Accepted Manuscript

Chemostratigraphy of the Ordovician Sarah Formation, North West Saudi Arabia: An

integrated approach to reservoir correlation

Neil W. Craigie, Andrew Rees, Keith MacPherson, Seth Berman

PII: S0264-8172(16)30235-5
DOI: 10.1016/j.marpetgeo.2016.07.009
Reference: JMPG 2616

To appear in: Marine and Petroleum Geology

Received Date: 3 January 2016

Revised Date: 3 March 2016
Accepted Date: 7 July 2016

Please cite this article as: Craigie, N.W., Rees, A., MacPherson, K., Berman, S., Chemostratigraphy
of the Ordovician Sarah Formation, North West Saudi Arabia: An integrated approach to reservoir
correlation, Marine and Petroleum Geology (2016), doi: 10.1016/j.marpetgeo.2016.07.009.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
CHEMOSTRATIGRAPHY OF THE ORDOVICIAN SARAH
FORMATION, NORTH WEST SAUDI ARABIA: AN
INTEGRATED APPROACH TO RESERVOIR CORRELATION

Neil W. Craigie, Andrew Rees, Keith MacPherson and Seth Berman

ABSTRACT

PT
Late Ordovician pre-glacial and syn-glacial deposits form economically important hydrocarbon
reservoirs throughout North Africa and the Arabian Peninsula. Internal subdivision of the unit is
difficult due to the short time span of deposition resulting in coarse biostratigraphic control. The
arenaceous character of the sediments has also contributed to poor biostratigraphic recovery. The

RI
present study focuses on the chemostratigraphic correlation of this interval encountered in ten wells
across a complex, multi-phased pro-glacial outwash fan dominated by hyperpycnal (jet-efflux)
turbidites fed by glacial outburst floods. The high resolution reservoir layering scheme presented here
utilizes detailed sedimentological core descriptions, borehole image-derived palaeoflow indicators, 3D

SC
seismic attributes and chemostratigraphy.

The Inductively Coupled Plasma – Optical Emission Spectrometry (ICP-OES) and Inductively Coupled
Plasma – Mass Spectrometry (ICP-MS) techniques were used to acquire data for 50 elements, with the

U
chemostratigraphic scheme based on changes in the following ‘key’ element ratios: Zr/Y, Zr/Lu,
Th/Nb, Th/P, Ta/P, Cr/P, P/Lu, Th/Ti, Cr/Yb and Th/Ta. G may give the impression that there are two
principal aims are defined by distinct values of key element ratios, with variations in these mainly
AN
reflecting changes in source/provenance. In order to avoid complications relating to changes in grain
size/lithology, separate schemes were proposed for sandstone and mudrock lithologies, with the former
comprising a hierarchical order of two zones, three subzones and seven divisions. Owing to the lower
number of analyzed mudrocks, the scheme produced for this dataset is less detailed but, nevertheless,
M

consists of three principal zones and five subzones.

At the formation level, chemical stratigraphy allowed for consistent identification and correlation of the
D

pre-glacial Qasim, syn-glacial Sarah and post-glacial Qalibah formations. The earliest post-glacial unit
referred to as the Uqlah Formation has an inherited signature from the syn-glacial chemozone (i.e.
uppermost Sarah Formation) below, and represents initial reworking of the post-glacial landscape
TE

following isostatic rebound. Evidence for glacial isostacy is recorded in a regional deformed zone at
the top of the Sarah Formation and immediately beneath an isostatic rebound unconformity at the base
of the Uqlah Formation. Chemical zonation of this unit clearly illustrates significant local truncation in
wells located close to NW-SE trending lineaments. It is proposed that chemozones are missing across
EP

reactivated fault block highs as a result of increased erosion on the shallow marine post glacial shelf.

High resolution reservoir layering of the hydrocarbon-bearing syn-glacial Sarah Formation highlights a
complex depositional system characterized by an initial phase of deeply entrenched pro-glacial
C

turbidite channels and younger sheet-like depositional lobes. Cores and image logs from both elements
of the outwash fan system record extensive climbing dune cross-bedding, an unusual sedimentary
structure described from many Upper Ordovician glacial successions in North Africa, and considered
AC

indicative of quasi-steady sediment-charged flows initiated by glacial outburst floods. There is no

shared chemical link between the lowest channels and overlying lobes, suggesting a large fan system
may be found downdip of this deeper channel system. Integration of the zonation scheme with 3D
seismic attribute analysis clearly illustrates a second and even third phase of small channels cross
cutting the lower larger vintage that fed the sheet-like younger lobe system. This is substantiated by a
shared chemical signature and provenance between these small channels and their attached lobes.

Saudi Aramco: Company General Use

 ACCEPTED MANUSCRIPT
1. Introduction and Objectives

Hydrocarbon exploration of the Upper Ordovician Sarah Formation in Northwestern Saudi

Arabia has relied upon simplistic models, integrating conventional sedimentological datasets acquired
from well bores, traditional lithostratigraphic correlation and various 3D seismic attribute extraction
techniques. The geological model used for static and dynamic reservoir simulation suffered from poor
resolution due to inherent uncertainties regarding reservoir layering in this rapidly deposited,
arenaceous phase of glacially-influenced sedimentation. Geochemical data were acquired from core
and cuttings samples, with the objective to produce a high resolution layering scheme for the Sarah
Formation encountered in 10 wells in Northwestern Saudi Arabia (Fig. 1). The study focuses on the
chemical zonation of clastic sediments that are beyond the resolution of traditional biostratigraphical

PT
techniques, and includes integration of seismic, sedimentological and borehole image data to
understand the temporal and spatial evolution of this complex pro-glacial depositional system.
There is limited published material on the Sarah Formation throughout Saudi Arabia, with this
contribution being one of the first geological account of these hydrocarbon-bearing reservoirs in the

RI
subsurface. The Late Ordovician glacial event is widely recognized across Africa (Ghienne, 2003; Le
Heron et al., 2006; Ghienne et al., 2007) and Arabia (Melvin, 2015; Rees, 2015), occurring during the
late Ashgillian Hirnantian stage (Fig. 2) and persisting for approximately one million years (Sutcliffe et

SC
al., 2000). There is general consensus that the Sarah Formation across Saudi Arabia was deposited in a
range of fluvio-glacial and glaciomarine environments (Vaslet, 1990; Janjou et al., 1997; Melvin and
Miller, 2009; Melvin, 2015). Palynological data and detailed facies analysis from cores in the study
area indicate wholly glacio-marine conditions (Rees, 2015), with tentative models suggesting glacial
outwash fans dissected by narrow, elongate ribbon channels (Figs 3b and 4). Stratal slices through 3D

U
seismic relative acoustic impedance volumes highlight this network of submarine channels (Fig. 1c),
which developed basinward, and to the northeast of the regionally defined maximum grounding line of
AN
the ice (Fig. 4b). These channels are largely oriented towards the northeast (Fig. 1c), with a subset
aligned broadly to the southeast. Some wells, such as 2, 4 and 6, targeted channelized reservoir sands,
while others (e.g. Wells 1, 5, 8 and 9) were drilled to test inter-channel play concepts.
The complexity of Hirnantian age glacial systems and the reservoirs that subsequently
M

developed have received considerable attention in recent years, particularly the large petroliferous
provinces in Southeast Algeria and Southwest Libya (Hirst et al., 2002; Le Heron et al., 2004, 2006;
Ghienne et al., 2010). While large sub-glacial tunnel valleys remain the primary exploration target in
these areas (e.g. the Tiguentourine Field – Lang et al., 2012), their pro-glacial downdip outwash fan
D

extensions (the “Cordons” and “Grès á rides” units in Algeria – Beuf et al., 1971) have appreciable
volumes of reservoir grade material (Deschamps et al., 2013). This sediment is sourced by pressurized
TE

meltwater incision of the pre-glacial stratigraphy during tunnel valley formation (Rees, 2015), with jet-
efflux deposits forming a series of ribbon channels and lobes on rapidly constructed lowstand pro-
glacial fans. The ultimate aim of this study is to describe for the first time in Saudi Arabia the facies
that characterize this fan system, how the fan evolves through time and space and to demonstrate how
chemical stratigraphy can only be fully utilized when integrated with other geological/geophysical
EP

disciplines.

2. Geological Setting
C

According to Sharland et al. (2001), the Arabian Plate occupied an intra-cratonic location at
relatively high southern latitudes (c. 45° S), with a passive margin interpreted to have been present
AC

between this part of the Gondwana margin and Paleo-Tethys in Ordovician times. During this time, the
plate drifted into lower southern latitudes of less than 40° S. An ice sheet covered much of the African
platform and Western Arabia (Vaslet, 1989, 1990; Eyles, 1993; Sutcliffe et al., 2000), and was
responsible for depositing the glaciogenic sediments of the Sarah Formation (Clarke-Lowes, 1980;
Williams et al., 1986). An unconformity exists between the Qasim and Sarah Formations (Fig, 2),
produced by glacial advance, and with some of the former locally removed by erosion. This period of
ice sheet expansion resulted in deposition of intensely deformed glaciotectonites of the Zarqa
Megafacies (Melvin, 2015), which is in turn overlain by erosive, sediment gravity-flow dominated
sandstones laid down during glacial retreat (Fig. 3a).
During glacial maximum or early glacial retreat (Fig. 3b), pressurized release of water resulted
in the cutting of deep incisions into the underlying Qasim Formation, eroding a network of
palaeovalleys (Melvin, 2009) and liberating vast quantities of sand to be deposited in the form of pro-
glacial ribbon channels and fans at the terminus of tunnel valleys (Rees, 2015). These may be
somewhat similar to modern day trough mouth fans (O´Cofaigh et al., 2003; Dowdeswell et al., 2008).

Saudi Aramco: Company General Use 2

 ACCEPTED MANUSCRIPT
Palynological data indicates a late Katian to early Hirnantian age for these fans (Fig. 4e), which is older
than the sediments that typically occupy the palaeovalley fills. It is likely that the valley system
operated as a bypass zone, funneling sediment onto a lowstand apron. The main phase of glacial retreat
(Fig. 3c) is marked by ice contraction to the south and southwest with the main zone of ice proximal
deposition migrating into the exposed valleys, leading to more condensed fine grained deposition on
the abandoned lowstand fan. Palaeovalleys are dominated by a fining upward retrogradational
architecture (Fig. 3a) of ice proximal to ice distal high and low density turbidites (Melvin, 2009).
The Sarah Formation is capped regionally by relatively condensed dropstone diamictite and
low density turbidites that are locally organic-rich. This unit essentially marks deglaciation and
breakup of the ice sheet in the area (Fig. 3d). The uppermost 5 to 15 ft form a regionally correlatable
deformed zone (Figs. 5a and 6a) characterized by slumping, sliding, dewatering and local shearing

PT
(Fig. 5f). This regional instability is caused by isostatic rebound and is variably truncated by a post-
glacial rebound unconformity (PGRU) (Melvin and Miller, 2009). Differential unloading and fault
reactivation led to source area rejuvenation and deposition above the rebound unconformity of the
fluvial to shallow marine, post-glacial Uqlah Formation (Janjou et al., 1996). In the study area, this

RI
formation is wholly shallow marine, completely undeformed and dominated by tidal features (Fig. 5a).
The chemical fingerprint of this unit strongly suggests that much of the sediment was sourced by
erosion and reworking of the uppermost Sarah Formation beneath the unconformity. This is discussed

SC
in more detail in later sections of the present paper.
Transgressive sandstones of the Uqlah Formation are overlain by grey to black, locally
organic-rich, mudstones and claystones of the Qusaiba Member (Fig. 2). These often condensed
sediments represent widespread flooding of the glaciated margin during the Lower Silurian as eustatic
sea level rose in response to the retreating Gondwanan ice sheet. Although sampling of this member

U
was limited, a small number of samples were taken in each well in order to define the top of the Sarah
Formation. Some samples were also taken from the underlying Qasim Formation in order to
AN
characterize the Qasim:Sarah formation boundary. The Qasim Formation is Middle-Upper Ordovician
in age and was deposited during a period of tectonic quiescence resulting in stable marginal marine
shelf conditions (Senalp and Al-Duaji, 2001). Although mudrocks, sandstones and siltstones have been
recorded in the Qasim Formation, only the uppermost part of the interval (Quwarah Member) has been
M

sampled in the present study, where it is comprised of mudrocks. For further information on the Qasim
Formation, the reader is referred to the works of McGillivray and Husseini (1992); Williams et al.,
(1986); Vaslet (1990); and Senalp and Al-Duaiji (2001).
D

3. Sedimentology
TE

Conventional cores form the primary database for sedimentological description and
interpretation, with all ten wells in the study having at least 60 ft of coverage. Cores were described at
1:120 scale and coded for lithofacies and then grouped into genetic facies associations (see Fig. 5). A
cross section across the turbidite fan system is illustrated in Figure 6, and highlights the broad
EP

architecture of deeply incised ribbon channels (e.g. Wells 2, 6 and 4) and thinner inter-channel
sandstones (Wells 5, 9, 10 and 17). Six facies associations have been identified based on texture, fabric
and palaeoflow direction resulting in the lithostratigraphic correlation outlined in Figure 6 and
described in the following paragraphs. An example core description from Well 6 is presented in Figure
5, outlining the vertical stacking of facies associations with descriptions and core photographs.
C

3.1 FA1 – Storm influenced mid to outer shelf heterolithic (Quwarah Member of the Qasim Formation)
AC

The oldest sediments in the study comprise laminated, locally pyritic and carbonaceous silty
mudstone interbedded with thin coarse silt to very fine grained sandstone starved ripple trains. These
thin discontinuous lenses comprise current ripple cross lamination, loaded sunken bases and rare pyrite
nodules. The lack of wave ripples implies deposition below wave base, while the starved lenticular
character of bedding reflects weak unidirectional currents and limited sediment supply. The laminated
mudstone records low energy suspension sedimentation, while the thin event beds may reflect distal
storm processes operating on the mid to outer shelf. These distal Qasim facies are variably truncated by
syn-glacial processes associated with deposition of the overlying Sarah Formation.

3.2 FA2 – Climbing dune cross bedded channel fill sandstones (Sarah Formation)

Saudi Aramco: Company General Use 3

 ACCEPTED MANUSCRIPT
FA2 lithofacies form 50 - 300 ft of fining and ‘muddying’ upward sequences comprising thick
bedded, moderately to well sorted, medium to fine grained variably stratified, unbioturbated sandstones
with local mudstone rip-up clasts and conglomeratic lags. These deposits form the base of the Sarah
succession in wells that demonstrate significant incision into the Qasim Formation below, and occupy
the shoestring or ribbon geometry channels (Fig. 4). Although local massive intervals occur
sporadically, FA2 sandstones are dominated by dune scale, aggrading undulatory lamination
interpreted as climbing dune cross stratification (Figure 5d). This conspicuous sedimentary structure
had garnered much attention in recent years (Ghienne et al., 2010,) and has become synonymous with
Late Ordovician peri-Gondwanan glacial outburst deposits in flood dominated pro-glacial systems
(Deschamps et al., 2013). These have been particularly well documented in Hirnantian age sediments
occupying ribbon channel (Cordon Unit) and attached depositional lobes (Gres a Rides Unit) in SE

PT
Algeria (Hirst et al., 2002; Lang et al., 2012; Deschamps et al., 2013).
The deeply incised character of this unit implies high energy erosional flows, while climbing
dune cross stratification reflects lower to transitional upper flow regime processes with very high sand
fallout (Ghienne et al., 2010). An usual feature of cross bedded sands in FA2 seen on oriented image

RI
logs is apparent dune progradation perpendicular to the long axis of the channel (Fig. 6a). Similar
across channel palaeoflow orientations were described from climbing dune facies in the Murzuq Basin,
Libya by Ghienne et al., (2010), which they attributed to bank attached eddy-related sidebars (Fig. 6c).

SC
The local conglomeratic sands in this model occupy the channel thalweg, which is eventually covered
by the migrating sidebar resulting in fining upward cycles.
The sustained accretionary nature of climbing dune cross bedding, coupled with very high
sediment supply, high velocity flows and association with other pro-glacial lithofacies, has led to many
interpretations invoking hyperpycnal jets emanating from the ice front, possibly via more proximal

U
sub-glacial tunnel valleys (Fig. 3b) during catastrophic outburst floods.
AN
3.3 FA3 – Climbing dune cross bedded sheet sandstones (Sarah Formation)

The main characteristics of this facies are similar to those that occupy the ribbon channels
described above. Moderately to well-sorted fine to medium grained sandstones dominate exhibiting
M

very well-developed climbing dune and climbing ripple cross lamination. An image log example of the
former is illustrated in Figure 5b, and shows aggrading undulating lamination in a 9 ft thick bed, with
no obvious internal erosion or reactivation surfaces. These features are consistent with long-lived
steady state hyperpycnal flows operating over fairly long timespans (Deschamps et al., 2013), typically
D

in the range of weeks to months. Waxing flow features reflecting deposition from currents that are
accelerating over time are often considered primary indicators (along with steady-state bedforms) of
TE

hyperpycnal underflows. An example of this inversely graded fabric is seen in Figure 5e, where over a
1 inch interval; muddy laminae die out and are replaced by increasingly thicker sandstone layers as the
current gained velocity and carrying capacity.
The main differences between FA2 and FA3 are related to vertical architecture and lateral
geometry. While FA2 displays the salient attributes of a confined channel, FA3 climbing dune
EP

sandstones typically show 30 – 80 ft thick blocky to coarsening / cleaning upward profiles, that occur
above typical channel fills and in inter-channel areas (Fig. 6a). This sheet like geometry also has a
consistent palaeoflow direction, with sediment transport directed towards the Northeastern Saudi
Arabia in all wells. This lack of confinement, consistent flow direction as well as local indictors of
C

weak wave scour and tidal modulation suggest deposition on a broader fan, possibly as amalgamated
progradational lobes. A similar relationship between lower feeder channels and upper lobe sheets were
AC

described from age equivalent “Cordon” units in Southeastern Algeria by Beuf et al., (1971), Hirst et
al., (2002) and Deschamps et al., (2013).

3.4 FA4 – Slumped and dewatered regionally deformed heterolithics (Sarah Formation)

Facies FA4 is dominated by a sequence of cross laminated to parallel laminated normally

graded sandstones and siltstones with local mudstone interbeds. The upper part of the succession is
variably truncated by an erosional surface at the base of the overlying Uqlah Formation (Fig. 2). The
lower part of FA4 is relatively undeformed with minor convolute bedding and disharmonic folding
indicative of syn-sedimentary soft-sediment deformation. The most pervasive deformation is within the
uppermost part of the interval and is characterized by inclined to recumbent asymmetrical folds, more
complex disharmonic folds, liquefaction and injection, small scale brittle extension (normal faults) and
localized brittle shearing (Fig. 5f).
This predominantly ductile style of deformation forms a regionally mappable unit across the

Saudi Aramco: Company General Use 4

 ACCEPTED MANUSCRIPT
study area (Fig. 6a), and has been described 300 km to the south beneath the post-glacial Uqlah
Formation in the outcrop belt (John Melvin, pers. comm. 2015). While an end Sarah glacial re-advance
cannot be ruled out, the dominantly ductile stress response coupled with the lack of an overlying retreat
phase, suggest that this regional event had origins other than ice sheet expansion. A more viable
alternative is glacial isostasy, where differential ice unloading led to fault reactivation and regional
instability. This is consistent with the base Uqlah erosion surface being interpreted as a post glacial
rebound unconformity in the outcrop belt (Melvin and Miller, 2009; Melvin 2015). Similar styles of
deformation have been described beneath post-glacial shelf deposits in Libya (Le Heron et al., 2006),
and are also attributed to glacial isostasy.

3.5 FA5 – Tidally influenced, locally bioturbated cross bedded sandstones (Uqlah Formation)

PT
The base of FA5 is marked by significant erosion and facies change, with the sharp contact
(littered with granule to pebble sized intraclasts and exotic lithoclasts) marking a change from
deformed low density turbidites of the Sarah Formation to undeformed cross bedded and variably

RI
bioturbated sandstones of the Uqlah Formation. This broadly fining upward unit is characterized by
coarse to fine grained cross-bedded and current rippled pyritic sandstone with abundant black shale
drapes and local Planolites, Skolithos and Palaeophycus burrows. Double mud drapes, complex bi-

SC
direction ripple forms, synaeresis cracks and local fluid muds reflect tidal modulation on a shallow
marine shelf (Reineck and Wunderlich, 1968). The abundance of pyrite and black organic-rich shale
drapes suggest that periods of anoxia were common, a characteristic that dominates the overlying
Silurian hot shales of the Qalibah Formation (facies FA6).
The geometry of FA5 is sheet-like forming a regionally correlatable unit in both the outcrop

U
belt (Janjou et al., 1996) and subsurface (Fig. 6a). The facies dislocation documented at the base of the
formation has been interpreted as a PGRU (Melvin, 2015), separating deformed late syn-glacial
AN
turbidites of the Sarah from undeformed post glacial tidal sandstones of the Uqlah. The unit is
transgressive in character with an increase in density and diversity of marine trace fossils towards the
contact with FA6. This architecture is typical of post glacial sediments deposited during rising eustatic
sea level linked to a demise of polar icecaps.
M

3.6 FA6 – Organic-rich graptolitic black shale/chert (Qusaiba Member of the Qalibah Formation)

The sub-oxic to anoxic conditions documented in facies FA5 reach a peak in the Lower
D

Silurian black shales of FA6. This world class source rock interval is characterized by organic-rich,
pyritious, graptolitic, massive to laminated shales, cherts and argillaceous siltstones deposited on a low
TE

energy anoxic shelf. These condensed sediments reflect marine flooding of the glaciated landscape
(Melvin, 2015) and are consequently diachronous in character, reflecting extreme topographic variation
following ice retreat and isostatic rebound.

4. Chemostratigraphy
EP

4.1 Materials and methods

C

A total of 1100 core and cuttings samples were analyzed in the present study. In general,
cuttings samples were taken every 10 ft, while core material was selected at 3 ft intervals. The cuttings
samples were first washed in liquid detergent to remove drilling additives and then dried in an oven.
AC

The samples were then sieved to remove the finest dust fragments (< 10 microns) and the largest
fragments (> 3mm). The latter were most likely to have been 'caved' from sediments occurring above
the depths from which the cuttings were taken. A magnet was then placed over the surface of each
sample to remove any magnetic metallic fragments. The samples were later examined under a
microscope and ‘picked’. This process involved the selection of fragments that were considered to be
representative of a given lithology at a given depth. By examining conventional wireline and mud log
signatures, it was possible to determine the dominant lithology in any given sample and only fragments
of such a lithology were picked in each sample. Approximately 60 - 80 fragments were selected in each
sample. This material was then ground to a fine powder (of particles less than 10 microns) using an
agate grinder. Core samples were also micronized in this way.
ICP-OES and ICP-MS techniques were used to acquire data for 50 elements, including all of
the major elements and a large number of trace and rare earth elements. The powdered samples were
liquefied using the alkali fusion preparation technique. This involved mixing the samples with a lithium
metaborate flux, heating each sample to 1000° C for a period of 30 minutes and then dissolving this

Saudi Aramco: Company General Use 5

 ACCEPTED MANUSCRIPT
mixture in hydrochloric acid, prior to insertion into the ICP machines. The ICP-OES technique
measures abundances of major (e.g. K, Na, Si, Al) and high concentration trace elements (e.g. Zr, Sr,
Ba), while ICP-MS is employed to acquire data from the low abundance trace elements (e.g. Rb, Cs,
Ta) and rare earth elements (e.g. La, Lu, Yb). Data for all major elements are provided as wt% oxide
values but, for brevity, the symbols, rather than the full oxide symbols are quoted in the present study.
For example Mn is quoted instead of MnO.
In each sample, data were acquired for the major elements Al, Si, Ti, Fe, Mn, Mg, Ca, Na, K
and P, the trace elements Ba, Be, Co, Cr, Cs, Cu, Ga, Hf, Mo, Nb, Ni, Pb, Rb, Sc, Sn, Sr, Ta, Th, Tl, U,
V, W, Y, Zn and Zr, and the rare earth elements (REEs). All major elemental data are quoted as %
oxide values, while trace and rare earth element concentrations are in ppm (parts per million). The
REEs comprise the elements La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb and Lu, and decrease

PT
in ionic radii from La to Lu. They are normally subdivided into three main groups: light rare earth
elements (LREEs) – La to Nd; middle rare earth elements (MREEs) – Sm to Tb; and heavy rare earth
elements (HREEs) – Dy to Lu. Pm is another REE but data were not acquired for this element in the
present study.

RI
On acquiring the data, an attempt was made to establish the mineralogical affinities of
elements by employing graphical and statistical techniques. After this, data for elements and elemental
ratios were plotted in the form of geochemical profiles, with separate profiles produced for mudrock

SC
and sandstone lithologies. Chemozones (a general term used to describe any chemostratigraphic zone,
subzone or division- see section 4.4 for further explanation) were distinguished using a number of
‘key’ elemental ratios, variations in which largely reflect changes in source/provenance. Details of the
element:mineral links and chemostratigraphic scheme are provided in the following paragraphs.

U
4.2 Geochemistry and mineralogy
AN
Prior to proposing a chemostratigraphic scheme, it is necessary to establish the associations of
elements and minerals and to understand the controls on geochemistry and mineralogy (e.g.
provenance, weathering, diagenesis etc.). The mineralogical affinities of elements were first established
by comparing the geochemical data with available petrographic data. However, it was also possible to
M

establish the element:mineral links using graphical and statistical techniques. Details of
element:mineral links established in this study and the techniques used to establish these are outlined
below.
D

4.2.1 Comparison with Mineralogical (Petrographic) Data

TE

In some studies it is possible to compare mineralogical data (e.g. XRD, petrography etc.) with
the geochemical data acquired from the same samples to establish element:mineral links.
Unfortunately, a comprehensive comparison between the mineralogical and geochemical data was not
possible in the present study as the petrographic dataset was limited and samples were not taken from
EP

the same depths as the ones subjected to ICP analysis. In spite of this drawback, the recording of
biogenic phosphate in many of the thin sections is significant. The element P was used for
chemostratigraphic purposes in the present study, but this element may either occur in the form of
biogenic phosphate (e.g. fossil fish bones and scales) or in P-bearing heavy minerals such as apatite,
C

xenotime and monazite. The fact that no P-bearing heavy minerals were recorded in thin section, but
biogenic phosphate was, may be taken as evidence that P has the strongest mineralogical affinities with
AC

the latter.

4.2.2 Statistical Techniques (Principal Component Analysis)

Principal Component Analysis (PCA) is a statistical technique used to identify important

element associations, as elements occurring together in the same area of the eigenvector plots are likely
to have similar mineralogical affinities. The principal component (PC) scores for the original vectors
(eigenvectors) indicate the association of certain elements and the relationship between some elements
and mineralogy. This technique has been used for this purpose by Craigie (2015) and others. The
principal element associations predicted by the interpretation of PCA results are discussed in the
following paragraphs. In order to avoid any associations with grain size/lithology, PCA is normally
performed separately on sandstone and mudrock datasets. However, owing to the low number of
analyzed mudrocks, the technique has only been applied to the sandstone dataset in the present study.
EV1 (eigenvector 1) and EV2 (eigenvector 2) represent 46.13% and 16.18% of the variance in the

Saudi Aramco: Company General Use 6

 ACCEPTED MANUSCRIPT
geochemical dataset, respectively. When the EV1 and EV2 values are plotted, the elements form
several broad associations (Fig. 7), as follows:

Group 1: includes Si. This element is mainly associated with quartz.

Group 2: includes Ca and Mn. These elements are normally concentrated in carbonate minerals,
particularly calcite.

Group 3: includes MREE, HREE, LREE, Ta, Y, U, Nb, Ti and Th. These are generally considered to
be ‘high field strength’ elements linked with a variety of heavy minerals.

PT
Group 4: includes Mg, Fe, Cs, Be, V, Cr, Rb, Sc, Ga, Al and K. These elements are concentrated in
various minerals associated with the fine sand, silt and clay fractions.

Group 5: includes Zr and Hf. These elements have strong mineralogical affinities with zircon.

RI
Group 6: includes P. This element is concentrated in P bearing heavy minerals such as apatites and
monazites and/or in biogenic phosphate.

SC
Group 7: includes Cu, Ni and Co. These elements may be linked with pyrite and/or the development of
anoxia. A discussion on the precise mineralogical affinities of these elements is beyond the scope of
this paper but for more detailed information the reader is referred to the works of Craigie (1997 and
2015).

U
Group 8: includes Ba and Sr. These elements are associated with drilling additives.
AN
Figure 7 shows that the EV1 axis is most strongly influenced by variations in grain size as Si
(group 1), plots with low values of this parameter. In contrast, EV2 is affected by grain size, drilling
additives and carbonate content. Si plots with low EV2 values, as with Zr and Hf (concentrated in
M

zircons which may be associated with slightly coarser grained sandstones), while the group 2
(carbonate association) and 8 (drilling additives) elements yield elevated values of EV2.

4.2.3 Graphical Techniques (Binary Diagrams)

D

Although PCA may be used to identify broad element:mineral links, the technique should be
TE

used with caution, as the presence of one or two samples with anomalously high or low values of a
particular element may exert an ‘overriding’ influence on eigenvector values. Furthermore, the fact that
elements plot in the same field does not necessarily mean that they are linked to the same minerals. The
conclusions drawn on the mineralogical affinities of specific elements, from this statistical technique,
EP

should be supported by either comparing the geochemical dataset with available mineralogical data
(e.g. petrography, XRD, heavy mineral data) or by employing binary diagrams. The binary diagrams
presented in Figure 8 are used to establish certain element:mineral links.
The Al vs. Si binary diagram (Figure 8a) shows a negative association between the two
C

elements, explained by the fact that Si and Al are mainly concentrated in quartz and clay minerals,
respectively. It is noted that a minority of samples plot with low values of both elements. These are
likely to be enriched in carbonate minerals (probably calcite) as shown by their elevated concentrations
AC

of Ca (Figure 8b). Distinct positive trends are observed in the graphs of Al vs. K (Figure 8c) and Al vs.
Ga (Figure 8d). It is probable that Ga and K (along with Al) are mainly associated with clay minerals,
but may also be concentrated in K feldspars and micas to a lesser degree.
Figures 8e and 8f display very strong relationships between Ti, Nb and Ta, suggesting that
these elements are concentrated in the same heavy minerals, probably rutile, anatase, sphene,
titanomagnetite, magnetite and/or ilmenite. The fact that there is a reasonably strong association
between Ta and Th (Figure 8g) may suggest that Th is also associated with these heavy minerals.
Figure 8h shows a very strong positive relationship between Zr and Hf, confirming that both elements
are linked with zircon.

4.3 Key elements and ratios

Saudi Aramco: Company General Use 7

 ACCEPTED MANUSCRIPT
The chemostratigraphic scheme proposed for the present study involved using the
geochemical data to model subtle changes in provenance and then to use that as a basis for reservoir
correlation. Although data may be obtained for more than 50 elements in any given chemostratigraphy
study, most schemes are based on variations in 4-12 ‘key’ or ‘index’ elements. In order to avoid
mineralogical changes related entirely to variations in paleoenvironment/grain size, key element ratios,
as opposed to individual elements, were employed for chemostratigraphic analysis. When choosing the
key elements and ratios, care was also applied to avoid using elements that are primarily sourced from
minerals susceptible to post-depositional weathering/diagenesis, such as Na and K (both abundant in
feldspars). The list of key element ratios and their mineralogical significance are as follows:

Zr/Y: zircon vs. Y-bearing heavy minerals

PT
Th/P: Th-bearing heavy minerals vs. P-bearing heavy minerals (e.g. apatite, monazite) and/or biogenic
phosphate
Ta/P: Ta-bearing heavy minerals (i.e. rutile, anatase, sphene, titanomagnetite, magnetite, ilmenite) vs.
P-bearing heavy minerals (e.g. apatite, monazite) and/or biogenic phosphate

RI
Cr/P: Cr-bearing heavy minerals (e.g. chrome spinel) vs. P-bearing heavy minerals (e.g. apatite,
monazite) and/or biogenic phosphate
P/Lu: P-bearing heavy minerals (e.g. apatite, monazite) and/or biogenic phosphate vs. Lu-bearing

SC
heavy minerals
Th/Ti: Th-bearing heavy minerals vs. Ti-bearing heavy minerals (i.e. rutile, anatase, sphene,
titanomagnetite, magnetite, ilmenite)
Cr/Yb: Cr-bearing heavy minerals (e.g. chrome spinel) vs. Yb-bearing heavy minerals
Th/Ta: Th-bearing heavy minerals vs. Ta-bearing heavy minerals (i.e. rutile, anatase, sphene,

U
titanomagnetite, magnetite, ilmenite)
Zr/Lu: zircon vs. Lu-bearing heavy minerals
AN
Th/Nb: Th-bearing heavy minerals vs. Nb-bearing heavy minerals (i.e. rutile, anatase, sphene,
titanomagnetite, magnetite, ilmenite)

Variations in most key elements and ratios are thought to be largely dependent on changes in
M

source/provenance as the elements Zr, Nb, Ti, Ta, Cr, Y, Lu and Yb are highly immobile, being almost
exclusively concentrated in heavy minerals. With the exception of Zr (concentrated in zircon), these
elements may have multiple mineralogical affinities (e.g. Ti may be concentrated in rutile, anatase,
sphene, titanomagnetite, magnetite or ilmenite), but they are nearly always associated with heavy
D

minerals. Further information on the mineralogical affinities of elements associated with heavy
minerals and their application to chemostratigraphy are discussed by Mange and Morton (2007), Racey
TE

et al., (1995), Pearce et al., (1999), Pearce et al., (2005), Ratcliffe et al., (2010), Craigie (2015) and
others. In spite of the fact that some heavy minerals, such as zircon and monazite, are more resistant to
physical and chemical attack than others (e.g. apatite), variations in Th/Ti, Cr/Yb, Th/Ta, Zr/Lu and
Th/Nb are considered to represent changes in provenance. If hydraulic sorting or variations in
depositional conditions had an effect on these ratios, a close negative or positive association would
EP

have been expected with trends plotted for the Si/Al ratio (representing quartz vs. clay minerals and
primarily dependent on changes in depositional environment). Close observations have revealed no
significant similarities in the trends shown in profiles plotted for these ratios and Si/Al. The possibility
of post depositional diagenetic alteration of heavy minerals cannot be completely ruled out but is
C

unlikely as heavy minerals are normally highly resistant to weathering/diagenesis under ‘standard’
aqueous conditions. Furthermore, the petrographic study revealed no observable diagenetic alteration
AC

of heavy minerals.
The element P may be associated with either heavy minerals (e.g. apatite, monazite) and/or
biogenic phosphate (Craigie, 1998). However, the somewhat limited petrographic dataset indicates that
P is more likely to reflect the distribution of the latter. In spite of the fact that P may be mostly, if not
exclusively, concentrated in biogenic phosphate, it is very unlikely that this would be ‘in situ’ as
turbidite sediments would have been deposited rapidly under high energy conditions. This material is
much more likely to have been ‘sourced’ from hinterland areas. Consequently, variations in P and
ratios involving this element (e.g. Ta/P) are considered to reflect changes in source/provenance, rather
than depositional environment. It is possible that biogenic phosphate may have been sourced from
erosion of the underlying Qasim Formation or from shallow marinesediments occurring in hinterland
areas during the deposition of the Sarah Formation. In both instances, there may have been a
pal;eoenvironmental control on the original distribution and occurrence of biogenic phosphate. Given
that the biogenic phosphate occurring within the Sarah Formation in the study wells is unlikely to have

Saudi Aramco: Company General Use 8

 ACCEPTED MANUSCRIPT
accumulated in situ, however, the distribution of P is considered to have been influenced by changes in
provenance.
Details of the chemostratigraphic schemes proposed for the sandstone and mudrock datasets
are shown in Figures 9 - 16. It is noteworthy that the Si/Al profile is plotted in Figures 9, 11 and 14.
This may be considered a ‘quality control check’ to provide an indication of grain size. As Si and Al
are associated with quartz and clay minerals respectively, this ratio is known to increase with grain
size. Any key element ratio profile that shows a very similar trend to that of Si/Al should be used with
caution, as it may be dependent on changes in grain size and depositional environment, rather than
source/provenance.

4.4 Chemostratigraphic characterization

PT
The chemostratigraphic framework presented herein comprises a hierarchical order of zones,
subzones and divisions based on changes in key element ratios. Each zone, subzone and division is

RI
defined using specific values of one or more key element ratio. For example, zone JL1 is characterized
by Zr/Y ratios lower than 20 and Th/P values of less than 35. These changes may be subtle but reflect
significant changes in the proportion of particular heavy minerals (and detrital biogenic phosphate),
hence, variations in provenance. In order to avoid variations in geochemistry relating to changes in

SC
grain size/lithology, separate schemes have been proposed for the sandstone and mudrock datasets, the
details of which are outlined in the following paragraphs.
Details of the chemostratigraphic scheme proposed for sandstone lithologies are shown in the
zonation/correlation panels of Figures 9 - 11, while the principal geochemical characteristics of each

U
zone, subzone and division are summarized in Figure 12. The mudrock scheme is illustrated in Figures
15 and 16. Chemozones are recognized in each study well; but, for brevity, key element ratios are only
included for a small selection of wells in this paper in order to illustrate their geochemical
AN
characteristics.

4.4.1 Chemostratigraphic Zones- sandstones

M

Of the two chemostratigraphic zones, JL1 occurs at the base and generally produces lower
values of Zr/Y and Th/P than in the overlying zone JL2. However, it is noted that the JL1:JL2
boundary is placed with a higher degree of certainty using the Th/P ratio than Zr/Y. Figure 9 shows that
D

it is possible to differentiate the two zones using Th/P in Well 3, but this is more difficult using Zr/Y,
as many of the JL2 samples plot with low values of this ratio. Similarly, Zr/Y ratios are highly variable
in Well 8, where it is only possible to determine the absence of JL1 by the fact that Th/P values
TE

consistently exceed 35 throughout the study interval of this well.

4.4.2 Chemostratigraphic Subzones- sandstones

EP

Chemostratigraphic subzones are only identified in zone JL2 and are labelled JL2-1, JL2-2
and JL2-3 in ascending stratigraphic order. Subzones JL2-1 and JL2-3 are very similar geochemically,
yielding higher values of Ta/P and Cr/P than in JL2-2 (Figures 9 - 11). Note that the ‘cut off’ values
used to define these subzones are higher for both ratios in Well 2. This inconsistency may be explained
C

by local variations in source/provenance that only persisted in the immediate vicinity of this well. It is
also possible to employ binary diagrams to differentiate the chemostratigraphic subzones. Figure 13
AC

illustrates that, though there is some ‘overlap’ of fields, the samples assigned to subzone JL2-2
generally yield lower values of Cr/P and Ta/P than those of JL2-3.

4.4.3 Chemostratigraphic Divisions- sandstones

Chemostratigraphic divisions are recognized in all three zone JL2 subzones (Figures 9-12).
The divisions of subzone JL2-1 are labelled JL2-1a, JL2-1b and JL2-1c in ascending stratigraphic
order, with the former characterised by the highest P/Lu ratio values. Division JL2-1b yields lower
P/Lu and Th/Ti values than in JL2-1a and Jl2-1c respectively, while JL2-1c is defined by the highest
Th/Ti ratios. A twofold subdivision is proposed for JL2-2, with JL2-2a occurring at the base of this
subzone and producing higher Cr/Yb and Zr/Y values than in the overlying division JL2-2b. The fact
that a lower Zr/Y ‘cut off’ value has been used to place the JL2-2 division boundaries in Well 9 may be
explained by local variations in provenance in the vicinity of this well location. Of the two subzone
JL2-3 divisions, JL2-3a is recognised at the base and is characterised by lower Th/Ta values than in the
overlying division JL2-3b.
Saudi Aramco: Company General Use 9
 ACCEPTED MANUSCRIPT
The differentiation of division JL2-1b from JL2-1c are shown in Wells 3, 4 and 8 (Figures 9 -
13), though it is notable that JL2-1a is only identified in Well 6. Figures 9 and 10 illustrate that Zr/Y
values are generally higher in division JL2-2a than in JL2-2b in Wells 8 and 4, respectively. In some
wells, Cr/Yb ratios are higher in JL2-2a, though this parameter is considered less reliable as this does
not hold true from some wells, including Wells 4 and 8. Division JL2-2a is considered to be absent in
Well 3 (Figure 9) as low Zr/Y values persist throughout subzone JL2-2. Figure 9 illustrates that JL2-3a
produces lower Th/Ta values than in the overlying subzone JL2-3b in Well 8, while Figure 10 shows
that the same holds true for Wells 4 and 10. Values of Th/Ta are consistently high throughout JL2-3 in
well 3, suggesting that these samples should be assigned to JL2-3b, with JL2-3a being either missing or
unsampled.
Division JL2-1b’ is only identified in Wells 6 and 9, where it is considered to represent a local

PT
change in source/provenance that probably only occurred in the immediate vicinity of these wells.
Figures 11 and 16 illustrate that this chemozone is very similar to JL2-1b, being characterised by
slightly higher Zr/Y ratios.

RI
4.4.4 Definition of “A” Sand

The 5 - 10 ft thick Uqlah Formation is recognised beneath the Qusaiba Member shales in most

SC
wells, though it has not been sampled in every well (e.g. Wells 3, 4, 8 and 10). Where it has been
sampled, it yielded very similar geochemical signatures to the sands occurring immediately below it.
For example, in Well 6, it cannot be distinguished from JL2-3 based on values of key elements alone
(Fig. 11). In most wells this sand is geochemically similar to subzone JL2-3b, hence the labelled AJL2-
3b. The exception is in Well 9, where subzone JL2-3 is absent and the uppermost sandstone overlies

U
JL2-2. In this well, the sand bed is geochemically similar to the underlying sandstones of JL2-2b
(though there is a sample gap between JL2-2a and AJL2-2b, and JL2-2b has not been sampled in this
AN
well) and is also labelled AJL2-2b (Fig. 14). The only geochemical parameter that can be used to
distinguish the uppermost sandstones of AJL2-2b and AJL2-3b from JL2-2b and JL2-3b is Si/Al.
Values of this ratio are very much higher in this AJL2-2b and AJL2-3b, reflecting the fact that these
chemozones are more mineralogically mature than most other sandstones of the underlying Sarah
M

Formation (Figs 9 - 11 and 14). The fact that chemozones AJL2-2b and AJL2-3b produce similar
values of key element ratios to those of JL2-2b and JL2-3b respectively, suggests that they were
derived from the erosion/reworking of turbidite fan deposits of the uppermost Sarah Formation.
Sedimentological data indicate that the A sand was deposited in a wave and tide influenced shoreface
D

environment. Deposition in such a high energy shelf regime provides an explanation for the
mineralogically mature character of these sandstones and their elevated values of Si/Al.
TE

4.4.5 Chemostratigraphic Zones- mudrocks

The scheme proposed for the mudrock dataset is less detailed, owing to the relatively low
EP

number of samples taken from this lithology. Nevertheless, it is possible to recognise a hierarchical
order of three zones and five subzones. Details of the scheme are shown in Figure 15 (correlation
panel), with the principal geochemical characteristics of each chemozone presented in Figure 16.
A broad threefold zonation is proposed, with zones JLM1 and JLM3 occurring at the base and
top of the study sections respectively, and defined by lower Th/Nb values than in the intervening zone
C

JLM2. It is also noted that Zr/Lu values are generally higher in zone JLM1 than in JLM2 (although this
is less obvious in Well 1). Figure 15 illustrates that the basal samples of Wells 1 and 8 yield low Zr/Lu
AC

and high Th/Nb ratios - typical of JLM2, with JLM1 either missing or occurring in an unsampled part
of each well below JLM2. In fact, zone JLM-1 has only been sampled in Well 2 (profiles of this well
are not included in this paper).

4.4.6 Chemostratigraphic Subzones- mudrocks

Chemostratigraphic subzones are only recognized in zone JLM3, where they are labelled
JLM3-1, JLM3-2, JLM3-3, JLM3-4 and JLM3-5 in ascending stratigraphic order. Subzones JLM3-1,
JLM3-3 and JLM3-5 yield higher Ta/P ratios than in JLM3-2 and JLM3-4 (Figs. 15 and 16). Perhaps
owing to the relatively low number of analysed mudrocks, it is not possible to identify
chemostratigraphic divisions in any of the aforementioned subzones.

4.5 Chemostratigraphic correlation

Saudi Aramco: Company General Use 10

 ACCEPTED MANUSCRIPT
Separate chemostratigraphic schemes have been produced for the sandstone and mudrock
datasets; but, in order to produce a correlation to honor both, a combined scheme has been produced
and summarised in Figure 17. This diagram shows that definition of the two basal chemozones is based
entirely on the mudrock dataset. These were labelled JLM-1 and JLM-2 in the original scheme
proposed for mudrock lithologies, but have been renamed J1 and J2 respectively in the final/combined
framework. Similarly, chemozone JL1 represents sandstones occurring at the base of the Sarah
Formation that is equivalent to chemozone J3 in the finalised scheme. Figure 18 demonstrates the
correlation between the study wells based on this finalized/combined scheme.
The principal benefit of employing chemostratigraphy is that it was possible to construct high
resolution correlations of individual sandstone beds (compare the Fig. 18 with a correlation prior to
chemostratigraphic analysis presented in Fig. 6). Some chemozones, such as J4-2b and J4-3b are

PT
correlative between almost every well. In contrast, others are more restricted in their lateral extent, only
occurring in four wells or less. For example, division J4-1b’ is only identified in Wells 6 and 9, while
J4-1c is only present in Wells 3, 6 and 8. In spite of a number of sample gaps, the chemostratigraphic
scheme is considered robust and many of the questions relating to variations on the thickness of the

RI
Sarah Formation can now be explained. Previously, the reasons for the relatively thin expressions of
the turbidite fan sediments of the upper Sarah Formation were unclear in Wells 1, 5 and 10. This study
revealed the uppermost chemozones to be present, but the lower ones (i.e. subzone J4-1 and associated

SC
divisions) are clearly missing (Fig. 18). In contrast, the basal chemozones of J4-1 and J4-2 are clearly
identified in Well 9, with J4-3 absent, probably as a result of erosion/non-deposition. With the
exception of Well 9, chemozones J4-1b, J4-1b’, J4-1c and J4-2a are generally only identified where the
turbidite fans at the top of the Sarah Formation attain their greatest thickness (e.g. in Wells 2, 6, 8 and
3). Where the turbidite fan deposits are relatively thin (e.g. Wells 1, 5 and 10), this is usually explained

U
by the absence of these basal chemozones.
AN
5. Geological Implications

Prior to this study, hydrocarbon exploration of Sarah Formation outwash fan sediments has
relied upon simplistic geological models integrating conventional sedimentological datasets acquired
M

from well bores and various seismic amplitude extraction techniques. The geological model used for
static and dynamic reservoir simulation suffered poor resolution due to inherent uncertainties regarding
reservoir layering in this rapidly deposited, arenaceous succession. At the formation level, chemical
fingerprinting of the Late Ordovician succession enabled consistent identification and correlation of the
D

pre-glacial Qasim (zones J1 and J2), syn-glacial Sarah (J3 and J4) and post-glacial Qalibah (J5)
formations (Fig. 18). As discussed in section 4.5, the “A sand” has an inherited signature from the
TE

chemical zone below and represents initial reworking of the post glacial landscape prior to deposition
of J5 hot shales.
Within the Sarah succession, the chemical fingerprint is typically provenance controlled, with
secondary climatic / diagenetic variables. As a result of the short time span of deposition, lack of major
climatic changes, multiple switches in provenance and short / rapid transfer between source and sink,
EP

chemostratigraphy was found to be particularly useful in producing a high resolution scheme for the
Sarah reservoir. The reservoir layering scheme based on chemostratigraphy (Fig. 18) highlights a
complex depositional system characterized by an initial phase of deeply entrenched pro-glacial
turbidite channels (zone J3) and younger sheet-like sandstones (zone J4), that occur above the J3
C

channels and in apparent inter-channel areas. The broadly layer-cake correlation of zone J4 across the
study area reflects a younger fan system that was not linked to the J3 feeder systems. The fans
AC

associated with this early phase of channelization may occur to the east of the studied seismic
acquisition area (Fig. 1).
The thin basal channel sand in Well 6 (Fig. 5a) has the same chemical signature (J4-1) as
some of the earliest deposited lobes (e.g. Well 9), suggesting shared system links tied to a common
provenance. A strict chemical correlation between the two may lead to confusion between
distinguishing confined feeder elements and attached fans, leading to errors populating geocellular
models. The true nature of the channel lobe system can only be understood when the chemical
stratigraphy is fully integrated with seismic attribute maps and borehole image palaeoflow data. An
example of this is illustrated in Figure 19 and covers the stratigraphic expression seen in Wells 4 and 5,
with an additional Well 17 (not sampled for chemostratigraphic analysis). Two discrete 3D seismic
stratal slices have been generated (deep and shallow) and their positions marked on a Northwest-
Southeast oriented 2D section. The deep slice highlights a broadly Northeast-Southwest trending
channel (labelled as (a) in Fig. 19) that is penetrated by Well 4. Palaeoflow direction in the channelized
portion (Fig. 6) is orthogonal to the long axis, which is consistent with the across channel sidebar mode

Saudi Aramco: Company General Use 11

 ACCEPTED MANUSCRIPT
of fill. There is no evidence along the channel length that lobes or splays in inter-channel areas are
being deposited. We may expect to find terminal lobes towards the NE beyond the seismic volume. On
the 2D section, the channel penetrated by Well 4 is clearly visible incising into the pre-glacial Qasim
Formation below (chemozones J1 and J2).
The shallow stratal slice imaging the younger evolution of the fan system (J4 chemozone)
captures the development of a terminal fan or lobe (labelled as (b) in Fig. 19) prograding to the north
(based on palaeoflow). This can be seen on the 2D section as a bright red sheet-like attribute that is also
penetrated by Well 17. The feeder system to the Southeast is a continuation of the channel penetrated
in Well 6 (see Fig. 1); thus a strong systems link can be proven between the J4-1 bearing channel fill
and its terminal fan. Well 5 at the periphery of the lobe has only the younger J4-2 and J4-3
chemozones, which suggests the lobe prograded into this area over time. Further complexity across the

PT
fan system can be seen by a third phase of channelization cross cuttings the previously deposited lobe,
with a prominent feature occurring between Wells 4 and 17 (labelled as (c) in Fig. 19).
Turbidite fan lobes form the most prospective reservoir intervals both in terms of
porosity/permeability and gas charging. Understanding the potential thickness variations in this unit is

RI
critical for optimum well placement, a task made clearer after applying a chemical zonation scheme.
Thickening of high quality lobe sands typically occur above the deeply incised zone J3 channels (Fig.
18), suggesting additional accommodation space was created from under-filled ribbon channels.

SC
Targeting deeper channels in locations with proven lobes offers the best chance of success, as lobe
thickness is often expanded, while J3 channel thalwegs offer secondary reservoir targets.
The high resolution correlation afforded by chemical stratigraphy also highlights the role of
isostatic rebound on the preservation of the uppermost part of the Sarah Formation. The regional
deformed zone interpreted as a product of glacioisostasy occurs towards the top of chemozone J4-3b

U
(Fig. 18). Although well-developed across much of the study area, there are a few locations (e.g. Well
7) where this zone is truncated beneath the PGRU at the base of the Uqlah Formation (A-sand). Both of
AN
these wells were drilled in close proximity to Northwest-Southeast trending basement lineaments with
an activation history dating back to at least the Cambrian. It is proposed that parts of chemozone J4-3b
were lost across reactivated fault block highs by increased erosion on the shallow marine post glacial
shelf.
M

6. Conclusions

This contribution to the stratigraphic layering of glaciogenic hydrocarbon bearing reservoirs in

D

Northwest Saudi Arabia embraces a true multidisciplinary approach, where geochemical, 3D seismic,
borehole image and traditional sedimentological techniques aid high resolution zonation and modelling
TE

of this complex deposition system. The following key conclusions have been noted based on the study
of ten wells penetrating Hirnantian age pro-glacial outwash fan sediments:

• Detailed geochemical sampling of core and cuttings material and the generation of different
EP

chemical zonation schemes for arenaceous and argillaceous rock types provided a sound
platform for high resolution chemostratigraphy that was not biased by grain size. The inherent
nature of the pro-glacial system, with a short transfer time between source and sink, multiple
small scale changes in provenance and coarse biostratigraphic zonation make the succession
C

particular conducive to chemical fingerprinting.

• The ‘key’ elemental ratios used to define chemostratigraphic boundaries include Zr/Nb,
AC

Nb/Cr, Zr/Y, Th/P, Ta/P, Cr/P, P/Lu, Th/Ti, Cr/Yb, Zr/Lu and Th/Nb. Variations in these
ratios are largely dependent on changes in source/provenance as Zr, Nb, Cr, Y, Th, Yb and Lu
are high field strength elements that are very stable in ‘standard’ aqueous conditions, and are
largely unaffected by post depositional weathering/diagenesis. The element P is are considered
to be influenced by changes in provenance. By employing ratios, as opposed to individual
elements, it is possible to reduce the effects of hydraulic sorting and depositional processes, so
that the resulting chemostratigraphic boundaries reflect variations in source/provenance. Most
of the aforementioned elements are concentrated in heavy minerals.

• At the formation level, chemical fingerprinting adequately constrained the pre-glacial, syn-
glacial and post-glacial stratigraphy. As well as using the technique to identify major
formation boundaries, it was possible to propose a high resolution scheme for the Sarah
Formation, comprising a hierarchical order of zones, subzones and divisions. The

Saudi Aramco: Company General Use 12

 ACCEPTED MANUSCRIPT
identification of missing zones high in the syn-glacial Sarah Formation provided evidence for
early post-glacial erosion during isostatic rebound. Proximity to deep-seated NW-SE
lineaments suggest that differential isostasy may have been accommodated along faults. The
inherited chemical signature of early post glacial tidalites (Uqlah Formation) represents initial
reworking of the structurated post glacial landscape, prior to deposition of Silurian organic-
rich “hot” shales.

• In depth understanding of stratal layering and provenance from the chemostratigraphy study
allowed for reconstruction of the syn-glacial outwash fan system in both space and time. The
oldest chemozones identified in the Sarah Formation occur within elongate, deeply entrenched
ribbon channels (zone J3). This provenance-controlled signature had no expression in wells

PT
drilled between channels, suggesting unconfined lobate portions of the fan system may be
found downdip of the ribbon element. The relationship between lobate facies above ribbon
channels and in inter-channel areas can only be resolved when integrated with 3D seismic
geometries from multiple time slices. Deep time slices highlight large scale NE-SW channels

RI
(chemozone J3) incised into the pre-glacial Qasim Formation (chemozone J2), with very little
sedimentation outside of channel margins. Zone J4 sheet-like lobe facies found in non-
channelized wells are geochemically tied to a second phase of small channels, that often cross

SC
cut the deeper lying ribbons. Thickening of gas-bearing high quality J4 lobe sandstones can be
demonstrated in under-filled J3 ribbons.

Acknowledgements

U
The authors thank Saudi Aramco for granting permission to publish this work and for providing
stratigraphic columns and maps. Words of thanks also go to Chemostrat Ltd for generating ICP-OES
AN
and ICP-MS data used in this study, Daniel Franken for providing seismic basemaps and John Melvin
for lengthy discussions on all things glacial.

References
M

Beuf, S., Biju-Duval, B., de Charpal, O., Rognon, P., Gariel, O., Bennacef, A., 1971. Les Grès du
Palaeozoïque inférieur au Sahara. Editions Technip, Paris.
Clark-Lowes, D.D., 1980. Sedimentology and mineralization potential of Saq and Tabuk Formations.
D

Imperial College of Science and Technology Cover Rock Contract Openfile Report CRC/IC
7Open-file Report DGMR 767. Saudi Arabian Deputy Ministry for Mineral Resources, Jiddah 88.
TE

Craigie, N.W. 1998. Chemostratigraphy of Middle Devonian lacustrine sediments, N.E. Scotland.
Unpublished PhD thesis, University of Aberdeen.
Craigie, N.W. 2015. Applications of chemostratigraphy in Middle Jurassic unconventional reservoir in
eastern Saudi Arabia. GeoArabia 20, 79-110.
EP

Deschamps, R, Eschard, R., Rousse, S. 2013. Architecture of Late Ordovician glacial valleys in the
Tassili N’Ajjer area (Algeria). Sedimentary Geology, 289, 124-147.
Dowdeswell, J.A. Cofaigh, C. Ó., Noormets, R. Larter, R.D., Hillenbrand, Claus-Dieter, Benetti, S.
Evans, S, J., Pudsey, C.J. 2008 A major trough-mouth fan on the continental margin of the
C

Bellingshausen Sea, West Antarctica: The Belgica Fan. Marine Geology, 252 (3-4), pp. 129-140.
Eyles N. 1993 Earth’s glacial record and it’s tectonic setting. Earth Science Review, volume 35, pp1-
248
AC

Ghienne, J.F., 2003. Late Ordovician sedimentary environments, glacial cycles, and post-glacial
transgression in the Taoudeni Basin, West Africa. Palaeogeography, Palaeoclimatology,
Palaeoecology 189, 117–145.
Ghienne, J.F., Le Heron, D., Moreau, J., Denis, M., Deynoux, M., 2007. The Late Ordovician glacial
sedimentary system of the North Gondwana platform. In: Hambrey, M., Christoffersen, P., Glasser,
N., Janssen, P., Hubbard, B., Siegert, M. (Eds.), Glacial Sedimentary Processes and Products.
Special Publication, 39. International Association of Sedimentologists, pp. 295–319.
Ghienne, J.F., Girard, F., Moreau, J., Rubino, J.L., 2010. Late Ordovician climbing-dune cross-
stratification: a signature of Outburst floods in proglacial outwash environments? Sedimentology,
57, 1175–1198.
Hirst, J. P. P., Benbakir, A., Payne, D. F., Westlake, I. R. 2002. Tunnel Valleys and Density Flow
Processes in the upper Ordovician glacial succession, Illizi Basin, Algeria: influence on reservoir
quality. Journal of Petroleum Geology, 25(3), 297-324.

Saudi Aramco: Company General Use 13

 ACCEPTED MANUSCRIPT
Janjou, D., Halawani, M.A., Al-Muallem, M.S., Robelin, C., Brosse, J-M, Courbouleix, S., Dagain, J.,
Genna, A., Razin, P., Roobol, M. J., Shorbaji, H., Wyns, R. 1996. Geologic map of the Al Qalibah
quadrangle, Sheet 28C, Kingdom of Saudi Arabia (with text). Saudi Arabian Deputy Ministry for
Mineral Resources Geoscience Map GM-135, 1:250,000.
Janjou, D., Halawani, M.A., Brosse, J-M., Al-Muallem, M.S., Becq-Giraudon, J.F., Dagain, J., Genna,
A., Razin, P., Roobol, M.J., Shorbaji H., Wyns, R. 1997. Geologic map of the Tabuk quadrangle,
Sheet 28B, Kingdom of Saudi Arabia (with text). Saudi Arabian Deputy Ministry for Mineral
Resources Map GM-137, 1:250,000.
Lang, J., Dixon, R. J., Le Heron, D. P., Winsemann, J. 2012. Depositional architecture and sequence
stratigraphic correlation of Upper Ordovician glaciogenic deposits, Illizi Basin, Algeria. Geological
Society, London, Special Publications, 368.

PT
Le Heron, D., Sutcliffe, O., Bourgig, K., Craig, J., Visentin, C., Whittington, R. 2004. Sedimentary
architecture of Upper Ordovician tunnel valleys, Gargaf Arch, Libya: Implications for the genesis
of a hydrocarbon reservoir. GeoArabia 9, pp. 137-160.
Le Heron, D.P., Craig, J., Sutcliffe, O.E., Whittington, R. 2006. Glaciogenic Reservoir Heterogeneity:

RI
an example from the Late Ordovician of the Murzuq Basin, SW Libya. Marine and Petroleum
Geology 23, pp. 655-677.
Mange, M.A., Morton, A.C. 2007. Geochemistry of heavy minerals. Developments in sedimentology,

SC
58, pp. 345-391.
McGilvray G.J., Hussaini, M. 1992. The Paleozoic petroleum geology of Central Saudi Arabia. Am.
Assoc. Pet Geol. Bull., 76, pp1473-1490.
Melvin, J. 2009. Heterogeneity in glaciogenic reservoirs: examples from the Ordovician and Permo-
Carboniferous of Saudi-Arabia. In Glaciogenic Reservoirs and Hydrocarbon Systems, Conference

U
Abstract Volume, Geological Society of London, pp. 37-38.
Melvin, J., Miller, M. 2009. Re-advance, retreat and rebound: stratigraphic response of the Sarah
AN
Formation (Hawban and Baq’a members) to the demise of the Late Ordovician (Hirnantian)
glaciation in central Saudi Arabia. In Glaciogenic Reservoirs and Hydrocarbon Systems,
Conference Abstract Volume, Geological Society of London, pp. 62-63.
Melvin, J. 2015. Lithostratigraphy and depositional history of Upper Ordovician and lowermost
M

Silurian sediments recovered from the Qusaiba-1 shallow core hole, Qasim region, central Saudi
Arabia. Review of Palaeobotany and Palynology, 212, pp. 3-21.
O´Cofaigh, C., Taylor, J., Dowdeswell, J. A., Pudsey, C. J. 2003 (March): Palaeo-ice streams, trough
mouth fans and high-latitude continental slope sedimentation. Boreas, 32, pp. 37–55.
D

Pearce, T.J., Besley, B.M., Wray, D.S., Wright, D.K. 1999. Chemostratigraphy: a method to improve
interwell correlation in barren sequences- a case study using onshore Duckmantian/Stephanian
TE

sequences (West Midlands, U.K.). Sedimentary Geology, 124, pp. 197-220.

Pearce, T.J., Wray, D.S., Ratcliffe, K.T., Wright, D.K., Moscariello, A. 2005. Chemostratigraphy of the
Upper Carboniferous Schooner formation, southern North Sea. In: Carboniferous hydrocarbon
geology: the southern North Sea and surrounding onshore areas. In: Colinson J.D., Evans, D.J.,
Holiday, D.W. and Jones, N.S. (eds) Yorkshire geological Society, Occasional Publication series, 7,
EP

pp. 147-164.
Racey, A., Love, M.A., Bobolecki, R.M., Walsh, J.N. 1995. The use of chemical element analysis in
the study of biostratigraphically barren sequences: an example from the Triassic of the central
North Sea (UKCS). In: Dunnay, R.E., Hailwood, E.A. (eds) Non-biostratigraphic methods of dating
C

and correlation. Geological Society, London Special Publication 89, pp. 69-105.
Ratcliffe, K.T., Wright, A.M., Montgomery, P., Palfrey, A., Vonk, A., Vermeulen, J., Barrett, M. 2010.
AC

Application of chemostratigraphy to the Mungaroo Formation, the Gorgon Field, offshore

northwest Australia. APPEA Journal 2010, 50th Anniversary issue pp. 371-388.
Rees A. J. 2015. The Late Ordivician glacial record in the subsurface of NW Saudi Arabia. Clastic
Reservoirs of the Middle East – American Association of Petroleum Geologists, 23rd – 25th March,
Kuwait.
Reineck, H.-E., Wunderlich, F. 1968. Classification and origin of flaser and lenticular bedding.
Sedimentology, 11, pp. 99–104.
Senalp, M. and Al-Duaiji, A.A. 2001. Qasim Formation: Ordovician storm- and tide-dominated
shallow-marine siliciclastic sequences, Central Saudi Arabia. GeoArabia, 6, 233 – 268.
Sharland, P.R., Archer, R., Casey, D.M., Davies, R.B., Hall, S.H., Heward, A.P., Horbury, A.D.,
Simmons, M.D. 2001. Arabian Plate Sequence Stratigraphy, GeoArabia Special publication 2.
Sutcliffe O.E., Dowdeswell, J.A., Whittington, R.L., Theron, J.N., Craig J. 2000. Calibrating the Late
Ordovician glaciation and mass extinction by the eccentricity cycles of Earth’s orbit. Geology 28,
pp. 967-970.

Saudi Aramco: Company General Use 14

 ACCEPTED MANUSCRIPT
Vaslet D. 1989. Late Ordovician glacial deposits in Saudi Arabia: a lithostratigraphic revision of the
Early Palaeozoic succession. Saudi Arabian Deputy Ministry for Mineral Resources, professional
papers, pp. 13-44.
Vaslet D. 1990. Upper Ordovician glacial deposits in Saudi Arabia. Episodes, 13, pp147-161.
Williams P.L, Vaslet, D., Jahnson, P.R., Berthiaux, A., LeStrat, P., Fourniguet, J. 1986. Geological
map of the Jabal Habashi quadrangle, Sheet 26F, Kingdom of Saudi Arabia (with text). Saudi
Arabian Deputy Ministry for Mineral Resources Geoscience Map GM-98 C, 1:250,000.

PT
RI
U SC
AN
M
D
TE
C EP
AC

Saudi Aramco: Company General Use 15

 ACCEPTED MANUSCRIPT

Figure Captions

Fig. 1. Location maps. (a) Map of the Arabian Peninsula showing the location of the study area in NW
Saudi Arabia. (b) Two way time basemap for the NW region showing deeper basin development towards
the NE, with thinning in more proximal southwesterly locations. The thicker linear feature on the
eastern side of the map is the Al Jauf Graben. The inferred Hirnanian maximum grounding line of the ice
is marked in white. (c) Stratal slice through 3D seismic relative acoustic impedance volume illustrating

PT
well locations and ribbon channel geometries of the Sarah Formation. The strong impedance contrast is
caused by the incision of sandy Sarah turbidite fan channels into the muddy pre-glacial Qasim shelf.

Fig. 2 General stratigraphy of Lower Palaeozoic strata in NW Saudi Arabia

RI
Fig. 3 Upper Ordovician geological setting. (a) Composite stratigraphic column showing typical gamma
ray and density/neutron cross over profile against the core. The stepped cold climate delta

SC
parasequences of the Lower Quwarah Member evolved into open marine bioturbated shelf heterolithics
with sharp based lowstand shorefaces of the Upper Quwarah. This unit is erosively truncated in proximal
sub-glacial locations by tunnel valleys, with less severe erosion in distal outwash fan areas. The syn-

U
glacial Sarah succession generally fines upwards, representing one major glacial retreat cycle above the
locally preserved glacial advance tectonite of the Zarqa Megafacies. A regional soft sediment deformed
AN
zone at the top of the Sarah is variably truncated by an erosion surface at the base of post-glacial
tidalites of the Uqlah Formation. This regional instability is caused by glacial isostasy and led to an
isostatic rebound unconformity between the Sarah and Uqlah formations. The transgressive,
bioturbated shelf shallow sediments of the latter evolved into condensed organic-rich black shales of
M

the Silurian age Qusaiba Member of the Qalibah Formation. (b) Depositional block model of pro-glacial
outwash fans fed by sub-glacial erosional of tunnel valleys in more proximal (southerly) locations. Fans
were developed during ice maxima lowstand phases. This depositional system is the current focus of this
D

study. (c) Depositional block model for the transgressive fill of exposed tunnel valleys during glacial
retreat. The main zone of deposition shifts away from fans and into valleys as the ice recedes, resulting
TE

in backfilling, a retrogradational valley fill and relative condensation on the fan. (d) Depositional block
model illustrating restructuration of the margin during isostatic rebound. Differential unloading
reactivates NW-SE trending faults creating local horst and graben features that provided short lived
EP

clastic source areas and local depocenters.

Fig. 4 Palaeogeography and stratigraphic setting. (a) Cross section flattened on base Qusaiba illustrating
broad architectural elements of sub-glacial tunnel valleys and their downdip pro-glacial fan extension.
C

The line of section (shown in map b) begins (Area A) and ends (Area C) in locations dominated by large
sub-glacial tunnel valleys with glacial advance Zarqa Megafacies in intervalley areas. Area B is the focus
AC

of this study, and lacks large scale erosional features and sub-glacial shearing reflecting its pro-glacial
setting. (b) Palaeogeographical map of the Hirnantian depositional system at close to glacial maximum.
The strong blue line running NW-SE marks the position of the maximum grounding line based on the
presence of sub-glacial tunnel valleys and glaciotectonic deformation. Pro-glacial outwash fans form a
NW-SE trending apron at the terminus of tunnel valleys that include the Risha gas field. (c) Relative sea
level curve outlining the relationship between eustatic sea level change and glacial depositional systems.
Ice advance (during falling sea level) has very little depositional expression in the region, with fans
formed during glacial maximum to early sea level rise (glacial retreat). This initial melting phase led to
tunnel valley incision and sediment transfer to pro-glacial fans. Increase in the rate of sea level rise

Saudi Aramco: Company General Use

 ACCEPTED MANUSCRIPT

coupled with ice sheet retreat exposed tunnel valleys with deposition shifting from the fans into valley
locations. (d) Depositional model illustrating the relationship between sub-glacial tunnel valleys,
grounding line and pro-glacial fans. This model runs approximately SW-NE across the (b)
palaeogeographic map. (e) Upper Ordovician stratigraphic column with asterisk denoting recovered
palynozones. Outwash fans are typically latest Katian to very early Hirnantian, while tunnel valley fills
are younger. This forms the basis for the sequence stratigraphic interpretation outlined in (c).

PT
Fig. 5 Lithological characteristics, depositional facies and image log-determined palaeoflow of the
Qasim, Sarah, Uqlah and Qusaiba in core and wireline logs. (A) 1:120 scale (ft) core description from
Well 6, shifted to conventional wireline logs highlighting lithology and facies associations. (B)

RI
Interpreted image log across a 6 ft interval towards the top of the Sarah Formation in Well 6 illustrating
climbing dune cross bedding. (C) Armored pebbly lag in Well 6 related to sediment bypass between
massive to cryptically bedded sandstones and wave influenced turbidites. Note the change in

SC
palaeoflow azimuth at this contact. (D) Climbing dune crossbedding developed in sandstones from Well
8. (E) Hyperpycnal turbidites from Well 3 showing a graded upper surface deposited during waning flow
followed by an inversely graded base to the next flow event. The distribution over 2 cm reflects waxing
steady state flows. (F) Regional sequence boundary (highlighted in black) between lower deformed

U
turbidites destabilized during isostatic rebound and upper post-glacial undeformed tidal sandstones in
AN
Well 6. This erosional contact is the distal expression of a regional post-glacial rebound unconformity
identified at the base of the Uqlah Formation in the outcrop belt to the southeast of the study area.

Fig. 6. (a) Lithostratigraphic wireline log correlation (in log depth) of selected study wells prior to
M

chemostratigraphic analysis. The facies described in Fig. 5 are illustrated along with major lithological
boundaries and image log-derived palaeoflow. (b) Sketch map of channel profiles imaged in 3D seismic
time slices. (c) Across channel sedimentological model (modified from Ghienne et al., 2010) illustrating
D

eddy-related sidebar growth across the channel in re-circulation zones. This model accounts for the
lateral fill (perpendicular to channel long-axis) demonstrated by oriented image log palaeoflow.
TE

Fig 7. Eigenvector (EV) crossplot used to aid the establishment of element:mineral links in sandstone
samples.
EP

Figure 8. Binary diagrams used to aid the establishment of element:mineral links.

Fig 9. Chemostratigraphic correlation proposed for Wells 3 and 8. Geochemical data acquired from
C

sandstone samples were used to construct this scheme. Definition of JL2-3a is made with little
confidence as only one sample is assigned to this division. All depths are log depths in feet.
AC

Figure 10. Chemostratigraphic correlation proposed for Wells 4 and 10. Geochemical data acquired from
sandstone samples were used to construct this scheme. All depths are log depths in feet. Note that
division JL2-2a is defined with a low level of confidence in Well 4.

Figure 11. Chemostratigraphic zonation proposed for Well 6. Geochemical data acquired from sandstone
samples were used to construct this scheme. All depths are log depths in feet.

Saudi Aramco: Company General Use

 ACCEPTED MANUSCRIPT

Figure 12. Summary of the principal geochemical criteria used to define each zone, subzone and division
with respect to the sandstone dataset. Note that details of “A sand” are not included in this figure but
are discussed in detail in the text.

Figure 13. Differentiation of subzones C2-2 and C2-3 (sandstone dataset) in Wells 2, 5, 6, 7 and 8 utilizing
selected binary diagrams.

PT
Figure 14. Chemostratigraphic zonation proposed for Well 9. Geochemical data acquired from sandstone
samples were used to construct this scheme. All depths are log depths in feet.

Figure 15. Chemostratigraphic correlation proposed for Wells 1, 7 and 8. Geochemical data acquired

RI
from mudrock samples were used to construct this scheme. All depths are log depths in feet.

Figure 16. Summary of the principal geochemical criteria used to define each zone and subzone with

SC
respect to the mudrock dataset.

Figure 17. Generalized comparison between the chemostratigraphic (mudrock and sandstone schemes)
and lithostratigraphic schemes.

U
Figure 18. Summary of the chemostratigraphic correlation (combined scheme for mudrock and
sandstone lithologies) proposed for Wells 1-10. All depths are log depths in feet. Depths of the cored
AN
intervals are indicated by the gray column next to the GR in each well. The location of the wells in
relation to turbidite channels in the lowermost Sarah Formation are also shown. Note that the “A sand”
represents the Uqlah Formation.
M

Fig. 19. 3D seismic relative acoustic impedance time slices and 2D cross section documenting the
D

geometric evolution of the pro-glacial fan system in the area penetrated by Wells 4, 5 and 17. The deep
time slice images a NE-SW oriented ribbon channel penetrated in Well 4 and characterized by the J3
TE

Chemozone. Sedimentation was constrained solely within the channel. The shallow time slice
(equivalent to Chemozone J4) highlights a second channel (connected to Well 6 to the SE) feeding a
terminal lobe. Both the channel signature in Well 6 and the lobe signature in Well 4 have the same
provenance character reflecting a strong systems link. A third channel oriented N-S can be seen cutting
EP

across the previously deposited lobe.

C
AC

Saudi Aramco: Company General Use

b ACCEPTED MANUSCRIPT a

PT
RI
SC
Thin

U
AN
Inferred maximum
Thick grounding line of ice
M

c 10 Channels
17 Faults
D

11
4 5
6
TE

2
3 13
9
EP

8
C

1
AC

12
7 14
15

Well
Volcanics
16

Figure 1
 O W E R
405 ACCEPTED MANUSCRIPT
JAUF STRATIGRAPHY

L SERIES
Subbat

SYST.
ERA
TIME STAGE Qasr LITHOLOGY
410 Pragian FORMATION MEMBER
Sha'iba WEST EAST

Lochkovian
415 TAWIL
PRIDOLI
Ludfordian

WENLOCK LUDLOW
420

S I L U R I A N
Gorstian
Homerian Sharawra
425
Sheinwoodian

430
LLANDOVERY

PT
Telychian Q ALIBAH Qusaiba
435
Aeronian
I C

440

RI
Rhuddanian
UQLAH
445 Hirnantian SARAH
P A L E O Z O

SC
Katian Quwarah
O R D O V I C I A N

450
UPPER

Ra'an
455 Q ASIM
Sandbian Kahfah

U
460
AN
Darriwilian Hanadir
MIDDLE

465
Dapingian
470
Floian
M

Sajir
LOWER

480 SAQ
Tremadocian

490
D
UPPER

Merioneth Risha
C A M B R I A N

500
QUWEIR A
TE

St. Davids
M

510 BUR J
SIQ
Lenan
520 Botomian
LOWER
EP

Atdabanian
530
Tommotian
Nemakit-
540 Daldynian
C

550 Ediacaran
AC

KURAYSHAH

600
650
700 Cryogenian
750
BASEMENT
Figure 2 800
TONIAN
 ACCEPTED MANUSCRIPT

Silurian
d. REGIONAL DEFORMED ZONE
UQLAH Fm
UPPER ORDOVICIAN (HIRNANTIAN)
SARAH FORMATION POST GLACIAL SHALLOW MARINE (ISOSTATIC REBOUND)
GLACIO-MARINE

PT
RI
“ZARQA”

SC
SHELF & LOWSTAND SHOREFACES

U
AN
M
QASIM FORMATION (QUWARAH MEMBER)
UPPER ORDOVICIAN (KATIAN)

D
TE
PRO-DELTA TO DELTA FRONT PARASEQUENCES

c. TRANSGRESSIVE TUNNEL VALLEY FILLS

C EP
AC

b. LOWSTAND PRO-GLACIAL OUTWASH FANS

(Focus of this study)
a. Figure 3
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC

Figure 4
 ACCEPTED MANUSCRIPT

Borehole Image Zone

5

PT
RI
U SC
AN

5
M
D
TE
EP
C
AC

Figure 5
(a) Well 5 Well 17
Well 2
Well 6
ACCEPTED MANUSCRIPT GR

Depth

Strat.
GR GR

Depth
Depth
Well 9

QUSAIBA MBR. Strat.

Strat.
GR Well 10

Depth

QUSAIBA MBR. Strat.

GR 0 150 0 150
0 150
Depth

UQLAH QUSAIBA Strat.

GR

Depth

UQLAH QUSAIBA Strat.

0 150
0 150 Well 4

QUSAIBA
QUSAIBA MBR.
0 150
GR

Depth

Strat.
0 150

UQLAH
SARAH FORMATION (turbidite fans) UQLAH

20ft
SARAH FORMATION ( fans)
20ft

SARAH FORMATION (turbidite fans)

20ft
SARAH FORMATION ( fans)

SARAH FORMATION (turbidite fans)

QASIM FM. SARAH FORMATION (turbidite fans)
20ft
20ft

20ft
0
0
00 450
0 3150 0
315
0 45 0
0
0 45
0
315
270 900
SARAH FORMATION (turbidite fans)

0 0
270 90
0
270 90
0
0 0
225 135
2250 135
0
20ft

0
180 0
180
0
225 135
0
00
0 0
3150 45 180

0 0
270 90

0 0
225 135
1800

PT
0
0
0
45

QASIM FORMATION
0
315

0 0
270 90

2250
QASIM FORMATION

1350
0
180
0
0
SARAH FM. (channels)
0
3150 45

0 0
270 90

0 0
225 135

RI
0
180
0 0
0
3150 45

Core 3
0 0
270 90

0 0
225 135
0
180
SARAH FORMATION (turbidite channels)

SARAH FORMATION (turbidite channels)

QASIM FORMATION

SC

QASIM FORMATION
0
0
0
315
0 45

0
270 900

0 0
225 135
1800

U
AN
Key
FA6- organic rich graptolitic black shale
(Qalibah Formation/Qusaiba Member)
FA5- tidally influenced, locally bioturbated
cross bedded sandstone (Uqlah Formation)
M

FA4- slumped and dewatered regionally

deformed heterolithics (Sarah Formation)
FA3- climbing dune cross bedded
sheet sandstones (Sarah Formation)
D
QASIM FORMATION

FA2- climbing dune cross bedded

channel fill sandstones (Sarah Formation)
FA1- storm influenced mid-outer shelf
TE

heterolithics (Qasim Formation)

00
0
45
3150

900 paleoflow
0
225 1350
1800

formation boundary
EP

change in depositional facies

C

(b) Location of wells and channels (c) Cross section model across outwash fan channels
10
AC

Eddy-related sidebar
17
4 5
N
6
2 Main flow
2D climbing dunes
9

0 10 km

Figure 6.
 ACCEPTED
0.4 EV2 MANUSCRIPT
Group 4: elements concentrated
Group 8: drilling additives in various minerals associated
0.3 with the fine sane, slit and/or
Zn clay fractions
Ba
Ca Sr
Group 2: carbonate minerals
Mn Cs
0.2
Mo Fe

Group 7: pyrite Mg
Co Cr
0.1 Be V
Cu Ni
Sc
Group 6: P-bearing heavy minerals P Rb Ga
and/or biogenic phosphate K Al EV1
0.0
-0.2 -0.1 0.0 0.1 0.2 0.3

PT
Na MREE
Ta
-0.1
Y U
Nb
HREE

RI
-0.2 LREE Ti
Th

Hf Group 3: heavy minerals

SC
Si
-0.3 Zr
Group 1: quartz Group 5: zircon

-0.4

U
Figure 7.
AN
M
D
TE
C EP
AC
 (a) (b)
100 25

90 ACCEPTED MANUSCRIPT
20

80

15

Ca %
70
Si %

60
10

50

negative trend 5 calcareous samples

40
calcareous samples

PT
30 0
0 5 10 15 20 25 30 30 40 50 60 70 80 90 100
Al % Si %
(c) (d)
7 40

RI
6 35

30
5

SC
25
4 strong positive trend
Ga ppm
positive trend
20
K%

3
15

U
2
10
AN
1 5

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Al % Al %
M

(e) (f)
40 2.5

35
D

2
30
TE
Ta ppm

25
1.5
Nb ppm

strong positive trend

20
strong positive trend 1
15
EP

10
0.5
5
C

0 0
0 0.5 1 1.5 2 0 10 20 30 40
Ti % Nb ppm
AC

(g) (h)
40 80

35 70

30 positive trend 60

25 50
Th ppm

strong positive trend

Hf ppm

20 40

15 30
Key
10 20 mudrock sample

10
sandstone sample
5

0 0
0 0.5 1 1.5 2 2.5 0 500 1000 1500 2000 2500 3000
Ta ppm Zr ppm

Figure 8.
 20ft Depth

0
GR
API
150

Figure 9.
QASIM FM. SARAH FORMATION (turbidite fans) QUSAIBA MBR. Strat.

0
UQLAH

Si/Al
A sand

25 0
Zr/Y

JL1 JL2
60 0
JL1

JL2
Th/P
1500

JL2-2
Ta/P

JL2-3
JL2-1
12 0

JL2-2
Well 8
Cr/P

JL2-3
JL2-1
400 0

JL2-1a
JL2-1b
P/Lu

JL2-1c
0.75 15

JL2-1a and JL2-1b

Th/Ti

JL2-1c
35 7

JL2-2b
Cr/Yb

JL2-2a
AC
180

C JL2-2b
Zr/Y

JL2-2a
60 9

JL2-3a
EP
JL2-3b
Th/Ta
20

JL2-1c JL2-3b chemo

TE
JL

JL
2-

2 -1
D
2a

c
JL

JL

JL2-1b
2-
JL2-3b

2-3

2b

M
a?

20ft Depth
0

AN
GR
API
ACCEPTED MANUSCRIPT

150

QASIM FM. SARAH FM. (channel)

U
SARAH FORMATION (turbidite fans) QUSAIBA Strat.
0

UQLAH
SC
Si/Al

A sand
40 0

JL1
Zr/Y

RI
JL2
60 0

JL1
PT
JL2
Th/P
150 0

JL2-2
Ta/P

JL2-3
JL2-1
12 0

JL2-2
Cr/P

JL2-3
JL2-1
400 0
Well 3

JL2-1a
JL2-1b
P/Lu

JL2-1c
0.75 15

JL2-1a and JL2-1b

Th/Ti

JL2-1c
35 7

JL2-2b
Cr/Yb

JL2-2a
18 0

JL2-2b
Zr/Y

JL2-2a
609

JL2-3a

JL2-3b
Th/Ta
20

JL2-1 JL2-2b JL2-3b chemo

JL1
 ACCEPTED MANUSCRIPT

PT
zone JL1 zone JL2
Zr/Y generally < 20 Zr/Y generally > 20

RI
Th/P generally < 35 Th/P generally > 35

U SC
subzone JL2-1 subzone JL2-2 subzone JL2-3
Ta/P > 5 (> 9 in Well 2) Ta/P < 5 (< 9 in Well 2) Ta/P > 5 (< 9 in Well 2)
AN
Cr/P > 183 (> 250 in Well 2) Cr/P < 183 (< 250 in Well 2) Cr/P > 183 (< 250 in Well 2)
M

division JL2-2a division JL2-2b division JL2-3a division JL2-3b

Cr/Yb > 12.4 Cr/Yb < 12.4 Th/Ta < 13.1 Th/Ta > 13.1
D

Zr/Y > 25 (11 in Well 9) Zr/Y < 25 (11 in Well 9)

TE

division JL2-1a division JL2-1b division JL2-1c

P/Lu > 0.36 P/Lu < 0.36 P/Lu < 0.36
Th/Ti < 21.97 Th/Ti < 21.97 Th/Ti > 21.97
EP

division JL2-1b’
same as JL2-1b
but higher Zr/Y
C

Figure 12.
AC
 a) 25 Well 2 b) 25 Well 7
ACCEPTED MANUSCRIPT
20 20

15 subzone JL2-3 15

Ta/P
Ta/P

10 10 subzone JL2-3

5 5

PT
subzone JL2-2
0 0 subzone JL2-2
0 200 0 200 400 600
Cr/P 400 600

RI
Cr/P

c) 25 Well 6

SC
d) 25 Well 8

20 20

U
AN
15 15
Ta/P

subzone JL2-3
Ta/P

subzone JL2-3
10
M

10

5 5
D

subzone JL2-2
TE

subzone JL2-2
0 0
0 200 400 600 0 200
Cr/P Cr/P 400 600
EP

e) 25 Well 5
C

20
AC

15 Key
Ta/P

subzone C2-2
subzone JL2-3
subzone C2-3
10 sample off scale

subzone JL2-2
0
0 200 400 600
Cr/P
Figure 13.
Figure 15.
10ft Depth
0
GR
200

SARAH FORMATION QUSAIBA MEMBER Strat.

200

JLM2
Zr/Lu

JLM1
650 0.5
Well 7

JLM3
JLM1 &
Th/Nb

JLM2
17

JLM3-4
JLM3-2
Ta/P

JLM3-1,
15

JLM3-3 & JLM3-5

Chem.
JLM3-4

JLM3-3

AC
0

10ft Depth
0

C
GR
200

EP
QASIM FORMATION SARAH FORMATION QUSAIBA MEMBER Strat.
JLM2
200
Zr/Lu

JLM1
TE
ACCEPTED

D
650 0.5
Well 1

JLM3
JLM1 &
Th/Nb

JLM2
M
17

JLM3-4
JLM3-2
Ta/P

JLM3-1,
AN
15

JLM3-3 & JLM3-5

MANUSCRIPT

U Chem.
JLM-2
JLM3-2
JLM3-3

SC
RI
10ft Depth
0

PT
GR
200

QASIM FORMATION SARAH FORMATION QUSAIBA MEMBER Strat.

JLM2
200
Zr/Lu

JLM1
650 0.5

JLM3
Well 8

JLM1 &
Th/Nb

JLM2
17

JLM3-4
JLM3-2
Ta/P

JLM3-1,
15

JLM3-3 & JLM3-5

Chem.
JLM2
undiff.

JLM3-2
JLM3-3
 ACCEPTED MANUSCRIPT

PT
zone JLM1 zone JLM2 zone JLM3
Zr/Lu > 355 Zr/Lu generally < 355 Th/Nb < 0.87

RI
Th/Nb < 0.87 Th/Nb > 0.87

U SC
AN
M
D

subzone JLM3-1 subzone JLM3-2 subzone JLM3-3 subzone JLM3-4 subzone JLM3-5
Ta/P > 12 Ta/P < 12 Ta/P > 12 Ta/P < 12 Ta/P > 12
TE

Figure 16.
C EP
AC
 ACCEPTED MANUSCRIPT

Final Chemo
Sandstone

Formation
Mudrock

Group
Scheme
Scheme

Scheme
undiff. sample gap

Qalibah Group
Qusaiba Mbr.
JLM3-5 J5

PT
RI
A
AJL2-2b or Uqlah
AJL2-3b A sand Fm.

SC
JLM3-3

JLM3-4

JL2-3b J4-3b

U
JL2-3a J4-3a
JLM3-2 J4-2b
AN
JL2-2b

JL2-2a J4-2a
possible
JLM3-2
M

JL2-1c or undiff. J4-1c

Sarah Formation
D

JL2-1b’ J4-1b’
TE

JLM3-1
JL2-1b J4-1b
EP

JL2-1a J4-1a
undiff. sample gap

undiff. undiff.
C

sample gap sample gap

AC

JL1 J3
Qasim Formation
undiff. sample gap

JLM2 J2

JLM1 J1

Figure 17.
 20ft Depth

0
GR
150
Well 7
QASIM FORMATION SARAH FM. SARAH FM. (fans) QUSAIBA Strat.
sample gap J4-2b J4-3b Chemo

N
(lower fans)

0
TURBIDITE FANS

A sand
20ft Depth
0

Figure 18.
GR
150
Well 1

QASIM FORMATION SARAH FORMATION (fans) QUSAIBA MEMBER Strat.

J2

7
J4-2 J4-3 Chemo

1
TURBIDITE FANS

10 km
10
J4-3b
A sand

8
9
20ft Depth
0

4
GR

3
150

6
Well 8

QASIM FORMATION SARAH FORMATION (fans) Strat.

5
J4-2b

J2
J4-1c J4-3b sample QUSAIBA MEMBER Chemo
gap
J4-2
a
A sand

J4

J4-1b
-1c
J4-3a

20ft Depth
0
GR
150
Well 3

QASIM FM. SARAH FORMATION (fans) Strat.

RB
A

J4-1 QUSAIBA MEMBER

TU J3
ID

J4-2b J4-3b Chemo

TI
EC
H AN

TURBIDITE FANS
NE
L

Well
in lower Faults
Channels
Sarah
AC
20ft Depth
0

C
GR
150

o
Well 9

QASIM FORMATION SARAH FORMATION (fans) QUSAIBA Strat.

J2
A

J4-1b J4-1b' J4-2b Chemo

undiff.

EP
J4-2a

TE
TURBIDITE CHANNEL
SHALLOW MARINE

D 20ft Depth
0
GR
150

M
Well 2

QASIM FORMATION SARAH FORMATION (turbidite channels) SARAH FORMATION (turbidite fans) QUSAIBA MBR. Strat.
sample gap J4-3b sample gap Chemo
J1
J2
J3

J4-2b
TURBIDITE CHANNEL
J4-3a

AN
ACCEPTED MANUSCRIPT

FANS

J4-2a

U
TURBIDITE

J4-1c

20ft Depth
0

SC
GR
150
Well 6

QASIM FORMATION SARAH FORMATION (turbidite fans) QUSAIBA Strat.

A

J4-1b' J4-1c J4-3 Chemo

J2

J4-1b
-1a

J4
RI
-1a
J4-2b

PT
J4-3
a

J4-2b
SHALLOW MARINE

20ft Depth
0
GR
150
Well 5

QASIM FORMATION SARAH FM. fans QUSAIBA Strat.

A

J2
J5

J4-3b Chemo
FANS
TURBIDITE

TURBIDITE CHANNEL
20ft Depth
0
GR
150
Well 4

SARAH FORMATION (turbidite channels) SARAH FORMATION (fans) Strat.

sample gap J4-1b J4-3b J4-3b gap Chemo
J4-2b
J4-3a

J4-2a

TURBIDITE CHANNEL
SHALLOW MARINE

20ft Depth
0
GR
150
Well 10

QASIM FORMATION SARAH FORMATION (turbidite fans) QUSAIBA Strat.

J4-3a
J1
J2

J4-3b undiff. Chemo

 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 20ft Depth

0
GR

150

Figure 10.
SARAH FORMATION ( channel) SARAH FORMATION (turbidite fans) Strat.
0
Zr/Y
60 0

JL1 JL2 JL1

Th/P

JL2
150 0

JL2-2
Ta/P

JL2-1 & JL2-3

12 0

JL2-2
Cr/P

JL2-1, JL2-3
400 0
Well 4

JL2-1b, JL2-1c
P/Lu

JL2-1a
0.75 15

JL2-1a, JL2-1b
Th/Ti

JL2-1c
35 7

JL2-2b
Cr/Yb

JL2-2a
18 0

AC
C JL2-2b
Zr/Y

JL2-2a
60 9

JL2-3a
EP
JL2-3b
Th/Ta
20

JL2-1b
TE
JL2-3b Chemo.
JL2-3a

JL2-2b

D
M
20ft Depth
JL2-2a?
0

AN
GR
ACCEPTED MANUSCRIPT

150

U
SARAH FORMATION (turbidite fans) Strat.
0

SC
Zr/Y
60 0

JL1 JL2 JL1

RI
Th/P

JL2
150 0

PT
JL2-2
Ta/P

JL2-1 & JL2-3

12 0

JL2-2
Cr/P

JL2-1, JL2-3
400 0
Well 10

JL2-1b, JL2-1c
P/Lu

JL2-1a
0.75 15

JL2-1a, JL2-1b
Th/Ti

JL2-1c
35 7

JL2-2b
Cr/Yb

JL2-2a
18 0

JL2-2b
Zr/Y

JL2-2a
60 9

JL2-3a

JL2-3b
Th/Ta

JL2-3a JL2-3b Chemo.

 ACCEPTED MANUSCRIPT

Well 6

Chemo.
Depth

Strat.

GR Si/Al Zr/Y Th/P Ta/P Cr/P P/Lu Th/Ti Cr/Yb Zr/Y Th/Ta
0 API 150 0 40 0 60 0 150 0 12 0 400 0 0.75 15 35 7 18 0 60 9 20

PT
QUSAIBA MEMBER

RI
20ft

UQLAH FM.

AJL2-3b

U SC
JL2-3a
AN
SARAH FORMATION (turbidite fans)

JL2-2b
M

JL2-2a
D

JL2-1c
TE

JL2-1b'
EP
SARAH FORMATION (turbidite channel)

JL2-1b
AC

JL2-1a
QASIM FORMATION

JL2-3a
JL2-1b
JL2-1a

JL2-2b

JL2-1 JL2-2 JL2-1 JL2-1b JL2-1c

A JL1 JL2 JL2-2 JL2-3 JL2-1c JL2-1a JL2-2b JL2-2a JL2-2a JL2-3b
JL1 JL2 JL2-3

Figure 11.
 ACCEPTED MANUSCRIPT

Well 9

Chemo.
Strat.
Depth

GR Si/Al Zr/Y Th/P Ta/P Cr/P P/Lu Th/Ti Cr/Yb Zr/Y Th/Ta
0 API 150 0 25 0 60 0 150 0 12 0 400 0 0.75 15 35 7 18 0 60 9 20
QUSAIBA MBR.

PT
UQLAH FM.

AJL2-2b
RI
SC
SARAH FORMATION (turbidite fans)

JL2-2a

U
AN
JL2-1b’
M

JL2-1b
20ft

D
TE
QASIM FORMATION

C EP
AC

undiff.
JL2-3a
JL2-1b
JL2-1a

JL2-2b

JL2-1 JL2-2 JL2-1 JL2-1b JL2-1c JL2-2a

A JL1 JL2 JL2-2 JL2-3 JL2-1c JL2-1a JL2-2b JL2-2a JL2-2b’ JL2-3b
JL1 JL2 JL2-3

Figure 14.
 ACCEPTED MANUSCRIPT

The highlights for this paper are as follows:

1. Produced a chemostratigraphic scheme comprising a hierarchical order of zones,

subzones, divisions and subdivisions.
2. Zonation is largely based on based on variations in elemental ratios reflecting

PT
changes in source/provenance
3. Using sedimentological, seismic and paleoflow data, it was possible to determine

RI
that the basal part of the Sarah Formation comprises turbidite channels which
are succeeded by fan deposits.

SC
4. Chemostratigraphy was not only utilized to identify the channel and fan deposits,
but enabled a much higher level of resolution and the correlation of individual

U
beds. AN
M
D
TE
C EP
AC