nemaha-Woodford and Southern Oklahoma-Woodford, and has been estimated to contain 0.24 × 10 12 ft 3 (6.8 × 10 9 m 3 ) of natural gas and 70 × 10 9 bbl (11 × 10 9 ton) of oil, in place and potentially producible, on the basis of mass balance calculations, indicating a huge potential as an unconventional hydrocarbon production target (Comer 2005). In general, well control is better than seismic data control in terms of vertical formation resolution. Therefore, to predict key geochemical parameters, especially organic richness and maturity, from wireline logs is of practical significance from the perspective of unconventional shale play exploration. In other words, it is necessary to revisit "old" well logs to assess several key geochemical parameters based on the "old" well logs. The endeavor to look for a relationship between wireline logs and geochemical parameters if any and apply that relationship to predict geochemical parameters without measuring core data but simply based on wireline logs has been years (Meissner 1978;Smagala et al. 1984;Zhao et al. 2007;Passey et al. 2010). Meissner (1978) noticed that the Bakken Shale in the mature areas of the Williston Basin had much higher resistivity than in the immature areas. This was also noticed by Goff (1983) working with the Kimmeridge Clay in the northern North Sea of the UK. Meyer and Nederlof (1986) to some extent combined the observations of the above researchers by producing density versus resistivity (and sonic versus resistivity) cross-plots in order to separate source rocks from non-source rocks on a global scale. Smagala et al. (1984) found a linear relationship between resistivity and R o that allowed the mapping of organic maturation levels over a large part of the marine siliciclastic basin through the use of a large number of available electric logs. Passey et al. (2010) reported that in some shale-gas reservoirs that are at very high maturities (R o > 3), the overall rock resistivity can be 1-2 orders of magnitude less than that is observed in the same formation at lower thermal maturities (R o between 1 and 3) perhaps because the carbon in the organic matter is recrystallizing to a precursor of mineral graphite, which is electrically conductive. The work reported hereby was a "by-product" of the research which initially aims to derive quantitative relationships between wireline logs and geochemical parameters if any and to develop a scheme for calculation of source rock richness and maturity from wireline logs which can be applied on a "global" basis. Here, we display a rarely reported finding that deep resistivity readings decrease as R o increases when R o is over 0.90%. The new findings on the relationship between deep resistivity and R o may re-investigate the previously found linear relationship between resistivity and R o of source rock formation.

Geological settings of study area and data sets
The Woodford Shale of Late Devonian to Early Mississippian age is an organic-rich black shale widely distributed over the southern Midcontinent from the Iowa Basin in Kansas to the Permian Basin in West Texas (Comer and Hinch 1987;Comer 1991b). It was found to be distributed in most of Oklahoma, including the Anadarko Basin, the Anadarko Shelf, Cherokee Platform and the Arkoma Basin. In the early Paleozoic time, three major tectonic/depositional provinces existed in Oklahoma: the Oklahoma Basin, the southern Oklahoma Aulacogen and the Ouachita Trough. The Oklahoma Basin, initially formed during the continental breakup in the Late Precambrian (Miall 2008), was a shelflike area that received widespread and thick shallow-marine carbonates interbedded with thin marine shale and sandstones (Johnson 1989;Northcutt et al. 2001). The southern Oklahoma Aulacogen, a west-northwest-trending trough derived from one of the failed rifts during the breakup of the supercontinent Rodinia (Miall 2008), was the depocenter for the Oklahoma Basin and the precursor of the Anadarko Basin (Johnson 1989;Northcutt et al. 2001). The Ouachita Trough received deep-water sediments deposited along a rift located in the southern margin of the North American craton (Johnson 1989;Northcutt et al. 2001). From Silurian to Middle Devonian clean-washed skeletal limestone, argillaceous and silty carbonates, referred to as the Hunton Group in Oklahoma, were deposited in a shallow marine setting (Northcutt et al. 2001). Epeirogenic uplifts interrupted deposition, resulting in two regional unconformities. One unconformity arose during pre-middle Early Devonian (pre-Frisco-Sallisaw unconformity) and the second one during pre-Late Devonian (pre-Woodford-Chattanooga unconformity; Johnson 1989). In southern Oklahoma, the pre-Woodford-Chattanooga unconformity eroded to the Upper Ordovician, and in northern Oklahoma, the erosion sculpted out Upper Cambrian-Lower Ordovician rocks (Kirkland et al. 1992). The Woodford Formation in the Anadarko Basin was thought to be deposited in a restricted basin developed in an epeiric sea within the passive margin (Amsden 1975;Wang and Philp 1997;Kim and Philp 2000;Blakey 2008;Comer 2008;Haq and Schutter 2008;Miall 2008). On the Cherokee Platform, the Woodford Shale was deposited on a major regional unconformity developed during the Late Devonian (Amsden 1975). It is conformably overlain by limestone and shale of Early Mississippian age (Fig. 2). The predominant lithology of the Woodford Shale is black shale. Other common lithologies include chert, siltstone, sandstone, dolostone and light-colored shale (Amsden 1967;Amsden 1975;Comer 1991b). A typical core from the Woodford can contain 30%-50% quartz, 0%-20% calcite/dolomite, 0%-20% pyrite and 10%-50% total clay, a variance in mineralogy that occurs on a regional scale and within the stratigraphic section. These differences can have an effect on the porosity and permeability of the interval as they are reported to range from 3% to 9% and 100 nd-0.001 md, respectively (Comer 1991a, b).
A number of Woodford core samples from 52 wells and corresponding wireline logs (GR, SP, resistivity, density, etc.) provided by ConocoPhillips were collected for this study (well location shown in Fig. 1). Accurate values of vitrinite reflectance (R o ) present in the Woodford Shale penetrated by 52 control wells were measured directly. The specific details concerning analytical techniques are summarized in Sect. 2.2.

Analytical Methods
Woodford Shale section was identified and subdivided into three informal stratigraphic members: upper, middle and lower Woodford members (Amsden 1975;Slatt et al. 2009a) from GR, SP, RHOB (density log) and deep resistivity logs based on the Woodford electronic log characteristics in the study area (Hester et al. 1988(Hester et al. , 1990 Table 1); A-A′ cross section is to display the representative log characteristics of three Woodford members (log characteristics shown in Fig. 2); "MWF Ave. R o " represents "Average R o value of the middle Woodford member"; data points are classified into three groups whose geographic location is separated by dashed lines 1 3 (R o ) values from all of the Woodford cores were obtained from the organic petrographic pellets (either made from a whole rock or a kerogen concentrate if it is not easy to look for good vitrinite from a whole rock) prepared at the Oklahoma Geological Survey Organic Petrography Laboratories in Norman, Oklahoma, and measured at the University of Oklahoma Organic Geochemistry Laboratories by the author of this study. Wireline logs were provided by Mr. David Jacobi from the Geology & Geophysics & Reservoir Engineering Department of ConocoPhillips Company, Houston. Wireline logs were loaded onto Schlumberger Petrel and treated to calculate geometric average deep induction log (ILD) values for the middle Woodford member (M-WF) for each well. (2) lower bulk density (2.2-2.4 g/cm 3 ); (3) lower SP; (4) shorter DT (sonic log) reading; and (5) extremely high GR (over 300 API) (Hester et al. 1988(Hester et al. , 1990. Pristane and phytane (Pr/Ph) and biomarker ratios suggest the establishment of stronger anoxic conditions during deposition of the middle Woodford member than the upper Woodford member, where the latter may have received an additional siliciclastic organic matter input (Miceli and Philp 2012;Wang 2016;Wang and Philp 2019). In the Cherokee Platform in the proximity of the Nemaha Uplift, Pr/Ph ratios indicate deposition under suboxic to dysoxic conditions for the Woodford Shale interval analyzed (Wang 2016;Wang and Philp 2019). Based on these previous studies, the middle Woodford member was used to represent the whole Woodford section in this study.   Fig. 1 By cross-plotting sample depth versus corresponding vitrinite reflectance (R o ) values, middle Woodford member rock samples from the studied wells in the Anadarko Basin (Fig. 3) show a very good linear relationship between measured depth and R o values (in logarithm scale). This finding is consistent with that found on the Niobrara "K" zone organicrich calcareous shale by Smagala et al. (1984). Lockridge and Scholle (1978) noticed a reduction in porosity with increased burial depth for the Niobrara chalks in the Denver Basin. Smagala et al. (1984) attributed the co-increases in resistivity of the chalk with burial depth to the consequence of the porosity reduction with increased overburden according to the empirical Archie equation (Archie, 1942). This finding is consistent with that the Woodford section in the Anadarko Basin was undertaking a "simple" burial history without multiple stages of subsiding and uplifting (Amsden 1975;Johnson 1989;Hester et al. 1988Hester et al. , 1990Comer 1991bComer , 2005Comer , 2008, which may lead to that both the burial depth and vitrinite reflectance "recorded" the thermal stress applied on the rock (Pepper and Corvi 1995). In Fig. 4, the starting point for the Woodford Shale entering early oil window in the Anadarko Basin is around 35 Ω m, which is consistent with that reported by Schmoker and Hester (1990). The reason may be attributed to that there are many uncertainties in terms of relationship between resistivity and thermal maturity during the stage in which the source rock had not yet entered the oil window (Schmoker and Hester 1990). By plotting average formation resistivity readings of the middle Woodford member (geometric average value of ILD log readings of the middle Woodford member used in this study) versus average R o value of the corresponding middle Woodford member, it was shown that formation resistivity increases as maturity increases before the oil generation peak (R o ~ 0.85%) and decreases as maturity further increases beyond the oil generation peak.

Results and discussion
The resistivity of a rock is directly related to those components that are electrically conductive. In conventional reservoirs, formation water, usually thought to allow for ionic conduction, is the primary conductor of electricity. Low resistivity is observed when the amount of saline water filled in porosity is high-the larger the volume of formation water, the lower the resistivity of the fluid-filled rock. From the perspective of traditional reservoir petrophysics, which did not treat too much organic-rich components like kerogen, hydrocarbon fluids (oil or gas) are non-conductive, and when they are present in sufficient quantities, they displace the amount of formation water, resulting in resistivity values higher than the same rock fully filled with electrically conductive formation water (Archie 1942). There are many variants to the interpretation of resistivity in conventional reservoirs (e.g., clay conductivity and shaly sand analysis; thin-bed effects due to interbedded shale; Waxman and Smits 1968;Worthington 1985;Passey et al. 2006), but these are beyond the scope of the current paper. Based on previous consideration, a model was proposed to interpret the finding in this study. As source rock enters "oil window" approaching "oil peak" (R o less than 0.85%), it is generating greater amount of aliphatic compounds than aromatic and resin fractions. Aliphatic compounds are less conductive than aromatic and resin fractions since the latter contain a larger number of conjugated π-bonds, which allow delocalized electrons to flow through the π-bonds. However, as source rock maturity further increases beyond oil peak, it  Fig. 4 Relationship between the middle Woodford member resistivity and vitrinite reflectance (R o ) of the studied wells in the Anadarko Basin; numbers within the data points are well codes in Table 1 is generating a greater amount of aromatics and resins than aliphatic compounds, which leads to the decrease in formation resistivity.
In the resistivity versus R o diagram, the data points in the high-maturity area (beyond oil peak when R o is greater than 0.85%) are distinctly clustered into two groups and a good negative linear relationship between formation resistivity and R o values was displayed in each group (Fig. 4). By displaying the samples' location of three groups onto the Anadarko Basin structural province map, the wells of group 3 samples were shown to be adjacent to the depocenter (Fig. 1), where type II marine shale was deposited dominantly. The wells of group 2 samples approach the basin flank (Fig. 1), where type III terrestrial organic matter get a greater chance to be received. The Late Devonian (385 Ma) paleogeography map of the study area (Fig. 5) shows that the Woodford Shale was getting started to deposit in the ancient Oklahoma basin, which was originally derived from a failed rift formed back to Precambrian. During Late Devonian, it was a restricted basin within the passive margin (Johnson 1989;Northcutt et al. 2001;Miall 2008). The Anadarko Basin flank, compared to the Anadarko Basin depocenter, was more likely to receive fluvial terrestrial organic matters from the paleohighs from northeast (Blakey 2013;Blakey and Ranney 2018). Further investigating the resistivity versus R o diagram (Fig. 4), at the same maturity level, the data point of group 3 is less conductive than that of group 2, probably attributed to less aromatics/resins generated by group 3 type II marine kerogens than group 2 ones which contain a greater amount of type III terrestrial kerogens, resulting in more structural kerogens (producing a greater amount of aromatics/resins). Passey et al. (2010) found a similar phenomenon that resistivity decreases as source rock becomes mature as well, especially when R o goes beyond 3.0%. Passey et al. (2010) proposed an interpretation that perhaps the carbon in the organic matter is recrystallizing to the mineral graphite, which is electrically conductive. However, previous studies indicate that pure mineral graphite is not present in abundance at these thermal maturities. Thus, it is likely that a precursor to graphite is forming. It is sufficient to state that in extremely high-maturity organicrich rocks (R o > 3), the rock may be much more electrically conductive due to other mineral phases being present rather than solely formation water, clay and pyrite (as usually considered) (Passey et al., 2010).

Conclusions
As a famous traditional highly organic-rich hydrocarbon source rock, the Devonian Woodford Shale in the Anadarko Basin is used to delineate the relationship between wireline log responses and thermal maturity.  (2013) penetrated by 52 control wells were measured directly to obtain an average R o value of the middle Woodford member, which is the most organic-rich member of Woodford. The geometric average number of ILD log readings of the corresponding middle Woodford member was used as its formation resistivity. These vitrinite reflectance values, when plotted against borehole resistivity for the middle member of the Woodford Shale in the wells, display a rarely reported finding that an increase in R o values (source rock maturity scale) is positively related to increased resistivity in R o range of 0.60%-0.90% and decreased resistivity in R o range of 0.90%-2.00%. This phenomenon may be attributed to that aromatic and resin compounds containing conjugated pi bonds generated within source rocks are more electrically conductive than aliphatic compounds. And aromatic and resin fractions were generated more than aliphatic fraction when source rock maturity further increases beyond oil peak. The finding of the relationship between deep resistivity and R o may re-investigate the previously found linear relationship between source rock formation and unconventional play exploration.