Relationships between Sonic Compressional and Shear logs in ...
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
Relationships between Sonic Compressional and Shear logs in Unconventional Formations John Quirein*, Mehdi Far, Ming Gu, Deepak Gokaraju, and Jim Witkowsky; Halliburton *Corresponding Author Society of Petrophysicists and Well Log Analysts Copyright 2015, held jointly by the Society of Petrophysicists and Well Log Analysts (SPWLA) and the submitting authors. This paper was prepared for presentation at the SPWLA 56th Annual Logging Symposium held in Long Beach, CA, USA, July 18-22, 2015.
ABSTRACT Relationships between compressional and shear wave velocities were established in the classic paper by Castagna et al. (1985). In particular, linear trends for correlating the compressional and shear wave velocities are presented, including the Castagna mudrock line derived from in-situ sonic and seismic measurements. The work of Brie et al. (1995) extended the work of Castagna et al. (1985) to include predicting gas saturation from a Vp/Vs vs. compressional traveltime crossplot. In this crossplot, Brie proposes one non-linear trend for water-wet sands and another non-linear trend for shales. In addition, the Brie Vp/Vs vs. compressional traveltime crossplot has a greater sensitivity to low velocities or large compressional traveltimes. This paper compares some of the more famous rock physic trends, including the Castagna mudrock and Brie water-wet sands and shales, in both the Vp vs. Vs crossplot and the Vp/Vs vs. compressional traveltime crossplot. It also presents plots of laboratory data from the Bakken, Bazhenov, Barnett, Eagle Ford, Haynesville, Monterey, and Niobrara organic shales and descriptions of departures from the expected trends to the presence of kerogen and anisotropy.
are not commonly found in the literature. Passey (2010) points out that shales are more properly referred to as mudstones and exhibit a wide range in composition including clay, quartz, feldspar, heavy minerals, and carbonate. Moreover, the composition of a typical mudstone will vary much more than typical sandstones, even though, to the naked eye, many mudstones look similar. Passey (2010) presents an example of what he calls a “clay-rich gas-bearing mudrock” with the carbonate fraction varying from 0 to greater than 75%. In this paper, “Castagna mudrock” refers to clastic silicate rock composed primarily of clay or silt-sized particles, whereas “mudstone” refers to the more general case, which includes carbonate-rich shales. The Castagna mudrock equation suggests a linear relationship between the compressional velocity, Vp, and the shear velocity, Vs:
V p 1.16Vs 1.36
(1)
where the velocities are expressed in km/s. The mudrock equation provides a good fit for sonic and seismic data from the Devonian, Pierre, Grayson, and Japanese shales and an oil shale.
The paper also includes an investigation of in-situ well log data from the Eagle Ford, Haynesville, and Barnett formations in the context of the Castagna, Brie, and additional crossplots. The Eagle Ford data satisfies the carbonate Vp vs. Vs trend predicted by Castagna, and the Haynesville data clearly satisfies the hydrocarbon/kerogen effect predicted by the Brie Vp/Vs vs. compressional traveltime crossplot. INTRODUCTION Castagna et al. (1985) defined mudrocks as clastic silicate rock composed primarily of clay or silt-sized particles. Lithified muds are composed primarily of quartz and clay minerals. Owing to the difficulty associated with handling of most mudrocks, laboratory measurements on these rocks
Fig. 1 Vp and Vs for mudrocks (Castagna et al. 1985). In Fig. 1 the clay point is empirical and represents a dry clay mineral with no porosity. As porosity is added, the dry clay point moves towards the 100% water point at zero shear 1
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
velocity. The net result is that many quartz-clay-water ternary mixtures are spread along an elongate triangular region loosely defined by clay-water, quartz-clay, and quartz-water lines. Castagna concludes, to first order that shear velocity is linearly related to compressional wave velocity for water saturated rocks. For a given Vp, mudrocks tend toward slightly higher Vp/Vs than do clean porous sandstones.
trend in Figure 2, but the following empirically derived equation fits the trend:
Vs .0147V p2 .883V p .9886
(3)
A literature search reveals a large assortment of regressions relating the compressional and shear velocities. This study includes two from Han (1986). Han used ultrasonic data for 70 water-saturated shaley sand samples. The data are separated by clay volume fractions greater than 25% and less than 25%. The regressions to each part of the data set are as follows:
The work of Brie et al. (1995) extended the work of Castagna to include predicting gas saturation from a Vp/Vs vs. compressional traveltime tc crossplot. In this crossplot, Brie proposes one non-linear trend for water-wet sands and another non-linear trend for shales.
Vs .842V p 1.099 , clay > 25%
(4)
Vs .754V p 0.657 , clay < 25%
(5)
Figure 3 compares the Castagna mudrock, Brie water-wet sand, and the two Han regressions. It can be seen that the Castagna mudrock line essentially lies on top of the Han regression for clay >25%.
Fig. 3 Vp versus Vs crossplot for four regression models. Figure 4 compares the Castagna mudrock, Brie water-wet sand, and the two Han regressions, but now in the context of a Brie Vp/Vs vs. compressional traveltime tc crossplot.
Fig. 2 Vp/Vs versus tc crossplot for shear interpretation (Brie et al., 1995). Rewriting the Castagna mudrock equation gives the following:
Vs .8621V p 1.1724
(2)
Brie does not provide any equation for his water-wet sand 2
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
Fig. 5 Vp versus Vs crossplot for some organic rich shales. Fig. 4 Vp/Vs versus tc crossplot for four regression models.
Figure 6 presents the organic shale data in the context of a Brie Vp/Vs vs. compressional traveltime tc crossplot with the Castagna mudrock and Brie water-wet sand curves. In Figure 6, the majority of the data lies below the Brie waterwet sand data. Some of the data such as some of the Bazhenov, Barnett, and Monterey samples have a Vp/Vs ratio less than the 1.58 assumed for gas bearing sandstones by Brie. However, the work of Murphy et al. (1991) suggests a Vp/Vs ratio for dry or gas bearing sandstones of 1.5. Note in Figure 6 there are two samples with a Vp/Vs ratio greater than 1.8. Pickett (1963) suggests a Vp/Vs ratio for limestone of 1.9 and a Vp/Vs ratio for dolomite of 1.8.
Comparison of Figures 3 and 4 reveals the latter provides better discrimination between sands and shales, particularly in slow formations. In Brie crossplot space, the Castagna mudrock curve lies above the Brie water-wet curve until tc slows to less than about 70 microseconds/ft, which is about the lower limit of the actual Castagna mudrock data in Figure 1. There are other models available in the literature, but in general, they all lie between the Castagna mudrock and Han clay less than 25% curves of Figure 4. In general, it can be expected for any reservoir that mudrock and sand water-wet trends can be derived from the well log data in vertical wells. However, in unconventional formations, the trends can be expected to be more difficult to characterize due to the possible presence of carbonates and kerogen. Murphy et al. (2015) have compiled a collection of core data from organic shales in an attempt to characterize geomechanical properties of anisotropic shales, all with varying degrees of kerogen volume. The data of Bakken, Bazhenov, Niobrara, Monterey, Northsea, and Lockatong are from Vernik and Liu (1997). The data of Barnett, Eagle Ford, and Haynesville are from shale gas reservoirs from Sone (2012). Figure 5 presents this data in a Vp vs. Vs crossplot along with the Castagna mudrock and Brie waterwet sand trends. It can be seen in general the data lies off the Castagna and Brie trends.
Fig. 6 Vp/Vs versus tc crossplot for some organic rich shales. Figure 7 shows how the kerogen volume contributes to the departure of the data below the Brie water-wet trend. The samples with small amounts of kerogen clearly lie closer to 3
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
the Castagna and Brie trends than those with moderate to large amounts of kerogen.
Fig. 8 XRD data from the Haynesville and Bossier shale. Note that the two Bossier wells have more clay.
The depositional and diagenetic history created a complex pore geometry that complicates the reservoir characterization. Matrix microporosity resides in primary intercrystalline pores, as well as in organic matter. Where clastic dilution has occurred, intergranular porosity co-exists within the primary intercrystalline matrix porosity. In zones containing abundant, thermally mature organic matter, the transformation of kerogen to dry gas has created abundant secondary porosity within the organic matter. In Figure 9, Quirein et al. (2010), show a typical interpretation for a gas bearing Haynesville well. Track 1 compares the log interpretation and core gas filled porosity. Track 2 compares the core porosity (blue dots) with the log interpretation total and effective porosity (black and green curves respectively). The core porosity lies between the log interpretation total and effective porosity. Tracks 3 and 4 compare the core and log grain density and bulk density, respectively. Track 5 presents the mineralogy and volumetrics. There is a lot of clay, and the carbonate percentage increases as the clay and quartz percentages decrease as the well descends. Track 6 presents the fluid analysis, and as seen, the kerogen and gas volumes correlate and decrease as the well descends.
Fig. 7 Vp/Vs versus tc crossplot for some organic rich shales with the z-axis representing the volume of kerogen. LOG EXAMPLE 1: HAYNESVILLE SHALE The Haynesville shale is a highly productive gas shale of Late Jurassic age that was deposited in quiet water, within a restricted intrashelf (shallow ocean) basin. This intrashelf basin existed across east Texas and northwestern Louisiana. It was restricted from oceanic circulation by shallow-water carbonates to the north-northwest, west, and southwest by deltaic deposits to the northeast, and by remnant basement blocks to the southeast (Pope et al. 2009). The reservoir facies is a black, organic-rich mudstone, which consists of clay-sized particles, with minor amounts of very fine silt and carbonate mud. The clay content is generally less than 40 to 50%, and can be significantly lower in the calcite-rich areas. It is rich in organic matter; the total organic content ranges from 2% to more than 5% in more anoxic portions of the basin (Spain et al. 2010). The shale is thermally mature and highly overpressured.
Fig. 9 Interpretation for a Haynesville gas well (Quirein et al., 2010). Figures 10 a, b present the Brie crossplot for the data and interpretation displayed in Figure 9. The red curve is the Castagna mudrock trend and the blue curve is the Brie water wet sand trend. Similarly to the core data of Figure 6, most of the data in Figure 10b lies below the Brie water-wet curve with the departure increasing with gamma ray as shown in the z-axis. In the Haynesville, the gamma ray is 4
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
known to increase with either the volume of clay or kerogen. Here the gamma ray decreases as the well descends, and the Vp/Vs ratio approaches values greater than 1.75 in the lower interval of the well.
ORGANIC POROSITY ANALYSIS
FROM
SHALE
CORE
Fig. 10a Vp/Vs versus tc crossplot for a Haynesville gas bearing organic rich shales with the z-axis representing the Gamma Ray.
Results from the shale core analysis protocol (Galford et al. 2013), can be used to obtain an empirical value for the porosity associated with kerogen, k. Figure 11 shows how one can use the technique to obtain k for the Haynesville Shale by making a crossplot of gas-filled porosity from the shale core analyses vs. present-day kerogen volume. The data in Figure 11 are measurements from Haynesville core samples taken from wells in Northeast Texas and Northern Louisiana. The technique relies on the relationship for the total volume of hydrocarbon (the volume of hydrocarbon in porous minerals plus the volume of hydrocarbon in porous kerogen). In source rock reservoirs, the total volume of hydrocarbon approaches a minimum value when the volume of hydrocarbon in porous minerals approaches zero. Thus, a trend near the lower-right edge of the data cluster represents the minimum hydrocarbon volume. The red line in Figure 11 represents the minimum hydrocarbon volume for the Haynesville Shale, and the porosity associated with kerogen k is related to the slope so that Slope (9) . 1 k
Fig. 10b Vp/Vs versus tc crossplot for a Haynesville gas bearing organic rich shales with the z-axis representing volume kerogen plus gas.
Fig. 11 GRI gas-filled porosity vs. present-day kerogen volume crossplots can be used to determine the porosity associated with kerogen. The core data is from several Haynesville Shale wells.
Figure 10b displays the same data as Figure 10a, but now the z-axis corresponds to the volume kerogen plus gas. Gas has a Vp/Vs ratio in the range of 1.5 to 1.6, and kerogen has a Vp/Vs ratio in the range 1.7 to 1.8. In the plot it is difficult to separate the effects of kerogen from those of gas, but in general, both cause the data to lie below the Brie water-wet trend.
Figure 11 illustrates the strong correlation between the volumes of kerogen and gas filled porosity. This strong correlation is used to define the Vp/Vs ratio and tc an organic matter indicator similar to that proposed by Williams et al. (1984), for oil or gas. In this case, the organic matter indicator includes any of the volumes of gas, oil, and kerogen. Figure 10a illustrates how for a given tc the increasing distance delta of the Vp/Vs ratio from the Brie 5
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
kaolinite. In general, TOC ranges from 2.1 to 5.2 wt. %, vitrinite reflectance (Ro) ranges from an immature 0.68 to 1.5, core porosities vary from 1.5% to 9%, and log values average between 3% and 15%.
water-wet trend corresponds to increasing volume of kerogen plus gas. From equation 3, this distance can be computed as follows:
Vp /Vs
Vp
Vp
.014V .883V p .9886 Vs 2 p
(6)
Fig. 12 Haynesville well volumes kerogen + gas versus delta Vp/Vs below the Brie water-wet trend. Figure 12 displays on the Y-axis, the volumes kerogen plus gas versus the X-axis, Vp /Vs . Due to the varying
Fig. 13 Eagle Ford well interpretation.
mineralogy, gas, and kerogen volumes, the R-square is only a modest 0.5. However, the validity of the indicator is apparent and is obtained using only the sonic Vp/Vs ratio and the travel time tc.
In Figure 13, Quirein et al. (2012) present the results of an interpretation for an Eagle Ford well. The mineralogy and fluid analysis is presented in Track 1 with kerogen shaded in magenta and oil shaded in green. Track 2 compares the NMR porosity with core porosity. Track 3 shows how well the difference, the crossplot minus NMR porosity (blue curve), agrees with kerogen from the core analysis.
LOG EXAMPLE2: EAGLE FORD SHALE The Eagle Ford (known as the Boquillas formation in outcrop) is an Upper Cretaceous (Cenomanian to Turonian) formation comprised of limestone, marls, and claystones with some clays originating from volcanic ash (tuffs). It is a self-contained petroleum system, consisting of interstratified source, seal, and potential reservoir. The productive Eagle Ford extends from the Mexican border (Webb to Maverick Counties) striking northeast to Brazos County, covering approximately 19,500 mi2. The Eagle Ford varies from approximately 14,000 ft deep and 300 ft thick in the southeast to approximately 4,000 ft deep and 50 ft thick moving up the dip to the northwest. Bottomhole temperatures range from 340ºF to 240ºF; hence, production is mainly gas in the deeper Eagle Ford, gas condensate, condensate, and oil moving up the dip. Mineralogy consists of mainly calcite, quartz, and illite, with minor amounts of pyrite, dolomite, plagioclase, mixed-layer clay, chlorite, and
Figure 14 presents a Brie Vp/Vs vs. compressional traveltime tc crossplot for the data interval displayed in Figure 13. The
black curve is the Brie water-wet sand trend.
6
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
Fig. 14 Vp/Vs versus tc crossplot for an Eagle Ford oil bearing organic rich shale with the z-axis representing the volume kerogen + oil.
Fig. 15 Vp/Vs versus tc crossplot for a lower Barnett shale gas bearing organic rich mudstone with the z-axis representing the volume kerogen.
The samples within the blue ellipse are from the lowest interval (blue rectangle) of Figure 13, the Buda Limestone, with an average Vp/Vs of 1.9. The samples in the red ellipse are from the zone richest in kerogen and oil, within the red rectangle of Figure 13. It can be seen in Figure 14 that the volume of kerogen plus oil increases (z-axis) as DTC increases, and there is minimal reduction or change in the Vp/Vs ratio.
CONCLUSIONS The Castagna mudrock and Brie water-wet sand trends do not appear to apply in general to the Eagle Ford, Haynesville, and Barnett mudstone formations. This is attributed to the presence of carbonates (mainly limestone) and organic matter. Nevertheless, the gas bearing Haynesville and Barnett formations exhibit an expected Vp/Vs ratio of about 1.5 to 1.65.
LOG EXAMPLE3: BARNETT SHALE Laboratory data from organic shales from the Bakken, Bazhenov, Niobrara, Monterey, Northsea, Lockatong, Barnett, Eagle Ford, and Haynesville was observed to generally fall below the Brie water-wet trend in a Brie Vp/Vs vs. compressional traveltime tc crossplot. The distance beneath the trend curve was weakly correlated with the volume of kerogen. The same was observed for actual logging data from gas bearing zones of the Hainesville (carbonate rich) and Barnett (quartz rich) shales with the Vp/Vs ratio in the gas zones ranging from 1.5 to 1.65.
The Barnett Shale is composed of a combination of marine clays, primarily illite and chlorite, detrital quartz silt, silicified and carbonate bioclasts and fossils, interstitial organic carbon, and phosphate. Most Barnett Shale samples are siliceous mudstones. Many are so rich in quartz that they are really argillaceous siltstones. Figure 15 below presents a Brie Vp/Vs vs. compressional traveltime tc crossplot for the siliceous lower Barnett shale and underlying Viola limestone. The z-axis represents the interpreted volume of kerogen. The samples within the blue ellipse are from the Viola limestone, with some variation due to the presence of clay. The samples within the red ellipse are from the Lower Barnett formation, have between 40 and 60% clay, and are kerogen (and gas) rich. The gas and kerogen dominates the response, and it can be seen most of the samples lie below the Brie water-wet sand trend and the gas response of a Vp/Vs ratio of about 1.58 to 1.65.
Data from the Eagle Ford was from an oil-bearing well. It was observed that the volume of kerogen plus oil increases (z-axis) as DTC increases, and there is minimal reduction or change in the Vp/Vs ratio with a lower bound of about 1.8.
7
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
and Odumosu, T., 2010, Integrating core data and wireline geochemical data for formation evaluation and characterization of shale gas reservoirs: SPE 134559 presented at the Annual Technical Conference and Exhibition, SPE, Florence Italy, 19-22 September.
Work needs to be done to construct reference curves for the organic shales similar to the Brie water-sand of the Castagna mudrock trends. ACKNOWLEDGEMENTS
Sone, H., 2012, Mechanical properties of shale gas reservoir rocks and its relation to the in-situ stress variation observed in shale gas reservoirs: Ph.D. Thesis, Stanford University.
REFERENCES Brie, A., Pampuri, F., Marsala, A.F., Meazza, O., 1995, Shear sonic interpretation in gas-bearing sands: presented at the SPE Annual Conference and Exhibition, Dallas, Tx, 2225 October.
Spain, D.R., Dix, M., Sano, J., Ratcliffe, K., Hughes, S., and Buller, D., 2010, Exploiting whole-rock elemental data in shale-gas development programs with an example from the Jurassic Haynesville Formation of North America: European Association of Geoscientists and Engineer Shale Workshop, Nice, France, 26-28 April.
Castagna, J.P., Batzel, M.L., and Eastwood, R.L., 1985, Relationships between compressional wave and shear wave velocities in clastic silicate rocks: Geophysics Vol. 50, No. 4, 571.
Vernik, L., and Liu, X., 1997, Velocity anisotropy in shales: a petrophysical study: Geophysics, vol. 62, 521-532.
Galford, J., Quirein, J., Westacott, D., and Witkowsky, J., 2013, Quantifying organic porosity from logs: SPWLA 54th Annual Logging Symposium, June 22-26.
Williams, D.M., Zemenek, J., Angona, F.A., Dennis, C.L., and Caldwell, R.L., 1984, The long spaced acoustic logging tool: 25th SPWLA Annual Logging Symposium, Midland, TX, June.
Han, D.H., 1986, Effects of porosity and clay content on acoustic properties of sandstones and unconsolidated sediments: Ph.D. dissertation, Stanford University.
ABOUT THE AUTHORS Murphy, E., Barraza, S.R., Gu, M., Gokaraju, D., Far, M.E., and Quirein, J., 2015, New models for acoustic anisotropic interpretation in shale: SPWLA 56th Annual Logging Symposium, July 18-22.
John Quirein is the petrophysics team leader in the Halliburton Formation Evaluation Technology group, focusing in interpretation and software development with a recent emphasis on gas shale petrophysics, geochemical log interpretation, and multi-mineral solvers. He received a PhD degree from the University of Houston, and has worked for 10 years at Schlumberger, 12 years at Mobil, and the last 12 years at Halliburton. Quirein is a past SPWLA president and is currently a member of the SPWLA Foundation.
Murphy,W.F., Schwartz, L.M., Hornby, B, 1991, Interpretation physics of Vp and Vs in sedimentary rocks: 32nd SPWLA Annual Logging Symposium, Midland, TX, June. Passey, Q.R., Bohacs, K.M., Esch, W.L., Klimentidis, R., and Sinha, S., 2010, From oil-prone source rock to gasproducing shale reservoir – geologic and petrophysical characterization of unconventional shale-gas reservoirs: SPE131350.
Mehdi E. Far is a Principal Scientist in the Integrated Interpretation Group, Halliburton Technology Center in Houston, since 2014. He received a Ph.D. degree in geophysics from University of Houston. His research focuses on rock physics, petrophysics, advanced interpretation methods for sonic logs, and geomechanics. Mehdi E. Far is a member of SEG, SPWLA, SPE, and EAGE and a member of SEG research committee.
Pickett. G R., 1963, Acoustic character logs and their applications in formation evaluation: J Petr. Tech., Vol. 15, 650-667. Pope, C., Peters, B., Belton, T., and Palish, T., 2009, Haynesville Shale – one operator’s approach to well completions in this evolving play: SPE 125079 presented at the Annual Technical Conference and Exhibition, SPE, New Orleans, Louisiana, USA, 4-7 October. Quirein, J., Witkowsky, J., Truax, J., Galford, J., Spain, D., 8
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
research interests are geomechanics, rock physics, and advanced well logging interpretation techniques. He is currently a member of SPE and SPWLA.
Ming Gu is a Senior Scientist in the Integrated Interpretation Group, Halliburton Technology Center in Houston, since Feb. 2014. He graduated from the University of Texas at Austin with a Ph.D degree in petroleum engineering in 2013. His current research focuses on petrophysics, geomechanics, rock physics, formation testing and reservoir fluid analysis. He is a member of SPWLA.
Jim Witkowsky is the technical advisor for the Halliburton Petrophysics Technology Development group working with GEM elementals and mineral analysis in shale plays. He graduated from the University of Pittsburgh with a degree in chemical/petroleum engineering and began his career in 1986 as a general field engineer for Schlumberger, working primarily in the Gulf of Mexico. Witkowsky worked at the NUMAR corporation from 1992 to 1997, where he held various positions, including NMR field engineer for USA, Indonesia, and Australia, manager of Australia field operations, and petrophysicist. After the merger, he was assigned to the Halliburton Formation and Reservoir Solutions Center as a NMR petrophysicist, responsible for client products, interpretations, and NMR training. He is a member of SPE and SPWLA.
Deepak Gokaraju is a Senior Scientist in the Integrated Interpretation team of Halliburton Formation and Reservoir Solutions in Houston with a focus on providing integrated solutions using advanced interpretation methods. He graduated with a Master’s degree in Petroleum Engineering from Missouri University of Science and Technology. His
9
Relationships between Sonic Compressional and Shear logs in Unconventional Formations John Quirein*, Mehdi Far, Ming Gu, Deepak Gokaraju, and Jim Witkowsky; Halliburton *Corresponding Author Society of Petrophysicists and Well Log Analysts Copyright 2015, held jointly by the Society of Petrophysicists and Well Log Analysts (SPWLA) and the submitting authors. This paper was prepared for presentation at the SPWLA 56th Annual Logging Symposium held in Long Beach, CA, USA, July 18-22, 2015.
ABSTRACT Relationships between compressional and shear wave velocities were established in the classic paper by Castagna et al. (1985). In particular, linear trends for correlating the compressional and shear wave velocities are presented, including the Castagna mudrock line derived from in-situ sonic and seismic measurements. The work of Brie et al. (1995) extended the work of Castagna et al. (1985) to include predicting gas saturation from a Vp/Vs vs. compressional traveltime crossplot. In this crossplot, Brie proposes one non-linear trend for water-wet sands and another non-linear trend for shales. In addition, the Brie Vp/Vs vs. compressional traveltime crossplot has a greater sensitivity to low velocities or large compressional traveltimes. This paper compares some of the more famous rock physic trends, including the Castagna mudrock and Brie water-wet sands and shales, in both the Vp vs. Vs crossplot and the Vp/Vs vs. compressional traveltime crossplot. It also presents plots of laboratory data from the Bakken, Bazhenov, Barnett, Eagle Ford, Haynesville, Monterey, and Niobrara organic shales and descriptions of departures from the expected trends to the presence of kerogen and anisotropy.
are not commonly found in the literature. Passey (2010) points out that shales are more properly referred to as mudstones and exhibit a wide range in composition including clay, quartz, feldspar, heavy minerals, and carbonate. Moreover, the composition of a typical mudstone will vary much more than typical sandstones, even though, to the naked eye, many mudstones look similar. Passey (2010) presents an example of what he calls a “clay-rich gas-bearing mudrock” with the carbonate fraction varying from 0 to greater than 75%. In this paper, “Castagna mudrock” refers to clastic silicate rock composed primarily of clay or silt-sized particles, whereas “mudstone” refers to the more general case, which includes carbonate-rich shales. The Castagna mudrock equation suggests a linear relationship between the compressional velocity, Vp, and the shear velocity, Vs:
V p 1.16Vs 1.36
(1)
where the velocities are expressed in km/s. The mudrock equation provides a good fit for sonic and seismic data from the Devonian, Pierre, Grayson, and Japanese shales and an oil shale.
The paper also includes an investigation of in-situ well log data from the Eagle Ford, Haynesville, and Barnett formations in the context of the Castagna, Brie, and additional crossplots. The Eagle Ford data satisfies the carbonate Vp vs. Vs trend predicted by Castagna, and the Haynesville data clearly satisfies the hydrocarbon/kerogen effect predicted by the Brie Vp/Vs vs. compressional traveltime crossplot. INTRODUCTION Castagna et al. (1985) defined mudrocks as clastic silicate rock composed primarily of clay or silt-sized particles. Lithified muds are composed primarily of quartz and clay minerals. Owing to the difficulty associated with handling of most mudrocks, laboratory measurements on these rocks
Fig. 1 Vp and Vs for mudrocks (Castagna et al. 1985). In Fig. 1 the clay point is empirical and represents a dry clay mineral with no porosity. As porosity is added, the dry clay point moves towards the 100% water point at zero shear 1
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
velocity. The net result is that many quartz-clay-water ternary mixtures are spread along an elongate triangular region loosely defined by clay-water, quartz-clay, and quartz-water lines. Castagna concludes, to first order that shear velocity is linearly related to compressional wave velocity for water saturated rocks. For a given Vp, mudrocks tend toward slightly higher Vp/Vs than do clean porous sandstones.
trend in Figure 2, but the following empirically derived equation fits the trend:
Vs .0147V p2 .883V p .9886
(3)
A literature search reveals a large assortment of regressions relating the compressional and shear velocities. This study includes two from Han (1986). Han used ultrasonic data for 70 water-saturated shaley sand samples. The data are separated by clay volume fractions greater than 25% and less than 25%. The regressions to each part of the data set are as follows:
The work of Brie et al. (1995) extended the work of Castagna to include predicting gas saturation from a Vp/Vs vs. compressional traveltime tc crossplot. In this crossplot, Brie proposes one non-linear trend for water-wet sands and another non-linear trend for shales.
Vs .842V p 1.099 , clay > 25%
(4)
Vs .754V p 0.657 , clay < 25%
(5)
Figure 3 compares the Castagna mudrock, Brie water-wet sand, and the two Han regressions. It can be seen that the Castagna mudrock line essentially lies on top of the Han regression for clay >25%.
Fig. 3 Vp versus Vs crossplot for four regression models. Figure 4 compares the Castagna mudrock, Brie water-wet sand, and the two Han regressions, but now in the context of a Brie Vp/Vs vs. compressional traveltime tc crossplot.
Fig. 2 Vp/Vs versus tc crossplot for shear interpretation (Brie et al., 1995). Rewriting the Castagna mudrock equation gives the following:
Vs .8621V p 1.1724
(2)
Brie does not provide any equation for his water-wet sand 2
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
Fig. 5 Vp versus Vs crossplot for some organic rich shales. Fig. 4 Vp/Vs versus tc crossplot for four regression models.
Figure 6 presents the organic shale data in the context of a Brie Vp/Vs vs. compressional traveltime tc crossplot with the Castagna mudrock and Brie water-wet sand curves. In Figure 6, the majority of the data lies below the Brie waterwet sand data. Some of the data such as some of the Bazhenov, Barnett, and Monterey samples have a Vp/Vs ratio less than the 1.58 assumed for gas bearing sandstones by Brie. However, the work of Murphy et al. (1991) suggests a Vp/Vs ratio for dry or gas bearing sandstones of 1.5. Note in Figure 6 there are two samples with a Vp/Vs ratio greater than 1.8. Pickett (1963) suggests a Vp/Vs ratio for limestone of 1.9 and a Vp/Vs ratio for dolomite of 1.8.
Comparison of Figures 3 and 4 reveals the latter provides better discrimination between sands and shales, particularly in slow formations. In Brie crossplot space, the Castagna mudrock curve lies above the Brie water-wet curve until tc slows to less than about 70 microseconds/ft, which is about the lower limit of the actual Castagna mudrock data in Figure 1. There are other models available in the literature, but in general, they all lie between the Castagna mudrock and Han clay less than 25% curves of Figure 4. In general, it can be expected for any reservoir that mudrock and sand water-wet trends can be derived from the well log data in vertical wells. However, in unconventional formations, the trends can be expected to be more difficult to characterize due to the possible presence of carbonates and kerogen. Murphy et al. (2015) have compiled a collection of core data from organic shales in an attempt to characterize geomechanical properties of anisotropic shales, all with varying degrees of kerogen volume. The data of Bakken, Bazhenov, Niobrara, Monterey, Northsea, and Lockatong are from Vernik and Liu (1997). The data of Barnett, Eagle Ford, and Haynesville are from shale gas reservoirs from Sone (2012). Figure 5 presents this data in a Vp vs. Vs crossplot along with the Castagna mudrock and Brie waterwet sand trends. It can be seen in general the data lies off the Castagna and Brie trends.
Fig. 6 Vp/Vs versus tc crossplot for some organic rich shales. Figure 7 shows how the kerogen volume contributes to the departure of the data below the Brie water-wet trend. The samples with small amounts of kerogen clearly lie closer to 3
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
the Castagna and Brie trends than those with moderate to large amounts of kerogen.
Fig. 8 XRD data from the Haynesville and Bossier shale. Note that the two Bossier wells have more clay.
The depositional and diagenetic history created a complex pore geometry that complicates the reservoir characterization. Matrix microporosity resides in primary intercrystalline pores, as well as in organic matter. Where clastic dilution has occurred, intergranular porosity co-exists within the primary intercrystalline matrix porosity. In zones containing abundant, thermally mature organic matter, the transformation of kerogen to dry gas has created abundant secondary porosity within the organic matter. In Figure 9, Quirein et al. (2010), show a typical interpretation for a gas bearing Haynesville well. Track 1 compares the log interpretation and core gas filled porosity. Track 2 compares the core porosity (blue dots) with the log interpretation total and effective porosity (black and green curves respectively). The core porosity lies between the log interpretation total and effective porosity. Tracks 3 and 4 compare the core and log grain density and bulk density, respectively. Track 5 presents the mineralogy and volumetrics. There is a lot of clay, and the carbonate percentage increases as the clay and quartz percentages decrease as the well descends. Track 6 presents the fluid analysis, and as seen, the kerogen and gas volumes correlate and decrease as the well descends.
Fig. 7 Vp/Vs versus tc crossplot for some organic rich shales with the z-axis representing the volume of kerogen. LOG EXAMPLE 1: HAYNESVILLE SHALE The Haynesville shale is a highly productive gas shale of Late Jurassic age that was deposited in quiet water, within a restricted intrashelf (shallow ocean) basin. This intrashelf basin existed across east Texas and northwestern Louisiana. It was restricted from oceanic circulation by shallow-water carbonates to the north-northwest, west, and southwest by deltaic deposits to the northeast, and by remnant basement blocks to the southeast (Pope et al. 2009). The reservoir facies is a black, organic-rich mudstone, which consists of clay-sized particles, with minor amounts of very fine silt and carbonate mud. The clay content is generally less than 40 to 50%, and can be significantly lower in the calcite-rich areas. It is rich in organic matter; the total organic content ranges from 2% to more than 5% in more anoxic portions of the basin (Spain et al. 2010). The shale is thermally mature and highly overpressured.
Fig. 9 Interpretation for a Haynesville gas well (Quirein et al., 2010). Figures 10 a, b present the Brie crossplot for the data and interpretation displayed in Figure 9. The red curve is the Castagna mudrock trend and the blue curve is the Brie water wet sand trend. Similarly to the core data of Figure 6, most of the data in Figure 10b lies below the Brie water-wet curve with the departure increasing with gamma ray as shown in the z-axis. In the Haynesville, the gamma ray is 4
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
known to increase with either the volume of clay or kerogen. Here the gamma ray decreases as the well descends, and the Vp/Vs ratio approaches values greater than 1.75 in the lower interval of the well.
ORGANIC POROSITY ANALYSIS
FROM
SHALE
CORE
Fig. 10a Vp/Vs versus tc crossplot for a Haynesville gas bearing organic rich shales with the z-axis representing the Gamma Ray.
Results from the shale core analysis protocol (Galford et al. 2013), can be used to obtain an empirical value for the porosity associated with kerogen, k. Figure 11 shows how one can use the technique to obtain k for the Haynesville Shale by making a crossplot of gas-filled porosity from the shale core analyses vs. present-day kerogen volume. The data in Figure 11 are measurements from Haynesville core samples taken from wells in Northeast Texas and Northern Louisiana. The technique relies on the relationship for the total volume of hydrocarbon (the volume of hydrocarbon in porous minerals plus the volume of hydrocarbon in porous kerogen). In source rock reservoirs, the total volume of hydrocarbon approaches a minimum value when the volume of hydrocarbon in porous minerals approaches zero. Thus, a trend near the lower-right edge of the data cluster represents the minimum hydrocarbon volume. The red line in Figure 11 represents the minimum hydrocarbon volume for the Haynesville Shale, and the porosity associated with kerogen k is related to the slope so that Slope (9) . 1 k
Fig. 10b Vp/Vs versus tc crossplot for a Haynesville gas bearing organic rich shales with the z-axis representing volume kerogen plus gas.
Fig. 11 GRI gas-filled porosity vs. present-day kerogen volume crossplots can be used to determine the porosity associated with kerogen. The core data is from several Haynesville Shale wells.
Figure 10b displays the same data as Figure 10a, but now the z-axis corresponds to the volume kerogen plus gas. Gas has a Vp/Vs ratio in the range of 1.5 to 1.6, and kerogen has a Vp/Vs ratio in the range 1.7 to 1.8. In the plot it is difficult to separate the effects of kerogen from those of gas, but in general, both cause the data to lie below the Brie water-wet trend.
Figure 11 illustrates the strong correlation between the volumes of kerogen and gas filled porosity. This strong correlation is used to define the Vp/Vs ratio and tc an organic matter indicator similar to that proposed by Williams et al. (1984), for oil or gas. In this case, the organic matter indicator includes any of the volumes of gas, oil, and kerogen. Figure 10a illustrates how for a given tc the increasing distance delta of the Vp/Vs ratio from the Brie 5
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
kaolinite. In general, TOC ranges from 2.1 to 5.2 wt. %, vitrinite reflectance (Ro) ranges from an immature 0.68 to 1.5, core porosities vary from 1.5% to 9%, and log values average between 3% and 15%.
water-wet trend corresponds to increasing volume of kerogen plus gas. From equation 3, this distance can be computed as follows:
Vp /Vs
Vp
Vp
.014V .883V p .9886 Vs 2 p
(6)
Fig. 12 Haynesville well volumes kerogen + gas versus delta Vp/Vs below the Brie water-wet trend. Figure 12 displays on the Y-axis, the volumes kerogen plus gas versus the X-axis, Vp /Vs . Due to the varying
Fig. 13 Eagle Ford well interpretation.
mineralogy, gas, and kerogen volumes, the R-square is only a modest 0.5. However, the validity of the indicator is apparent and is obtained using only the sonic Vp/Vs ratio and the travel time tc.
In Figure 13, Quirein et al. (2012) present the results of an interpretation for an Eagle Ford well. The mineralogy and fluid analysis is presented in Track 1 with kerogen shaded in magenta and oil shaded in green. Track 2 compares the NMR porosity with core porosity. Track 3 shows how well the difference, the crossplot minus NMR porosity (blue curve), agrees with kerogen from the core analysis.
LOG EXAMPLE2: EAGLE FORD SHALE The Eagle Ford (known as the Boquillas formation in outcrop) is an Upper Cretaceous (Cenomanian to Turonian) formation comprised of limestone, marls, and claystones with some clays originating from volcanic ash (tuffs). It is a self-contained petroleum system, consisting of interstratified source, seal, and potential reservoir. The productive Eagle Ford extends from the Mexican border (Webb to Maverick Counties) striking northeast to Brazos County, covering approximately 19,500 mi2. The Eagle Ford varies from approximately 14,000 ft deep and 300 ft thick in the southeast to approximately 4,000 ft deep and 50 ft thick moving up the dip to the northwest. Bottomhole temperatures range from 340ºF to 240ºF; hence, production is mainly gas in the deeper Eagle Ford, gas condensate, condensate, and oil moving up the dip. Mineralogy consists of mainly calcite, quartz, and illite, with minor amounts of pyrite, dolomite, plagioclase, mixed-layer clay, chlorite, and
Figure 14 presents a Brie Vp/Vs vs. compressional traveltime tc crossplot for the data interval displayed in Figure 13. The
black curve is the Brie water-wet sand trend.
6
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
Fig. 14 Vp/Vs versus tc crossplot for an Eagle Ford oil bearing organic rich shale with the z-axis representing the volume kerogen + oil.
Fig. 15 Vp/Vs versus tc crossplot for a lower Barnett shale gas bearing organic rich mudstone with the z-axis representing the volume kerogen.
The samples within the blue ellipse are from the lowest interval (blue rectangle) of Figure 13, the Buda Limestone, with an average Vp/Vs of 1.9. The samples in the red ellipse are from the zone richest in kerogen and oil, within the red rectangle of Figure 13. It can be seen in Figure 14 that the volume of kerogen plus oil increases (z-axis) as DTC increases, and there is minimal reduction or change in the Vp/Vs ratio.
CONCLUSIONS The Castagna mudrock and Brie water-wet sand trends do not appear to apply in general to the Eagle Ford, Haynesville, and Barnett mudstone formations. This is attributed to the presence of carbonates (mainly limestone) and organic matter. Nevertheless, the gas bearing Haynesville and Barnett formations exhibit an expected Vp/Vs ratio of about 1.5 to 1.65.
LOG EXAMPLE3: BARNETT SHALE Laboratory data from organic shales from the Bakken, Bazhenov, Niobrara, Monterey, Northsea, Lockatong, Barnett, Eagle Ford, and Haynesville was observed to generally fall below the Brie water-wet trend in a Brie Vp/Vs vs. compressional traveltime tc crossplot. The distance beneath the trend curve was weakly correlated with the volume of kerogen. The same was observed for actual logging data from gas bearing zones of the Hainesville (carbonate rich) and Barnett (quartz rich) shales with the Vp/Vs ratio in the gas zones ranging from 1.5 to 1.65.
The Barnett Shale is composed of a combination of marine clays, primarily illite and chlorite, detrital quartz silt, silicified and carbonate bioclasts and fossils, interstitial organic carbon, and phosphate. Most Barnett Shale samples are siliceous mudstones. Many are so rich in quartz that they are really argillaceous siltstones. Figure 15 below presents a Brie Vp/Vs vs. compressional traveltime tc crossplot for the siliceous lower Barnett shale and underlying Viola limestone. The z-axis represents the interpreted volume of kerogen. The samples within the blue ellipse are from the Viola limestone, with some variation due to the presence of clay. The samples within the red ellipse are from the Lower Barnett formation, have between 40 and 60% clay, and are kerogen (and gas) rich. The gas and kerogen dominates the response, and it can be seen most of the samples lie below the Brie water-wet sand trend and the gas response of a Vp/Vs ratio of about 1.58 to 1.65.
Data from the Eagle Ford was from an oil-bearing well. It was observed that the volume of kerogen plus oil increases (z-axis) as DTC increases, and there is minimal reduction or change in the Vp/Vs ratio with a lower bound of about 1.8.
7
SPWLA 55th Annual Logging Symposium, May 18-22, 2014
and Odumosu, T., 2010, Integrating core data and wireline geochemical data for formation evaluation and characterization of shale gas reservoirs: SPE 134559 presented at the Annual Technical Conference and Exhibition, SPE, Florence Italy, 19-22 September.
Work needs to be done to construct reference curves for the organic shales similar to the Brie water-sand of the Castagna mudrock trends. ACKNOWLEDGEMENTS
Sone, H., 2012, Mechanical properties of shale gas reservoir rocks and its relation to the in-situ stress variation observed in shale gas reservoirs: Ph.D. Thesis, Stanford University.
REFERENCES Brie, A., Pampuri, F., Marsala, A.F., Meazza, O., 1995, Shear sonic interpretation in gas-bearing sands: presented at the SPE Annual Conference and Exhibition, Dallas, Tx, 2225 October.
Spain, D.R., Dix, M., Sano, J., Ratcliffe, K., Hughes, S., and Buller, D., 2010, Exploiting whole-rock elemental data in shale-gas development programs with an example from the Jurassic Haynesville Formation of North America: European Association of Geoscientists and Engineer Shale Workshop, Nice, France, 26-28 April.
Castagna, J.P., Batzel, M.L., and Eastwood, R.L., 1985, Relationships between compressional wave and shear wave velocities in clastic silicate rocks: Geophysics Vol. 50, No. 4, 571.
Vernik, L., and Liu, X., 1997, Velocity anisotropy in shales: a petrophysical study: Geophysics, vol. 62, 521-532.
Galford, J., Quirein, J., Westacott, D., and Witkowsky, J., 2013, Quantifying organic porosity from logs: SPWLA 54th Annual Logging Symposium, June 22-26.
Williams, D.M., Zemenek, J., Angona, F.A., Dennis, C.L., and Caldwell, R.L., 1984, The long spaced acoustic logging tool: 25th SPWLA Annual Logging Symposium, Midland, TX, June.
Han, D.H., 1986, Effects of porosity and clay content on acoustic properties of sandstones and unconsolidated sediments: Ph.D. dissertation, Stanford University.
ABOUT THE AUTHORS Murphy, E., Barraza, S.R., Gu, M., Gokaraju, D., Far, M.E., and Quirein, J., 2015, New models for acoustic anisotropic interpretation in shale: SPWLA 56th Annual Logging Symposium, July 18-22.
John Quirein is the petrophysics team leader in the Halliburton Formation Evaluation Technology group, focusing in interpretation and software development with a recent emphasis on gas shale petrophysics, geochemical log interpretation, and multi-mineral solvers. He received a PhD degree from the University of Houston, and has worked for 10 years at Schlumberger, 12 years at Mobil, and the last 12 years at Halliburton. Quirein is a past SPWLA president and is currently a member of the SPWLA Foundation.
Murphy,W.F., Schwartz, L.M., Hornby, B, 1991, Interpretation physics of Vp and Vs in sedimentary rocks: 32nd SPWLA Annual Logging Symposium, Midland, TX, June. Passey, Q.R., Bohacs, K.M., Esch, W.L., Klimentidis, R., and Sinha, S., 2010, From oil-prone source rock to gasproducing shale reservoir – geologic and petrophysical characterization of unconventional shale-gas reservoirs: SPE131350.
Mehdi E. Far is a Principal Scientist in the Integrated Interpretation Group, Halliburton Technology Center in Houston, since 2014. He received a Ph.D. degree in geophysics from University of Houston. His research focuses on rock physics, petrophysics, advanced interpretation methods for sonic logs, and geomechanics. Mehdi E. Far is a member of SEG, SPWLA, SPE, and EAGE and a member of SEG research committee.
Pickett. G R., 1963, Acoustic character logs and their applications in formation evaluation: J Petr. Tech., Vol. 15, 650-667. Pope, C., Peters, B., Belton, T., and Palish, T., 2009, Haynesville Shale – one operator’s approach to well completions in this evolving play: SPE 125079 presented at the Annual Technical Conference and Exhibition, SPE, New Orleans, Louisiana, USA, 4-7 October. Quirein, J., Witkowsky, J., Truax, J., Galford, J., Spain, D., 8
SPWLA 56th Annual Logging Symposium, July 18-22, 2015
research interests are geomechanics, rock physics, and advanced well logging interpretation techniques. He is currently a member of SPE and SPWLA.
Ming Gu is a Senior Scientist in the Integrated Interpretation Group, Halliburton Technology Center in Houston, since Feb. 2014. He graduated from the University of Texas at Austin with a Ph.D degree in petroleum engineering in 2013. His current research focuses on petrophysics, geomechanics, rock physics, formation testing and reservoir fluid analysis. He is a member of SPWLA.
Jim Witkowsky is the technical advisor for the Halliburton Petrophysics Technology Development group working with GEM elementals and mineral analysis in shale plays. He graduated from the University of Pittsburgh with a degree in chemical/petroleum engineering and began his career in 1986 as a general field engineer for Schlumberger, working primarily in the Gulf of Mexico. Witkowsky worked at the NUMAR corporation from 1992 to 1997, where he held various positions, including NMR field engineer for USA, Indonesia, and Australia, manager of Australia field operations, and petrophysicist. After the merger, he was assigned to the Halliburton Formation and Reservoir Solutions Center as a NMR petrophysicist, responsible for client products, interpretations, and NMR training. He is a member of SPE and SPWLA.
Deepak Gokaraju is a Senior Scientist in the Integrated Interpretation team of Halliburton Formation and Reservoir Solutions in Houston with a focus on providing integrated solutions using advanced interpretation methods. He graduated with a Master’s degree in Petroleum Engineering from Missouri University of Science and Technology. His
9
