SPWLA 54th Annual Logging Symposium, June 22-26, 2013

FORMATION EVALUATION CHALLENGES IN TAMAR FIELD,

OFFSHORE ISRAEL
Christopher J. Christensen and Garwin Powers, Noble Energy, Inc.
Copyright 2013, held jointly by the Society of Petrophysicists and Well Log Including the discovery well, five key appraisal wells
Analysts (SPWLA) and the submitting authors.
This paper was prepared for presentation at the SPWLA 54th Annual Logging
have been drilled and logged. Conventional cores were
Symposium held in New Orleans, Louisiana, June 22-26, 2013. acquired in three of these wellbores and a total of 143
meters of core has been captured to date.
ABSTRACT
Triple-combo wireline logs from the key appraisal well
The Tamar-1 gas discovery in early 2009 was the first used to build the core-calibrated log analysis workflow
in a series of successful exploration wells drilled by discussed in this paper are shown in Figure 2. As the
Noble Energy and its partners in the deepwater Levant discussion of reservoir and fluid properties that follows
Basin (Eastern Mediterranean Sea). Noble estimates will show, display of density and neutron porosity logs
recoverable gas resources from Tamar Field of 10 TCF. on a sandstone matrix / fresh water scale is appropriate.
Field appraisal has been integrated with the Phase 1 However, many of the quicklook log interpretations
development program and initial production from one might make on data displayed in this manner are
Tamar Field began on March 31, 2013. Phase 1 artifacts of either mudstone lithology variation or
production is supplied from five production wellbores, environmental effects on log responses associated with
each designed to produce up to 250 MMscf/day. the high-salinity, water-based drilling mud.

The Oligocene-Miocene reservoir section in Tamar ROCK AND FLUID PROPERTIES

Field is a sequence of deepwater turbidite sandstones
interbedded with variable amounts of siltstone and Temperature & Pressure - Both reservoir temperature
mudstone. Clean end-member sands are quartz arenites and pressure in the Tamar Field reservoir section are
(>=95% quartz), with porosity > 20% and permeability relatively benign. Formation temperature in the
> 500 mD. Although evaluation of Tamar reservoirs is reservoir section is less than 200 F and the objective
a classic shaly sand analysis problem, there are two interval has been logged with mud weights ranging
factors that make application of many industry-standard from 11.0 to 11.4 ppg.
evaluation methods problematic. The first complicating
factor is the impact that high-salinity, water-based mud Drilling Fluids - There is a ~1500 m thick evaporite
has on logging tool responses. The second major sequence in the shallow overburden above Tamar Field
interpretation issue is the presence of complex, variable that consists mostly of salt (halite), with some
mudstone lithology, which breaks the fundamental interbedded anhydrite and clastics (Figure 1). All
assumption of uniform shale properties used in many drilling to date in Tamar Field has been performed with
log-based evaluation techniques. Failure to understand a water-based mud that is pre-saturated with salt prior
and account for these issues has a significant, material to drilling the evaporite section. After the evaporite
impact on gas-in-place estimates. section has been drilled and cased off, mud salinity is
still maintained at a high level, for wellbore stability
Conventional core acquired in the first appraisal well purposes. There are smectite-rich, swelling clays in the
was instrumental in identifying the formation mudstones between the base of the evaporites and the
evaluation issues above. Addition of NMR logging top of the reservoir section and high mud salinity
measurements to formation evaluation programs in retards shale reactivity relative to lower salinity mud.
subsequent wells has led to significant improvements in As a result, the reservoir interval in Tamar Field has
log-based estimates of both lithology and porosity. been logged with mud filtrate salinity in the 230-240
kppm range in all five appraisal wells. At down-hole
INTRODUCTION temperature and pressure, this high salinity results in a
mud-filtrate with a calculated density of 1.17 to 1.18
The Tamar gas field is located in the Eastern g/cc and a calculated hydrogen index of 0.89.
Mediterranean Sea, offshore Israel, in approximately
1700 m of water (Figure 1). The trap for the field is a Reservoir Fluids - The gas in Tamar Field is dry gas
large four-way anticline, cross-cut by multiple faults. with low condensate yields (< 5 stbbl / MMscf). Gas

1
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

properties calculated using the gas gravity of methane correlation between clay volume from core analysis and
(0.554), reservoir temperature, and reservoir pressure, Th concentration reported in spectral gamma-ray
match insitu density estimates from pressure gradients service field deliverables. Neutron / density based
of 0.256 g/cc (0.111 psi/ft). methods cant be used in the gas leg due to hydrocarbon
effects on both logs. Elemental analysis via neutron-
Formation water salinity in Tamar Field is relatively induced, gamma-ray spectroscopy is not an option, as
fresh (< 20 kppm), but cannot be accurately measured logging tools will not work in high-salinity mud.
from samples captured to date with wireline testing
tools, due to unknown amounts of high-salinity mud- Within reservoir intervals, there is a linear relationship
filtrate contamination. between total gamma ray (GR) and core shale volume
defined on a grain size basis (Figure 3). However, the
Reservoir Properties - Within reservoir intervals, clay-rich shales interbedded with reservoir sands are
lithology and mineralogy is relatively simple, with never thick enough to be fully resolved by logging
reservoir units made up of variable mixtures of two tools. If a GRshale baseline is selected in the bounding
end-member lithologies: clean sands and shales (clay- marls above or below the reservoir sands, resulting GR-
rich mudstones). Clean end-member sands are greater based calculations significantly overestimate shale
than 95% quartz, with minor amounts of feldspar and volume and under-estimate net-to-gross.
other minerals. Clean sands have porosities greater
than 20% and permeabilities greater than 500 mD. Nuclear magnetic resonance (NMR) based lithology
Sands are unconsolidated to slightly-consolidated and identification is the only method tested to date that
there is no significant change in reservoir quality across allows for consistent lithology identification from logs,
the gas water contact. without the need for well-specific calibration to core.
Following fluid corrections to NMR total porosity
End-member shales are ~60% clay minerals. Within (methodology discussed in the next section of this
reservoir intervals, there is a good correlation between paper), the ratio of NMR clay bound water volume
shale volume defined on a grain size basis and total clay (standard 3 msec T2 cutoff) to total porosity, rescaled
mineral volume calculated from core X-ray diffraction between 0 (clean sand) and 1 (mudstone) does a good
(XRD) mineralogy data. job in matching core shale volume defined on a grain
size basis. (Figure 4). This is due to the strong
Mudstone Lithology - There is a major mineralogy relationship between grain size and pore size
change between mudstones interbedded with reservoir distributions in unconsolidated to slightly consolidated
sands and bounding mudstones above and below clastics.
reservoir intervals. While mudstones interbedded with
reservoir sands are true shales (clay minerals as the As NMR logs are expensive to acquire, due to slow
most abundant mineral fraction), bounding mudstones logging speeds and the high cost of associated
are marls (carbonate minerals as the most abundant deepwater drilling rig time, acquisition to date has been
mineral fraction). This lithology change invalidates the primarily limited to the gas column. Short sections of
primary assumption used in many shaly sand log the capping mudstones and underlying aquifer sands
analysis techniques, that parameters selected in have been included in logged intervals, for analysis
bounding mudstones can be applied in analysis of calibration and parameter selection purposes.
reservoir intervals.
In Nobles current log analysis workflow, final shale
LOG ANALYSIS ISSUES volume is defined on a grain size basis and is
calibrated to measurements made on core. As the
Lithology - The high-salinity water-based mud used in shale volume definition used is grain size rather than
all Tamar wells drilled to date severely limits wireline mineralogy based, it includes both true shales (clay-rich
logging service options for lithology identification. mudstones) and marls (carbonate-rich mudstones).
Mud weights and barite concentrations in drilling muds Final log-based calculations are gamma-ray (GR)
have been high enough that photoelectric (PE) based based, with the objective section zoned into separate
methods dont work (over-estimation of calcite volumes non-reservoir (marls) and reservoir (sands and shales)
due to barite effects). Natural gamma-ray spectroscopy intervals. Shale volume estimates based on NMR logs
methods are problematic, as invasion of potassium are used to guide GR baseline picks in the gas column,
chloride (KCl) in mud-filtrate causes a large tail in the on an interval-by-interval basis. Outside the gas
potassium (K) spectrum that overlaps with uranium (U) column, neutron / density porosity separation (NDS) is
and thorium (Th) spectrums. As a result, there is no used as a guide for GR-based interpretations.

2
 SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Porosity - Invasion of high-salinity mud-filtrate into the Water Saturation - Water saturation is the most
formation is deep enough to affect all logging significant remaining petrophysical source of
measurements used in porosity calculations. In the uncertainty in Tamar Field gas-in-place estimates. No
water leg, a good match between neutron and density direct core calibration data exists, as measurement of
porosity can be obtained in clean thick-bedded sands, if insitu (unflushed) water saturation is not possible with
both logs are calculated on a sandstone matrix, the water-based mud system currently in use and no
assuming all fluid seen by the tools is high salinity attempt at tracing the mud-filtrate and correcting for
mud-filtrate (Figure 5). core invasion has been made to date.

However, in shaly intervals in the gas leg, log-based The dominant shale habitat in Tamar reservoirs is
porosity calculations are a complicated, three fluid dispersed, pore-filling shale (Figure 7). With the
interpretation problem: (1) immobile formation water combination of dispersed shale and low salinity
with very low salinity, (2) high-salinity mud filtrate that formation water, the Waxman-Smits (1968) model is
invaded the formation, and (3) gas. The interaction of required for rigorous resistivity-based water saturation
non-linear gas effects, non-linear water salinity effects, estimates. Electrical properties measurements required
and clay effects on the neutron log are such that no for implementation of this method are currently in
deterministic or empirical neutron / density based progress, but the existing special core analysis (SCAL)
methods tested worked without well-specific calibration database is inadequate for application of the Waxman-
to core. Smits (1968) model.

Several other methods tested yield poor matches to core In the absence of core-calibrated electrical properties in
porosity. Sonic based methods are problematic, since shaly sand intervals, the normal log interpretation
fluid effects dominate log responses. Accurate approach would be to estimate shale properties from
estimates of porosity and flushed-zone fluid volumes bounding mudstones, with a method like the Juhasz
are required to correct sonic log responses to insitu (1981) Normalized Qv approach. However, it is
conditions, not vice-versa. Using bulk density and unlikely such approaches yield valid results in Tamar.
flushed-zone resistivity (Rxo) for simultaneous solution As noted earlier, bounding mudstones are marls with
of porosity and flushed-zone saturation yielded poor significantly lower clay content than the clay-rich
results, due to environmental effects on microlaterolog shales that are interbedded with reservoir sands (Figure
tool response. The microlaterolog tool used for Rxo 4).
estimates is hypersensitive to minor standoff from the
wellbore / formation interface, and reads values closer After gaps in the Tamar core electrical properties
to Rmud or Rmudcake, than Rxo. NMR-only processing database are addressed by ongoing SCAL programs,
products delivered by the logging vendor significantly there are still some sources of resistivity log-based
under call porosity relative to core. Vendor processing saturation uncertainty that cannot be resolved without
methodology, assumptions, and inputs are poorly switching mud systems. Issues with resistivity
documented, so it is not clear whether this issue is suppression by thin-bedded shales cannot be resolved
inherent in the method used or due to poor input with modern 3D array induction resistivity log methods,
assumptions. as these tools cannot be run in high conductivity muds.
Formation water salinity and resistivity estimates are
The Density Magnetic Resonance (DMR) porosity likely well constrained in clean, thick water bearing
method (Freedman, et al., 1998) is the only log-based sands, with both core-calibrated porosity and measured
method tested to date that yields consistent results electrical properties as inputs to apparent water
without well-specific calibration to core. For DMR resistivity calculations. However, this cannot be
calculation purposes, the standard NMR clay bound confirmed without salinity measurements from
water (T2 < 3 msec) and bulk volume irreducible (3 <= uncontaminated water samples.
T2 < 33 msec) volumes are given formation water
properties and the free fluid volume is assumed to be a A saturation / height modeling workflow, based on core
mixture of high-salinity water-based mud-filtrate and measurements of air-brine (porous-plate) capillary
formation gas. The true T2 cutoff for bulk volume pressure (Pc), has been employed to get around issues
irreducible (BVI) at Tamar has not been established via with resistivity-log based interpretation. Nobles
core-to-log integration to date; however, using the current saturation / height modeling workflow is based
standard value of 33 msec for sandstones appears to be on a Brooks-Corey (1964) capillary pressure model,
adequate for the purposes of fluid-corrected, total with model parameters (shape exponent, entry pressure,
porosity calculations via the DMR method (Figure 6). and irreducible water saturation) all estimated from

3
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

porosity and permeability. The fluid contact is well Freedman, R., Minh, C. C., Gubelin, G., Freeman, J. J.,
constrained, so the only significant sources of McGinness, T., Terry, B., and Rawlence, D., 1998,
uncertainty in Pc-based saturation estimates are the Combining NMR and Density Logs for Petrophysical
somewhat limited dynamic range in reservoir quality Analysis in Gas-Bearing Formations: SPWLA 39th
sampled to date (5% < Swirr < 40%) and minor errors in Annual Logging Symposium, Paper II.
modeled verses measured capillary pressure curves.
Saturation / height modeling is basis for Nobles current Juhasz, I., 1981, Normalized Qv the key to shaly sand
gas-in-place estimates and models will be refined / evaluation using the Waxman-Smits equation in the
updated when ongoing SCAL programs are complete. absence of core data: SPWLA 22nd Annual Logging
Symposium, Paper Z.
IMPACT OF CORE CALIBRATION ON GAS-IN-
PLACE ESTIMATES Thomas, E. C., and Stieber, S. J.,1975, The distribution
of shale in sandstones and its effect upon porosity:
Total hydrocarbon pore volume thickness (HCPV-H) SPWLA 16th Annual Logging Symposium, Paper T.
estimates in the Tamar wells have increased
significantly following development of the core- Waxman, M. H. and Smits, L.J.M, 1968, Electrical
calibrated log analysis workflow discussed in this conductivities in oil bearing shaly sands: SPE Journal
paper. HCPV-H in the Tamar-1 discovery well has (June), pp. 107-22.
increased ~12% relative to initial post-well evaluation.
ABOUT THE AUTHORS
However, not all input components to HCPV-H
calculations have gone up. Net-thickness estimates Christopher J. Christensen is a Petrophysical Advisor
have gone up, with knowledge from core data showing with Noble Energy, Inc., based in Houston, Texas
how clean sands are and development of a log analysis (USA). Since joining Noble in January 2011, he has
workflow to honor this data. Porosity estimates have been embedded in the Eastern Mediterranean business
gone down, as the impact invasion of high-salinity unit, providing petrophysical support for exploration,
water-based mud-filtrate has on porosity log response is appraisal, and development projects offshore Israel and
now understood and properly handled during log Cyprus. He has 22 years of industry experience and has
analysis. Gas saturation estimates have gone up, with worked for Amoco, BP, Newfield Exploration, and
implementation of a capillary pressure-based, saturation Shell. He received a MS degree in Geological Sciences
/ height modeling workflow, in place resistivity log- from the University Southern California in 1991 and
based estimates. joined Amoco Production Company as an exploration
geologist the same year. He attended Amocos
ACKNOWLEDGEMENTS Petrophysics Training Program in 1997-1998 and has
been a full-time, practicing petrophysicist ever since.
The authors would like to thank management of the As a petrophysicist, his primary focus has been core-to-
Tamar partner group (Noble Energy, Delek Drilling, log integration and other petrophysical aspects of multi-
Isramco, Avner Oil and Gas, and Dorgas) for discipline, integrated reservoir modeling projects. He
permission to publish this paper. This work would not has worked on clastic reservoirs ranging in quality from
have been possible without the core calibration dataset micro-Darcy tight gas sands, to multi-Darcy
included in initial appraisal well by former Noble unconsolidated sands. He has worked on projects in a
Tamar subsurface team members Owen Stephens and wide range of clastic depositional environments,
Jonathan ffrench. Special thanks go out to current including fluvial, deltaic, shallow marine, and
members of the Noble Tamar subsurface team, Dan deepwater marine systems, and a wide range of
Needham, Jim Hosler, and Stephanie Nowak, for geographical locations including SW Wyoming (Jonah
supporting both the technical work presented in this Field & Pinedale Anticline), the Anadarko Basin in
paper and recommended changes in subsequent logging Texas / Oklahoma (Granite Wash), Offshore Trinidad,
and core analysis programs. US Gulf-Coast (onshore, shelf, and deepwater), and
Nigeria (deepwater).
REFERENCES
Garwin Powers is Manager of Petrophysics at Noble
Brooks, R. H. and Corey, A. T., 1964, Hydraulic Energy and is based in Houston, Texas (USA). He
Properties of Porous Media: Colorado State University began his industry career in 1981 working for Amoco
Hydrology Paper Number 3, 37 p. Production Company at their Tulsa Research Center.
He completed his Master Degree in Geology from the
4
 SPWLA 54th Annual Logging Symposium, June 22-26, 2013

University of Tulsa in 1992 while working for Amoco. RTM Formation resistivity (field inversion
He started focusing on petrophysics after completing a provided with Rt Explorer service)
year in Amoco's petrophysics schools soon after that. ZCOR Bulk density correction
Since then he has had stops at Statoil, Landmark, and
ZDEN Bulk density
Burlington/Conoco-Philips working as a petrophysicist.
He landed at Noble Energy in 2007 where he worked ZDNC Bulk density (borehole corrected)
their string of large discoveries in West Africa, Eastern
Mediterranean and deep water Gulf of Mexico, before Calculated / Edited Curves
moving into his current position.
DIFCAL Differential Caliper (Caliper
APPENDIX: LOG NMENOMICS ON FIGURES Bitsize)
DPHIS_MF Density porosity calculated on a
Vendor-Supplied Wireline Logs (Baker Hughes) sandstone matrix with mud filtrate
fluid density.
CHAD Characteristic Hole Diameter NDS Neutron / Density Porosity Separation
CNCF Neutron Porosity (Series 2446 CN (NPHIS_MF DPHIS_MF)
Tool); Sandstone Matrix / Freshwater NPHIS_MF Environmentally corrected neutron
Scale porosity (sandstone matrix), with
GR Gamma ray (Total) mud-filtrate salinity used in formation
MBVIC Magnetic resonance bulk volume salinity correction
irreducible (MR Explorer Tool; PHIE_DMR_3F Effective porosity after fluid
NMR110 Processing Service) corrections using the Density
MLR#C Array laterolog resistivity Magnetic Resonance method
(shallowest=1; deepest = 4; Rt (modified 3 fluid version)
Explorer tool) PHIT_DMR_3F Total porosity after fluid corrections
MPHEC Magnetic resonance effective porosity using the Density Magnetic
(MR Explorer tool; NMR110 Resonance method (modified 3 fluid
processing service) version)
MPHSC Magnetic resonance total porosity VSH_NMR Shale volume calculated from the
(MR Explorer tool; NMR110 ratio of NMR clay bound water
processing service) volume to total porosity (after fluid
corrections). Unclipped version is
RD Deep resistivity (Dual Laterolog tool)
VSH_NMR_UNCL
RMLL Microlaterolog resistivity
*_clean Edited vendor-supplied curves with
RS Shallow Resistivity (Dual Laterolog values set to null in bad data intervals
tool)

5
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Figure 1: Overview of Tamar Field showing (A) location map, (B) structure map (top of reservoir), and (C)
structural cross-section (NW to SE) thru the Tamar-1 discovery well (T-1).

6
7
Figure 2: Triple-combo field data from the key appraisal well used to develop the core-calibrated formation evaluation workflow discussed in this
paper (same log plot left and right with and without annotations). Reservoirs and cored intervals are displayed on the left.
SPWLA 54th Annual Logging Symposium, June 22-26, 2013
8
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Figure 3: Plot on the left is a frequency distribution of core shale volume measurements and shows a bimodal distribution with separate clean sand and
grain-size-defined shale modes. Plot on the right is a cross-plot of shale volume calculated from NMR (rescaled ratio of clay-bound water to fluid-
corrected total porosity) vs total gamma-ray. Data points on the crossplot are colored by neutron / density porosity separation. Although there is a trend
of increasing GR with increasing shale volume in reservoir units, interbedded shales are never thick enough to make a GRshale pick in these intervals. If
a GRshale pick made in bounding mudstones (marls) is applied to reservoir intervals, net-to-gross is significantly underestimated.
9
Figure 4: Plots on the right highlight lithology / mineralogy difference between bounding mudstones (marls) and interbedded mudstones (shales). Shale
volume calculated from NMR data (VSH_NMR in right most log track) gets around this issue and is easily rescaled to match core shale volume from grain
size measurements (single linear rescale works in entire NMR log interval).
SPWLA 54th Annual Logging Symposium, June 22-26, 2013
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Figure 5: Density and neutron porosity (sandstone matrix), re-calculated assuming all fluid seen by tools is mud
filtrate (log track 5), compared to original field data provided on a fresh water scale.

10
11
Figure 6: Blind test of total porosity calculated via the Density Magnetic Resonance (DMR) method to new core measurements completed after initial log
calculations. Data points on the cross-plot are limited to just reservoir intervals, where matrix density estimates from initial core calibration are valid.
Best-fit line with forced intercept of zero has a slope very close to 1. Data scatter is mostly due to vertical resolution differences.
SPWLA 54th Annual Logging Symposium, June 22-26, 2013
12
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Figure 7: Modified Thomas-Stieber (1975) crossplot of core porosity vs cores shale volume from grain size measurements (plot on right) suggests
dispersed shale is the dominant shale habitat in Tamar reservoirs, at least at the core plug scale. This is verified by thin-section analysis of end-trim
material from core plugs (plot on left).