Late Ordovician deglaciation of the

Murzuq Basin (SW Libya): a core to

seismic-scale characterization of the
depositional environments and
sedimentary architecture F.J. Bataller
Visit to download the full and correct content document:
https://ebookmass.com/product/late-ordovician-deglaciation-of-the-murzuq-basin-sw-li
bya-a-core-to-seismic-scale-characterization-of-the-depositional-environments-and-se
dimentary-architecture-f-j-bataller/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Dynamic Sedimentary Environments of Mangrove Coasts 1st

Edition Daniel A. Friess (Editor)

https://ebookmass.com/product/dynamic-sedimentary-environments-
of-mangrove-coasts-1st-edition-daniel-a-friess-editor/

Church Architecture of Late Antique Northern

Mesopotamia Elif Keser Kayaalp

https://ebookmass.com/product/church-architecture-of-late-
antique-northern-mesopotamia-elif-keser-kayaalp/

Getting Started with Enterprise Architecture: A

Practical and Pragmatic Approach to Learning the Basics
of Enterprise Architecture Eric Jager

https://ebookmass.com/product/getting-started-with-enterprise-
architecture-a-practical-and-pragmatic-approach-to-learning-the-
basics-of-enterprise-architecture-eric-jager/

Trapped: Brides of the Kindred Book 29 Faith Anderson

https://ebookmass.com/product/trapped-brides-of-the-kindred-
book-29-faith-anderson/
The MENA Powers and the Nile Basin Initiative Okoth

https://ebookmass.com/product/the-mena-powers-and-the-nile-basin-
initiative-okoth/

Getting Started with Enterprise Architecture: A

Practical and Pragmatic Approach to Learning the Basics
of Enterprise Architecture 1st Edition Eric Jager

https://ebookmass.com/product/getting-started-with-enterprise-
architecture-a-practical-and-pragmatic-approach-to-learning-the-
basics-of-enterprise-architecture-1st-edition-eric-jager/

Petroleum Geology of Libya 2nd Edition Don Hallett

https://ebookmass.com/product/petroleum-geology-of-libya-2nd-
edition-don-hallett/

A Relational Ethics of Immigration: Hospitality and

Hostile Environments Bulley

https://ebookmass.com/product/a-relational-ethics-of-immigration-
hospitality-and-hostile-environments-bulley/

Late Europeans and Melancholy Fiction at the Turn of

the Millennium Ian Ellison

https://ebookmass.com/product/late-europeans-and-melancholy-
fiction-at-the-turn-of-the-millennium-ian-ellison/
 Marine and Petroleum Geology 134 (2021) 105335

Contents lists available at ScienceDirect

Marine and Petroleum Geology

journal homepage: www.elsevier.com/locate/marpetgeo

Late Ordovician deglaciation of the Murzuq Basin (SW Libya): a core to

seismic-scale characterization of the depositional environments and
sedimentary architecture
F.J. Bataller a, b, *, N.D. McDougall c, A. Moscariello a, M. Gil-Ortiz d, B. Cavalcanti de Paula e
a
Department of Earth Sciences, University of Geneva, Rue des Maraichers 13, CH-1205, Genève, Switzerland
b
Repsol Exploración S.A., Mendez Alvaro 44, 28045, Madrid, Spain
c
Independent Consultant, Madrid, Spain
d
Geomodels Research Institute, Facultat de Ciències de la Terra, Universitat de Barcelona, c/Martí i Franquès s/n, 08028, Barcelona, Spain
e
Petrobras, Petroleo Brasileiro S.A., Av. Henrique Valadares, 28, Rio de Janeiro, RJ, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: Glacial reservoirs have been the focus of numerous studies in various basins around the world, often with the aim
Sedimentology of supporting hydrocarbon exploration and development. These reservoirs present a significant challenge due to
Deglaciation their inherently complex internal architectures and significant lateral heterogeneities. This is of significant
Ordovician
importance, not only for both exploration and reservoir production optimization, but also as an opportunity for
Seismic
stratigraphic trapping and enhanced prospectivity in mature areas. This study is focused on a major glaciogenic
Stratigraphy
Depositional environment reservoir, of Late Ordovician age, in the Murzuq Basin (SW Libya). It aims to integrate subsurface datasets, such
Glacial as 3D seismic with well-based sedimentological data (core, conventional wireline logs and wellbore resistivity
images). The study proposes a comprehensive facies scheme optimized for the sub-surface but consistent with
outcrops, and aims to generate depositional models, characterize the complex internal architecture of the
reservoir, establish a genetic stratigraphic framework and evaluate facies in terms of reservoir properties and
stratigraphic trap potential. Results, integrating subsurface facies analysis and well-calibrated seismic stratig­
raphy support previous models, derived from outcrop analysis, suggesting that these complex reservoirs cannot
be characterised by purely lithostratigraphic criteria. To aid in the assessment of these formations, a facies
scheme is proposed, focused on subsurface data but also considering other published, outcrop-based, schemes.
Additionally, an interpretation is presented of the internal architecture of the Upper Ordovician based on a
succession of several ice advance-retreat cycles, not all of which are present across the study area. Each cycle
corresponds to a seismic package interpreted to comprise the deposits of several key sedimentation phases,
resulting in a complex mosaic of genetic packages generated by multiple, often areally-restricted, cut and fill
events. Predictive stratigraphic models are proposed for this complex package, based on both seismic stratig­
raphy and well-based sedimentology. The implications for future prospectivity and the significance for strati­
graphic trap potential are also discussed.

1. Introduction
Ancient, glacial sediments have been studied for many years. Ex­
amples occur throughout the scientific literature, ranging in age from
the late Archean and late Proterozoic or Cryogenian “snowball Earth”
events (e.g. Kirschvink, 1992; Fairchild and Kennedy, 2007; Vandyk
et al., 2019), through the Late Ordovician (e.g. Beuf et al., 1971;
McDougall and Martin, 2000; Bataller et al., 2019; Ghienne, 2003; Le
Heron et al., 2007) and the Permo-Carboniferous event (Le Heron et al.,
2019; Visser, 1987, 1990; Dietrich and Hofmann, 2019). Much of this
work has been supported by an ever-growing body of literature focused
on both modern and Pleistocene glacial sediments. Relevant examples
for the current study include the work of Dietrich et al. (2017b), focusing
on retreat of the Laurentide Ice Sheet in eastern Canada and

* Corresponding author.Department of Earth Sciences, University of Geneva, Rue des Maraichers 13, CH-1205, Genève, Switzerland.
E-mail addresses: fj.bataller@repsol.com (F.J. Bataller), neil85mcdougall@gmail.com (N.D. McDougall), andrea.moscariello@unige.ch (A. Moscariello),
amarquitus@gmail.com (M. Gil-Ortiz), br_cavalcanti@hotmail.com (B. Cavalcanti de Paula).

https://doi.org/10.1016/j.marpetgeo.2021.105335
Received 12 July 2021; Received in revised form 13 September 2021; Accepted 15 September 2021
Available online 21 September 2021
0264-8172/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
F.J. Bataller et al.
seismically-driven studies on the sub-surface of the southern North Sea
(Moreau et al., 2015; Benvenuti et al., 2018 and Benvenuti and
Moscariello, 2016). Such studies of ancient glacial systems have also
become increasingly relevant when considered within the framework of
climate change and the likely major impact of increasing temperatures
on global ice masses. In this respect the late Ordovician (Katian?-Hir­
nantian) glaciogenic sequence has been studied as a potential analogue
by several authors due to its similarities with this modern climatic
phenomenon (Ghienne et al., 2014; Zalasiewicz and Williams, 2014)
and thus the potential to improve understanding of modern processes of
deglaciation. Furthermore, this succession also provides us with a very
good opportunity to study an ancient event linked to one of the largest
mass extinctions in the geological record (Ordovician-Silurian).
The stratigraphy, facies and depositional environments associated
with the late Ordovician glaciation in both North Africa and the Arabian
Peninsula (Morocco, Algeria, Libya, Niger, Mauritania, Jordan, Saudi
Arabia, Oman), have been extensively investigated by many authors.
Much of the literature is based on high quality, high resolution outcrop
analysis (Beuf et al., 1971; McDougall and Martin, 2000; McDougall
et al., 2004; Blanpied et al., 2000; Deschamps et al., 2013; Ghienne and
Deynoux, 1998; Ghienne et al., 2007a, 2007b and 2010; Girard et al.,
2012a, b; Le Heron et al., 2004, 2009; Hirst, 2012; Douillet et al., 2012;
and Clerc et al., 2013; amongst others); In contrast, notably less work
has been done on the subsurface aspects of the same glaciogenic suc­
cession which, at least in SW Libya and SE Algeria, is a major hydro­
carbon reservoir, (i.e. Girard et al., 2019; Lang et al., 2012; Le Heron,
2010; Bataller et al., 2019; Abushalah and Serpa, 2016, 2018, McDou­
gall et al., 2003; McDougall and Gruenwald, 2011; Ron et al., 2016).
Additionally, some authors have sought an integration between the
subsurface and outcrop, both in Libya and Algeria (Roussé et al., 2009;
Moreau and Joubert, 2016; McDougall et al., 2004, 2005, 2006, 2007,
2008), and also with modern analogues (Bataller et al., 2019) by means
of seismic spectral decomposition and palaeogeomorphology. Such
subsurface-based studies are, however, notably less frequent and limited
in scope than outcrop-focused publications, a direct reflection of the
limitations imposed by the confidentiality requirements of both oper­
ating and national oil companies. A principal aim of the present work is
therefore to fill this significant gap in the literature by presenting the key
results of a high-resolution integration between 3D seismic, cores and
image log data through the glaciogenic succession in SW Libya.
The study area for this paper is located in the Murzuq Basin of SW
Libya and corresponds to the NC186 concession awarded by the Libyan
NOC to a partnership group comprising Repsol (REMSA), TotalEnergies,
OMV and Equinor. A complete, fully integrated dataset has been used for
the study; including 11 wells, both exploration and appraisal, most of
which offer conventional well logs, high-resolution core photographs,
detailed core descriptions and image logs. In addition, 3D seismic data
(time domain) has also been used extensively, to provide both a
framework for seismic stratigraphy and support for depositional
modelling in which the 11 wells act as calibration points. Although this
is a subsurface focused study, some complementary, unpublished
outcrop data and images have also been included to provide the link to
surface, outcrop-based description of these sediments.
The present work has 3 key objectives: (1) to present a practical
facies scheme and interpretation of depositional environments,
reviewing previously published outcrop-based schemes (taking
McDougall and Martin, 2000, as a starting point), updated with un­
published subsurface data, (2) to propose a genetic stratigraphic
framework and depositional architecture driven by the integration of the
key subsurface datasets (core, logs and seismic) and (3) to review and
assess the reservoir potential of the glaciogenic succession and its im­
plications for reservoir potential.
2
Marine and Petroleum Geology 134 (2021) 105335
2. Geological setting
2.1. Regional context
The area of interest for this paper is located within the Murzuq Basin,
SW Libya, one of a series of major intracratonic sag basins located across
the North African platform or Saharan Metacraton (Brahimi et al.,
2018). The basin is bounded by the Tihemboka Arch (Ghat and Tikiumit
outcrop areas) to the southwest, the Gargaf Arch outcrop to the north­
east and Jebel Ben Ghanimah (Mourizidie and Dor el Gussa outcrop
areas) to the southeast (Fig. 1). Elements of the basin framework origi­
nated as a result of the Proterozoic Pan-African orogeny with key ele­
ments later re-activated during several tectonic phases. Most relevant
were the end-Ordovician, glacio-isostatic rebound or “Taconic” event
(Perron et al., 2018) and the end-Silurian “Caledonian” both of which
served, respectively, to generate initial hydrocarbon traps for the Upper
Ordovician and the essential configuration of the current sag basin.
These initial traps play a key role since the source rock (and seal) are
located directly above, shortening the migration pathway, and simpli­
fying the charge of this initial trapping structures. Later Hercynian,
“Austrian” and “Alpine” events also impacted some elements of the
basin, notably in terms of uplift and the associated tilting of hydrocar­
bon traps (Hallet, 2002).
The studied sedimentary succession is associated with advances and
retreats of the Late Ordovician Saharan ice sheet; which extended across
approximately 4000 km in North Africa (as shown in Fig. 1b), from the
Atlantic coast of Morocco to Egypt and on to the Arabian Peninsula (Le
Heron and Craig, 2008; Ghienne et al., 2003, 2010bib_Ghienne_e­
t_al_2003, 2010; Hayton et al., 2017). It is thought to have formed as
part of the Western Gondwana Ice sheet (Beuf et al., 1971; Vaslet, 1990;
Sutcliffe et al., 2000a, 2000b; Ghienne, 2003; Le Heron et al., 2004),
located at high paleo-latitudes (Kent and Van der Voo, 1990; Smith,
1997; Scotese et al., 1999). The duration of the glaciation has been
described as short, around 0.5–1 m.y (Brenchley et al., 1994) corre­
sponding to the Hirnantian Stage (445.2ma-443.8ma), which is
certainly shorter than the overall ice-house period, above 5ma (Ghienne
et al., 2014), which would also include a significant segment of the
Katian Stage (453ma-445.2ma). The duration of this glacial episode in
North Africa thus remains uncertain, and is, in any case, likely highly
diachronous across both North Africa and indeed globally (Finnegan
et al., 2011).
During the Late Ordovician glaciation major changes occurred in the
paleogeography of the Murzuq Basin. This reflects the multiple advances
and retreats of ice sheets which, together with consequent changes in the
paleoclimate resulted in the development of a wide range of depositional
environments ranging from proximal glacial braidplain or sandur to
deep glaciomarine (Grubic et al., 1991; Pierobon, 1991, Girard et al.,
2012a,b).
2.2. Stratigraphic context
In terms of regional stratigraphy, the main glacial bedrock is
composed of the Hawaz and, locally, the Achebayat (Hamra Quartzite or
Unit III-2 in SE Algeria) Formations, of Lower to Mid Ordovician age
(Arenig to Llanvirn-Llandeilo Stages in UK usage or Floian to Darriwil­
lian Stages in international terms). These are themselves significant
secondary oil and gas reservoirs in, respectively, the Murzuq and Illizi-
Berkine basins. They are characterised by, often intensively burrowed,
shallow to marginal marine, sand-prone tidal channel and nearshore
facies (McDougall et al., 2008; Ghienne et al., 2013; Gil-Ortiz et al.,
2019), truncated by a major regional unconformity with significant
paleotopography (Ghienne et al., 2003) typically in the form of paleo­
valleys and bounding Hawaz paleohighs, marking the base of the Upper
Ordovician glacial episode.
Overlying this unconformity, there is often a mud-prone package of
argillaceous siltstones, sandy mudstones and argillaceous sandstones
F.J. Bataller et al. Marine and Petroleum Geology 134 (2021) 105335

Fig. 1. (A) Google Earth™ image highlighting the location of the study area (NC186 concession by the red rectangle) and the main outcrop locations for Late
Ordovician glacial sediments within the Murzuq Basin, Libya. (B) Modified from Bataller et al. (2019), highlighting previously published examples of evidence for the
end-Ordovician glaciation in North Africa. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

commonly known as the Melaz Shuqran Formation. These sediments
comprise two broad facies assemblages (McDougall and Martin, 2000);
(a) a “lower” assemblage, mainly proglacial-subglacial in origin (Blan­
pied et al., 2000; Le Heron et al., 2003), dominated by sandy mudstones
and heterolithic silty sandstones characterised by common soft sediment
deformation, including slumps, dewatering structures, slides, listric
faulting (McDougall and Martin, 2000; Le Heron et al., 2003) and
floating granules, pebbles and mudclasts (Collomb, 1962, El-Ghali,
2005); and, less commonly, (b) an “upper” assemblage more related
with late glacial advance and early retreat, characterised by silty, het­
erolithic, rippled and locally burrowed mudstones (McDougall and
Martin, 2000; El-Ghali, 2005). This formation is interpreted to record
the highest sea relative level within the Upper Ordovician deposition
(McDougall and Martin, 2000).
Overlying the initial mud-prone package is the main reservoir or
Mamuniyat Formation. An essentially sand-prone package, it corre­
sponds to a complex infill phase of the remaining paleorelief, often, but
not always, comprising several sedimentary sub-units, bounded by un­
conformities (McDougall and Martin, 2000; Ghienne et al., 2003,
2007abib_Ghienne_et_al_2003, 2007a; Le Heron et al., 2004,
2006bib_Le_Heron_et_al_2004, Roussé et al., 2009; Deschamps et al.,
2013) presumably recording several phases of ice advance and retreat
and deposition in braidplain to glaciomarine environments.
The mud-prone Bir Tlacsin Formation unconformably overlies the
Mamuniyat and is also present in the study area, although not contin­
uously. This final package has been described as a transgressive unit
between the Upper Ordovician sandstones and the ‘Hot Shales’ of the
overlying Tanezzuft Formation (Echikh and Sola, 2000), although the
unconformable stratal relationships observed, most notably in the sub­
surface, suggest this is most likely, an overly simplistic interpretation.

Fig. 2. Synthetic stratigraphic columns: (A) of the Paleozoic succession in the Murzuq Basin, the late Ordovician glacial unit and chronostratigraphy and; (B)
comparison, based on previous studies (McDougall and Martin, 2000; Blanpied et al., 2000; Echikh and Sola, 2000; Le Heron et al., 2009; and Ghienne et al., 2013) of
the key lithostratigraphic units.

3
F.J. Bataller et al.
More recently, the Bir Tlacsin has been re-interpreted as a late-stage
glacially-related cut and fill event (McDougall and Hussein, 2007;
McDougall et al., 2008; Moreau, 2011; Bataller et al., 2019), related to
the uppermost Mamuniyat Formation. Usually, it is found both within
topographic lows/paleovalleys, and locally, over the tops of paleohighs,
all of which suggests that, prior to the marine or transgressive erosion at
the base of the Tanezzuft, this unit was most likely widespread and well
developed across the study area. Fig. 2, is a synthetic stratigraphic col­
umn of the study area based on the work of McDougall and Martin
(2000); Blanpied et al. (2000); Echikh and Sola, 2000; Le Heron et al.
(2009); and Ghienne et al. (2013), This stratigraphic framework was
initially conceived (McDougall and Martin, 2000) as a simple five-fold
genetic scheme based on the apparent correlatability between outcrop
and subsurface of key stratal boundaries. However, on the basis of
increasing well data, this scheme was subsequently also assumed to be
effectively lithostratigraphic in character. In contrast, Ghienne et al.
(2003) introduced concepts, based on mapping and outcrop observa­
tions in the Qarqaf Arch, focused on erosive surfaces and widely
correlatable allostratigraphic, rather than lithostratigraphic, units
within the Murzuq Basin. More recently, another genetic-based scheme
has also been proposed (Girard et al., 2012a, 2012b), focused on sedi­
ment distribution patterns within clinoforms, controlled by ice-front
evolution and melt water related processes in proglacial outwash
plains and delta systems. These authors, together with Moreau (2011),
created more sedimentologically dynamic, genetic schemes that better
reproduced the variations in outcrop which this study intends to
compliment by adding 3D subsurface data.
3. Data and methods
The dataset utilized for this study comprises 11 exploration and
appraisal wells together with a 3D seismic cube. In Fig. 3, the study wells
are located on a TWT map showing the main or basal glacial uncon­
formity equivalent to Top Hawaz Formation. The map was generated by
flattening Top Hawaz (picked in TWT throughout a 3D seismic cube)
along a high amplitude, regional-scale, intra-Silurian reflector, high­
lighted in Bataller et al. (2019). This was done in order to minimize the
potentially distorting effect of the regional structural dip (3–5◦ towards
Fig. 3. TWT map of the top Hawaz Formation (sub-glacial unconformity) flattened on an intra-Silurian reflector (Bataller et al., 2019) highlighting the location of the
wells used in this study and the dataset available for each well.
4
Marine and Petroleum Geology 134 (2021) 105335
the south) and better highlight the paleogeography at the onset of Late
Ordovician glacial sedimentation as described in Bataller et al. (2019).
3.1. Core
The most important element of the dataset in this study is the core.
1054 ft of 6 inch diameter core were cut from 7 wells (Table 1, Fig. 3).
This data represents our only “ground truth” and sample of the Upper
Ordovician in the subsurface, which acts as the calibration point for both
conventional and image log interpretation. After an initial wellsite
description, the cores were all slabbed, photographed (later scanned at
very high resolutions) and described in detail at a scale of 1.50.
The initial description of the cores followed the McDougall and
Martin (2000) facies scheme. For this study a re-interpretation was
made, using both the original descriptions and high-resolution core
images, with a special focus on the following key elements: lithologies,
clast and visible cement types, bed thickness, grading trends, grain size,
sorting, a visual estimate of detrital clay content, sedimentary struc­
tures, bioturbation and colours. This reinterpretation aimed to update
the previous, outcrop based, facies scheme with newer subsurface core
data, within the framework of recent key publications from both SW
Libya and SE Algeria (Blanpied et al., 2000; McDougall et al., 2004;
Ghienne et al., 2007a, 2007b; Girard et al., 2012a, 2012b; Le Heron
et al., 2004, 2009) as well as considering advances in both our general
understanding of glacial sedimentology and other ancient glaciations
(Kurjański et al., 2020; Le Heron et al., 2019; Dietrich and Hofmann,
2019).
In addition, Conventional Core Analysis (CCA), measuring He
porosity, Horizontal Permeability (Kh), some Vertical Permeability (Kv)
and grain density (GD), was also carried-out on core plugs through most
of the cored intervals at a spacing of between 1 and 3 ft. This extensive
dataset forms the basis for the facies-based evaluation of porosity and
permeability presented in this study. The CCA data was acquired at at­
mospheric conditions, corrected for the Klinkenberg effect but uncor­
rected for overburden stress, although experience in the area, based on
Special Core Analysis or SCAL results, performed at reservoir pressure
and temperature (available from one of the wells reviewed; Well G),
shows that porosity and permeability are reduced only by an average of
F.J. Bataller et al. Marine and Petroleum Geology 134 (2021) 105335

Table 1
Data inventory detailing the available core and log datasets for the study wells.
Well Core Length (ft) CCAL SCAL SGR Density + PEF and Neutron Magnetic Resonance Sonic FMI
Well A Y 70 Y N Y Y Y Y Y
Well B no core N N Y Y Y Y Y
Well C Y 125 Y N Y Y Y Y Y
Well D Y 14 Y N Y Y N Y Y
Well E Y 331 Y N Y Y N Y Y
Well F Y 132 Y N Y Y N Y Y
Well G Y 266 Y Y Y Y N Y Y
Well H no core N N Y Y N Y Y
Well I no core N N Y Y N Y Y
Well J Y 116 Y N Y Y Y Y Y
Well K no core N N Y Y N Y Y

around 2% at current burial depths. dataset. All have a conventional log suite including Gamma Ray (GR),
Spectral GR (SGR), Caliper, Resistivity, Sonic, Density, Photoelectric
3.2. Conventional wireline and image logs Factor (PEF) and Neutron porosity. Additionally, a subset of the wells
(Table 1) has advanced petrophysical logs such as Nuclear Magnetic
Each of the 11 study wells also has a high-quality wireline log Resonance (NMR/CMR) which, together with the other conventional

Fig. 4. A: from left to right, Raw FMI data (Track 1), Static (Track 2) and Dynamic (Track 3) processed images. B: Magnetometer data showing the tool rotation based
on magnetic data in the borehole. C: Accelerometer data showing the movement within the borehole both on X and Y axis rotation (top) and Z axis (bottom)
indicating that no significant stick and pull events took place as the distribution of the values are centred in the gravity normal value.

5
F.J. Bataller et al.
logs, were mainly used for quality control of the measured core porosity
and permeability results provided by the laboratory.
The logs were acquired mostly in 12.25′′ and 8.5” vertical boreholes,
using Water Base Mud (WBM) at a standard vertical sample rate (0.1524
m). Core data has been depth shifted to wireline reference using GR logs
(with shifts ranging from 0.3 to 3 m). Boreholes are generally in good
condition and the core plus wireline logs dataset is considered to be of
high quality.
In addition, high resolution micro-resistivity image logs were also
available in all 11 study wells, and are a key piece of sedimentological
data as core was not taken, either in every well or indeed recovered
through the complete drilled glaciogenic succession in each of the study
wells. The Formation Micro Imager (FMI) tool was acquired by
Schlumberger with approximately 60% and 80% radial borehole
coverage (depending on the hole size).
The processing of the raw data comprised 3 key steps: (1) navigation
quality control through the magnetometer and accelerometer data as­
suring tool orientation centralization and stick and pull effects using the
Z axis accelerometer (Fig. 4b and c); (2) Speed correction assuring
correct depth using stick and pull effects through the cable tension; and
(3) generation of the final Static (constant values for the colour scale
end-points) and Dynamic (varying end-point values for the colour scale
aiming to maximize contrast between resistive and conductive features)
images to better facilitate interpretation. (Fig. 4, a).
Review of core descriptions, high resolution core photographs and
the image logs has been used to generate a selection, from across all 11
study wells, of key sedimentological features (Fig. 5) which were sub­
sequently used as a guide for the development of a facies association
scheme, depositional modelling and correlations calibrated to seismic.
Fig. 5. GR profiles through the Upper Ordovician section for the 11 well study wells showing: (a) the depths of cored intervals (black bars), and (b) the stratigraphic
position of the key sedimentological features, identified from FMI and/or core Hi-Res photographs, referenced in the text and Figs. 8–10. Additional information
about the data available for each well can be found in Table 1.
6
Marine and Petroleum Geology 134 (2021) 105335
3.3. 3D seismic
The seismic dataset utilized for this study is a pre-stack, time
migrated 3D cube, interpreted in Two Way Time (TWT) in order to avoid
any additional uncertainty due to time depth conversion artefacts.
Additionally, this dataset was flattened along a distinctive intra-Silurian
marker horizon to enhance reflector relationships and terminations.
Bataller et al. (2019) have demonstrated how this flattening improves
the visualization of the data and how it assists in the reconstruction of
depositional environments, although the limitations of this approach,
should always be kept in mind (Bataller et al., 2019).
To integrate the well data (depth reference) with the seismic volume
(time reference), a well tie process was carried-out for each of the 11
study wells (Fig. 6). The first step was to produce a synthetic seismogram
for each well, which is the calculated result of the expected seismic
response of the rock sequence in a well path generated by treating it as a
one-dimensional problem (Bacon et al., 2003) and calculating acoustic
impedance (Yilmaz, 2001) utilizing density and sonic logs.
Once the acoustic impedance has been calculated, zero-phase
wavelets were utilized to complete the synthetic seismograms (Wang,
2015). The seismic volume has a dominant frequency of around 30–35
Hz (ranging between 15 and 70 Hz, as shown in Bataller et al., 2019)
with an average of 25–30 m vertical resolution, at the depth of interest,
but ranging between 20 and 40 m depending on overlying Silurian Hot
Shale thickness. In terms of polarity, the standard polarity is considered
for this study, which is defined by the Society of Exploration Geo­
physicists (SEG) as positive for an increase in acoustic impedance and so
displayed as a peak, and negative for a decrease in acoustic impedance,
displayed as a trough.
F.J. Bataller et al.
Fig. 6. Example of a seismic well tie (Well A) showing the wavelet, power spectrum (division of power into frequency components), density and sonic logs, resultant
reflectivity, seismic data and synthetic seismogram.
3.3.1. Seismic facies and stratigraphy
One of the key aims of this paper is to extract sedimentological
(facies) and stratigraphic data from the seismic cube. The methodology
used in this paper focused on three key concepts: (1) reflector termi­
nations (stratal termination), similar to the ideas proposed by Ramsayer
(1979) and Mitchum et al. (1977), (2) reflector amplitude and seismic
facies variations and (3) morphology of the features. After a full review
of the Upper Ordovician interval in the seismic cube, both inlines,
crosslines and random seismic lines, a range of key features (Fig. 7) have
been identified. These have been used as input into the fully integrated
interpretation of depositional environments, stratigraphic development
and sedimentary architecture presented later as key conclusions. Each of
these key seismic stratigraphic elements are summarized as follows:
- Parallel reflectors: these indicate that sedimentation generally took
place without significant discordances. Smaller-scale un­
conformities, minor channels or associated features may be unre­
solved due to seismic resolution. Several examples of these patterns
can be seen in pre (Fig. 7, A and G) and post (Fig. 7A and B) gla­
ciogenic sediments (Fig. 7, A, H, D, F and I).
- Reflector truncations: within the context of the study area, these are
interpreted to correspond to erosional unconformities. These are
most likely of sub-glacial origin cutting down through both through
the underlying Hawaz Formation (resulting in the generation of
major incisions or paleovalleys up to 350 m in depth and between 1
and 3 km in width) forming a pre-Upper Ordovician paleotopog­
raphy (Fig. 7, A, D, G and I), and also intra-glacial sediments creating
wide U- (Fig. 7, D and H) or V-shaped (Fig. 7, A, F and I) incisions, the
latter possibly a response to downcutting of pro-glacial fluvial sys­
tems during sea level falls. These range in size from 1 or 2 km in
width, in the V-shaped case, and up to tens of km for the U-shaped
incisions.
- Downlaps: defined as a downwards-dipping reflector terminating
against an older, lower angle reflector. It is a morphology typically
7
Marine and Petroleum Geology 134 (2021) 105335
associated with progradation, specifically the toesets of clinoformal
bodies. Clear examples shown in Fig. 7 (B and E).
- Onlap: is defined as horizontal reflectors lapping against older,
steeper-dipping reflectors and paleohighs; be they of pre-glacial
bedrock or intraformational origin. In the context of this study,
four distinct cases can be distinguished: (1) onlap against pre-glacial
sediments (Fig. 7, G and I), (2) onlap against chaotic seismic facies
(Fig. 7, A, E and H), (3) onlap against intra-glacial incisions (Fig. 7D
and E and F) and (4) onlap against clear glaciogenic clinoforms
following a transgressive pattern (Fig. 7B and C and E).
- Chaotic facies: this seismic facies corresponds to sediments without
any consistent or continuous internal reflector structure. Examples
are most likely related to pro-glacial mass transport complexes,
generated by slope instability due to high sediment rates and/or sub-
glacial glaciotectonics. Extensive faulting may also be associated
with these packages. Some examples are shown in Fig. 7 (C, D, E and
F) which are typically remnant patches of chaotic features, later
incised and infilled by packages characterised by onlapping
reflectors.
- Pinch out: although is not a common feature in the study area, a very
clear example is seen in Fig. 7 (C). This is interpreted as sediment
infilling paleo-topographic depressions (mostly glacial erosion fea­
tures) and thinning towards the flanks.
4. Results
4.1. Facies analysis and general depositional environments
4.1.1. Proglacial Sandur-Active Braid Delta Front (PSB)
This GDE comprises 8 component facies associations (Table 2). It is
largely, but not exclusively sand-prone, dominated by the deposits of
glacio-fluvial and subordinate marine processes in two major linked sub-
systems; (a) a Pro-Glacial Delta Plain or sandur comprising Proximal
Channel/bar (PSB-fCH) and Sheetflood-Argillaceous Sandflat elements
(PSB-DFps, PSB-DPds & PSB-DPas) dominated by fluvial or unconfined
F.J. Bataller et al. Marine and Petroleum Geology 134 (2021) 105335

Fig. 7. Key seismic stratigraphic features identified within the Upper Ordovician succession the study area. These have been defined using as criteria: (1) reflector
facies and amplitude and (2) reflector terminations. The location of each example is shown on the seismic base map of Fig. 3 at bottom left. Also shown are the legend
for reflector types and the scale bars. In each of the cross-sections the left-hand and bottom scale bars respectively represent vertical (400 m) and horizontal scale (1
km) and the right-hand bars represent TWT scales (200 ms).

8
F.J. Bataller et al.
Fig. 8. Summary of key characteristics of the component facies associations comprising GDE PSB or Proglacial Sandur Braid Delta Front. Figure shows simplified
descriptions and GR curves for each association. Also shown are a selection of key images from core, FMI (see Fig. 5 for the depth and well of each image).
sheetflow processes, and occasional but volumetrically significant
outburst or Jokhulhaup events; passing downstream into; (b) an active
braid-delta front setting (fluvial-marine transition); comprising Dis­
tributary channels (PSB-dCH) and Proximal (PSB-pMB) to Distal (PSB-
dMB) mouth bar deposits. Less common are Mini- or Gilbert Delta de­
posits (APD-MD), infilling localised paleotopographic lows, such as
proglacial lakes or kettle holes (Kurjański et al., 2021).
Fluvial Channel (PSB–CH): This association typically occurs in dm-
scale stacked packages (<10 m thick) marked by an erosive base and an
aggradational or fining-upwards character. It comprises clean to slightly
argillaceous (0–5% detrital clay or Tmud), medium to coarse grained,
poor to moderately sorted sandstones often characterised by common,
moderate to poorly rounded, floating granules, pebbles and muddy
intraclasts Fig. 8 (H, n and M). The base of facies packages may also be
loaded or marked by pebbly lag deposits. Internal sedimentary struc­
tures are dominated by medium to large-scale sets of trough cross-
bedding, thick to very thick sets of low angle cross-stratification,
(locally interpreted, based on outcrop analogues, as climbing dune
cross-stratification or CDCS of Ghienne et al. (2013, 2010), planar or
undulatory parallel lamination and minor current ripples. Locally the
whole assemblage of structures may occur as a distinctive normal graded
facies sequence. Bioturbation is completely absent from these deposits.
The assemblage of lithofacies dominated by low angle cross-
9
Marine and Petroleum Geology 134 (2021) 105335
stratification-climbing megaripples cross-stratification could, arguably,
be distinguished as a separate facies association related to super high
energy outburst events (Ghienne et al., 2010) which is clearly different
to normal channel/bar events and associated cross-bedding. However,
given the inherent limitations of both core and image log data, when
compared to extensive 2 or 3D outcrops, it is not generally possible to
consistently differentiate such outburst events in this study and, as such,
it was decided to include these deposits within Facies Association
PSB-fCH.
Synthesis of both subsurface and outcrop (Annex I-A,C, Annex II,
Annex VIA-4 & 5), data suggests that geometries, are sheet-like to
broadly lenticular packages characterised by either: (a) stacked, low
width/depth (W/D) ratio multistorey/multilateral channel bodies, or
(b) discrete higher W/D ratio channel bodies, typically associated with
the outcrops of Tihemboka and the Tassili N Ajjers with the so-called
“Cordons” (Ghienne et al., 2010; Girard et al., 2012b; Deschamps
et al., 2013), some of which steeply incise the substrate, forming iso­
lated, bedrock-confined sand bodies ranging in width from <10 m to
several hundreds of metres. Where channel complexes are not so
confined they typically grade laterally into sheetflood deposits of As­
sociations DPps and DPds.
Interpretation: Facies Association PSB-fCH is mainly interpreted as
the deposits of sandy to gravelly bars and channel-fills formed by high
F.J. Bataller et al.
Fig. 9. Summary of key characteristics of the component facies associations comprising GDE ADP Active Pro-Glacial Delta Front. Figure shows simplified de­
scriptions and GR curves for each association. Also shown are a selection of key images from core, FMI (see Fig. 5 for the well and depth of each image).
energy, proximal, braided streams flowing across a pro-glacial braid-
delta plain or Sandur. In addition, as suggested above, some of these
deposits may also be the result of very high energy, sediment laden
glacial outburst events or Jokhulhaups; sometimes forming a distinctive
sequence of structures within single event beds, notably coarser grained
and/or dominated by climbing megaripples, as suggested by several
authors such as Ghienne et al. (2013) and Girard et al. (2012a). In
summary, FA PSB-fCH can be interpreted as a depositional continuum in
which end-members associated with low sinuosity, high energy fluvial
systems and outburst deposits are completely intergradational as a
function of flow regime and sediment supply.
Proximal Sheetflood Deposits (PSB-DPps): typically occurs in dm-
scale packages (<10 m thick) marked by a sharp or weakly erosive base
and an aggradational or weakly fining-upwards aspect (Annex I–C). It
comprises clean to slightly argillaceous (0–5% Tmud), fine to medium
grained, moderately to well sorted sandstones often displaying parallel
lamination as seen in Fig. 8 (J), small-scale cross-bedding, ripple cross-
lamination and dewatering structures. Bioturbation is completely ab­
sent. In outcrop, individual beds are typically; thin to medium bedded,
<30 cm thick and generally sheet-like although gutter casts and small
(metre-scale widths) shallow channel bodies are observed. Outcrops
(Annexes I and II, Annex VIA-1 and 2) also show this facies type to be
closely related, as a lateral equivalent, to Association PSB-fCH.
Distal Sheetfloods (PSB-DPds): are similar in many respects to PSB-
DPps. The Association is distinguished by weakly to non-erosive, sharp-
based metre-scale packages; (<5 m) characterised by an aggradational
to fining upwards stacking pattern. Packages comprise very fine to fine
grained, slightly to moderately argillaceous, moderate to moderately
good sorting, thin bedded sandstones and heterolithic sands (5–30%
Tmud); displaying common planar lamination and ripple cross-
lamination and dewatering structures. Bioturbation is completely
absent.
Interpretation: Both Associations PSB-DPps and DPds have a common
origin, recording deposition from poorly confined to unconfined,
decelerating sediment-laden sheet flows spreading across the pro-glacial
sandur, most notably during glacial flood or outburst events. Within this
framework, Association DPps would have been deposited from the most
10
Marine and Petroleum Geology 134 (2021) 105335
energetic flows and/or within relatively ice or channel proximal areas,
whereas Association DPds likely represents low energy deposition from
weaker flows or a more distal location. It is also probable that these
sheetflood deposits are gradational into climbing megaripples (Ghienne
et al., 2010) with increasing flow energy and sediment load.
Argillaceous sandflat (PSB-DPas): typically found in m-scale
packages (<5 m thick) with an aggradational and serrate GR profile. It
comprises moderately to highly argillaceous, very fine to fine, moderate
to well sorted, thin bedded sands and muddy heterolithics (30–50%
Tmud); (Fig. 8; ak). They are typically current rippled, flat laminated
and locally dewatered with no bioturbation present.
Interpretation: This association is best interpreted as the deposits of
weak, waning sheetfloods in a low energy, mud-prone, ice-channel distal
or by-passed environment.
Distributary Channel (PSB-dCH): forming dm-scale, fining-up­
wards or aggradational packages (<10 m thick), typically capping Facies
Associations pMB & dMB. Facies comprise clean (<5% Tmud), medium
to coarse grained, moderately poor to well sorted sandstones charac­
terised by parallel lamination, and small to medium-scale cross-bedding
(Fig. 8-I, ab, ac). Rare bioturbation, Skolithos tube type (Skolithos ich­
nofacies) also occurs locally.
Proximal Mouthbar deposits (PSB-pMB): typically form aggrada­
tional to coarsening upwards packages <7 m thick, gradational from
PSB-dMB. Facies packages comprise clean to moderately argillaceous
(0–20% Tmud), fine to medium grained, poor to moderately sorted
sandstones and sandy heterolithics. Sedimentary structures include flat
to low angle parallel lamination (Fig. 8, f and o), current, wave,
combined-flow ripples, possible Swaley Cross Stratification (SCS),
Hummocky Cross Stratification (HCS), and minor dewatering or soft
sediment deformation structures. Locally there is also a common but low
diversity ichnofauna including Zoophycus and indeterminate horizontal
burrows.
Distal Mouthbar deposits (PSB-dMB): comprise m-scale (<5 m
thick), aggradational to coarsening-upwards packages typically grada­
tional into PSB-pMB. The Association is characterised by very fine to fine
grained, poorly to moderately sorted, slightly to highly argillaceous
sandstones and sandy to muddy heterolithics (10–60% Tmud). Common
F.J. Bataller et al.
Fig. 10. Summary of key characteristics of the component facies associations comprising GDE IPD or Inactive Pro-Glacial Delta Front and the volumetrically
abundant FA BD-Hem. Figure shows simplified descriptions and GR curves for each association. Also shown are a selection of key images from core, FMI (see Fig. 5
for the well and depth of each image).
sedimentary structures include parallel lamination, ripple cross lami­
nation and widespread evidence of dewatering and soft sediment
deformation features (loads, flames, convolute lamination and slumps;
Annex VIA-3). Locally rare examples of Zoophycus are observed.
Interpretation: These three associations can be considered as a single
linked depositional system with an aggradational to coarsening-upwards
stacking pattern reflecting the localised progradation of pro-glacial
fluvial systems into a relatively high gradient glaciomarine setting.
Sedimentation rates are likely to have been high to very high, especially
during warmer seasons and, on a longer time-scale, during the initial
stages of ice retreat. This is essentially a pro-glacial, fluvial-dominated
braid-delta front system. The characteristically high accumulation and
progradation rates are reflected, as demonstrated in key outcrops of the
SW Gargaf Arch, by the widespread development (>25 km2) across
entire braid-delta front system, of listric growth faulting and associated
rollovers, with seismic-scale amplitudes of up to 50 m and wavelengths
of several hundreds of metres.
In the most distal, most marine-influenced setting represented by
Association PSB-dMB (Distal Mouth Bars), sedimentation was domi­
nated by abundant, fine grained deposition from glacially-derived mud
11
Marine and Petroleum Geology 134 (2021) 105335
plumes and waning distal hyperpycnal flows. Rapid accumulation
generated locally oversteepened, unstable slopes and an abundance of
dewatering. With continued progradation, thicker and more abundant
sands were deposited, as Proximal Mouth Bars (PSB-pMB), mainly from
flood-generated hyperpycnal flows but also there is evidence of storm-
wave influence in the form of HCS, SCS and combined flow ripples,
reworking the rapidly deposited pro-glacial output. Subsequently mouth
bars graded into or, more commonly, were truncated by the sand-prone
deposits accumulating in Distributary Channels (PSB-dCH). The latter
association is dominated by tractional deposits formed by fluvial or
possibly fluvio-tidal-dominated processes at the distal termination of the
Sandur. The limited occurrence of bioturbation may reflect largely
hostile environmental conditions, principally the high rates of sedi­
mentation and turbidity. When this depositional system is compared
with the work published by Girard et al. (2012a), there are significant
similarities between their DP3 (Terminal Distributary Channel) facies
and Facies Association PSB-dCH. However, mouth bar deposits associ­
ated in both studies with “lower energy” facies (DP2, Thin Bedded to
Heterolithic Mouth Bar deposits in Girard et al. (2102a) and PSB-dMB in
this paper) differ in that, in this paper, although not necessarily
F.J. Bataller et al. Marine and Petroleum Geology 134 (2021) 105335

Table 2
General Depositional Environments, Facies Associations and Lithofacies summary table.
Facies Code Visual Domintant Sorting Dominant Bioturbation GR profile Mean
Association clay lithology and Sedimentary Thickness
content Grain Size Structures

Fluvial Channel PSB- 0–5% Medium to coarse Poorly to Trough cross- No Blocky Thin m-
fCH sand. moderate bedding, climbing dm- scale
mega ripples, (<10 m)
parallel lamination
and minor current
ripples. Also,
climbing
megaripples.
Presence of floating
granules, pebbles
and muddy
intraclasts. The base
of facies packages
may also be loaded,
with gutter casts or
marked by pebbly
lag deposits
Proglacial Distributary PSB- 0–5% Medium to coarse Moderate Parallel laminated rare (Skolithos Aggradational Thin m-
Sandur-Active Channel Deposits dCH sand poor to good and small to ichnofacies) or fining dm- scale
Braid Delta medium-scale cross- upwards (<10 m)
Front bedded
Proximal PSB- 0–5% Fine to medium Moderate to sharp to locally No Aggradational Thin m-
Sheetfloods DPps sand good erosive based ssts; or weakly dm- scale
often graded, planar fining upwards (<10 m)
laminated, small-
scale cross-bedded,
rippled, dewatering
structures. Possible
climbing
megaripples.
Distal Sheetfloods PSB- 5–30% Slighly to Moderate to thin beds, common No Aggradational Thin m-
DPds moderatly moderately planar and ripple or fining scale (<5
argillaceous, fine good cross-laminated, upwards m)
to fine sand dewatering
structures
Argillaceous PSB- 30–50% Moderatle to Moderate to thin beds, current No Aggradational Thin en-
Sandflat Deposits Dpas highly good rippled, flat (serrate) scale (<5
argillaceous, very laminated & locally m)
fine to fine sand dewatered
Proximal Mouth PSB- 0–20% Clean to Poor to flat to low angle low diversity Aggradational Thin (<7
Bar Deposits pMB moderately moderatly parallel lamination, ichnofauna or coarsening m)
argillaceous, fine good current, wave, including upwards
to medium sand combined-flow Zoophycus and
or hetherolithics ripples, possible indeterminate
SCS/HCS, minor soft horizontal
sediment burrows
deformation
Distal Mouth Bar PSB- 10-60% Very fine to fine, Poor to characterised by locally rare coarsening Thin (<5-
Deposits dMB slightly moderate parallel lamination, burrowed upwards or m)
argillaceous sand ripple cross (Zoophycus aggradationa
to muddy lamination, common ichnofacies)
heterolithics dewatering and soft
sediment
deformation (loads,
flames, loads,
convoilute
lamination &
slumps)
Mini or Gilbert- APD- 0–10% Clean to slightly Moderate to sharp but not no Weakly 3–20 m
Delta Deposits MD argillaceous very poor erosively-based with coarsening
coarse sand, megaforeset upwards
locally pebbly stratification
and dipping at >30◦
conglomeratic to
coarse.
Tidal Channel IPD- 0-20% Clean to moderate to parallel laminated, local, rare fining upwards Thin en-
Deposits TDch moderatly good small to medium- bioturbation or less scale (<5
argillaceous, fine scale cross-bedded (Skolithos commonly m)
to coarse sand ichnofacies aggradational
dominated by
vertical
(continued on next page)

12
F.J. Bataller et al. Marine and Petroleum Geology 134 (2021) 105335

Table 2 (continued )
Facies Code Visual Domintant Sorting Dominant Bioturbation GR profile Mean
Association clay lithology and Sedimentary Thickness
content Grain Size Structures

suspension-
feeding tubes)
Tidal Flat IPD- 5-60% Very fine to moderate to small-scale cross- restricted low Aggradational Thin en-
Deposits TDf medium, slightly good bedding, parallel, diversity or fining scale (<5
argillaceous sand wave and current ichnofauna upwards m)
to muddy ripple cross- including the
heterolithics. lamination, flat- crawling trace
topped ripples Didymaulichnus
and shrinkage cracks
Inactive Upper Shoreface IPD- 0–10% Clean, fine to moderate to local stringers of Bioturbation Aggradational Thin dm-
Proglacial Deposits uSF medium, to good coarser sand and occur locally, or coarsening scale
Delta Front slightly fine gravel, flat to most commonly at upwards (<10 m)
argillaceous sand low angle parallel the base at these
laminated packages
(including possible (Planolites,
SCS), small to Paleophycus,
medium-scale cross- Thalassinoides).
bedding and both
current and wave
ripple cross-
lamination
Lower Shoreface IPD- 10-50% Slightly to very Moderate to flat to low angle burrowing can be Aggradational Thin dm-
Deposits ISF argillaceous, very good parallel laminated present or coarsening scale
fine to fine sand (including possible (Planolites, upwards (<10 m)
and sandy HCS), combined Paleophycus,
hetherolithics flow ripple cross- Thalassinoides)
lamination
Shelf Lobe IPD- 0-20% clean to Moderate to flat lamination, low No fining upwards Thin (<5
Deposits SHl moderately good angle lamination m)
argillaceous. (probably HCS) and
Very fine to Fine, combined flow
locally medium ripple cross-
lamination. Sharp or
locally erosive base
Muddy Shelf IPD- 50–70% Muddy 000000 present. rare bioturbation Aggradational Thin (<5
Deposits SHm heterolithics and (serrate) m)
silty mudstones
with very thin to
thin interbedded,
very fine grained
sandstones
Condensed IPD- elevated density and intense Thin
Horizons (Lags) CH high density values bioturbation (<lm)
(log response), often
Fe-rich cement and/
or coated grains.
Mostly marking
major sequence
stratigraphic events
Active Proximal Sandy APD- 5–30% slightly to highly Poor to floating coarser sand rare bioturbation Aggradational dm-scale
Proglacial Slope pSD argillaceous very moderate grains and or coarsening (<10 m)
Delta Front fine to fine subordinate debrite upwards
sandstones and layers, abundant,
heterolithic floating mudclast,
sandstones. Bed bases are
typically loaded,
massive or
dewatered aspect,
include slumps,
slides, dewatering
pipes, flame
structures, injectites
and convolute
lamination, is often
associated with
metre-scale ball &
pillow structures.
Rare cleaner
sandstone packages
show parallel and
ripple cross-
lamination
Distal Sandy APD- 20–50% highly Poor to Beds often show locally burrowed Aggradational metre- to
Slope dSD argillaceous, moderate dewatering or weekly dm-scale
(continued on next page)

13
F.J. Bataller et al. Marine and Petroleum Geology 134 (2021) 105335

Table 2 (continued )
Facies Code Visual Domintant Sorting Dominant Bioturbation GR profile Mean
Association clay lithology and Sedimentary Thickness
content Grain Size Structures

mixed to muddy structures with soft coarsening (<5–10

heterolithics, sediment upwards m)
intercalated with deformation. Less
silty to with very common are parallel
fine to fine sand. and ripple cross-
lamination
associated with the
cleaner sands.
Proximal APD- 0–5% very fine to poor to with angular to no Aggradational dm-scale
Outburst Sheet pOSD medium and Moderate rounded extraclasts or weekly (<20 m)
locally pebbly and/or mudclasts. coarsening
Sharp base, erosive upwards
or loaded,
sometimes on a
metre-scale
(“Megaloads”). low
and high angle cross
stratification, and
climbing dune cross
stratification
(CDCS). often large-
scale dewatering
structures including
loads, flames, load
balls, convolute
lamination and dish
& pillar structures
Distal Outburst APD- 5–40% very fine to fine Moderate locally displaying no Aggradational dm-scale
Sheet dOSD grained highly deformed or fining (<10 m)
sandstones and mudclasts, sharp, upwards
sandy non-erosive, bases,
heterolithics planar lamination,
less common ripple
lamination, and
common dewatering
structures
Glaciomarine Fringe Deposits BD- 20–70% Fine to very fine, Poor to good local dewatering minor Aggradational metre-
Basin Deposits FD moderatly structures, minor bioturbation or weekly scale (2-
argillaceous soft sediment coarsening <10 m)
sand, to muddy deformation or more upwards
heterolithics commonly planar
and/or ripple cross-
lamination
associated with
sharp, non-erosive
or loaded bases
Hemipelagic BD- 50–90% Muddy sharp or loaded massive Thin (<2
Basinal Deposits Hem heterolithics and bases, graded, m)
silty mudstones. parallel laminated
and rippled with
minor dewatering
structures, often
show scattered,
floating coarser
grains and
mudclasts, ranging
from sand to pebble-
size
Glaciotectonics Glaciotectonised GTR- Intraformational Upper Ordovician sediments associated with key surfaces or Middle to Lower Ordovician substrates, below the
and Sediments gTEC MGES, characterised by intense fracturing (“Step fractures’), injection strcutures, tilting and small to medium-scale folding
Remobilizatio
Slides & Mass GTR- Heavily faulted, dm to hundred metre-scale rotational blocks or olistostromes of pre-glacial Cambro-Ordovician substrate locally
Transport MTD mixed with sheared & folded packages derived rom basal Upper Ordovician deposits. Typically occur on paleohigh margins
Complexes

attributed to inter-outburst stages, the facies are in both cases described
as “flood-dominated mouth bar deposits”.
Mini- or Gilbert delta deposits (PSB-MD): are highly localised,
ranging between 3 and 20 m (McDougall et al., 2006; McDougall and
Gruenwald, 2011) in thicknesss. They form distinctive, sharp but not
erosively-based, coarsening-upwards packages characterised by steep
(30–32◦ ), uniformly-dipping, cm-scale foresets (Fig. 8, w).
Lithologically, APD-MD comprises clean to slightly argillaceous (0–10%
Tmud), coarse grained, moderately to well, even bimodally sorted,
pebbly and conglomeratic sandstones.
Interpretation: The deposits occur in only one of our study wells but
are slightly more common in nearby subsurface concessions (McDougall
et al., 2006; McDougall and Gruenwald, 2011) and locally in outcrops.
In all cases, the facies can be interpreted as the deposits of a mini- or

14
F.J. Bataller et al.
Gilbert delta recording high influxes of coarse grained sediment into a
standing body of water in an ice-proximal setting dominated by hyper­
pycnal flows; perhaps glaciomarine embayments or, in some cases,
glacial lakes.
4.1.2. Active pro-glacial delta front (APD)
This GDE comprises 4 component facies associations which are,
volumetrically, the most significant elements of the Upper Ordovician
succession in the study wells. Mostly sand-dominated, these facies were
deposited under fully glaciomarine conditions strongly influenced by
the high to very high sediment discharge from a rapidly retreating ice
margin. The facies from part of single linked depositional setting asso­
ciated with a steep (5–10◦ dip), rapidly prograding or aggrading, high
energy, active proglacial delta front setting (APD).
Within this framework, facies associations vary from proximal sub­
marine channels and associated overbank lobe deposits (APD-pSCH) to
poorly confined submarine channels (APD-dSCH) transporting signifi­
cant volumes of sand into a deeper water glaciomarine setting; both
forming perhaps the volumetrically most significant facies elements in
the study wells. Laterally, moving away from these primary clean sand
conduits the remainder of the braid-delta front comprised largely by-
passed proximal Sandy (APD-pSL) and distal Muddy (APD-dSL) Slope
deposits (Table 2). From subsurface data alone there is little we can say
about likely geometries and dimensions of these common facies. How­
ever, from abundant outcrop data, in SW Gargaf, Wadi Analelin
(Tihemboka Arch) and the Tassili N Ajjers of SE Algeria; notably the
well-studied Gara Ouarsasine and Iherir outcrops (e.g. Deschamps et al.,
2013) the component associations are intercalated in weakly
coarsening-upwards sequences. Such packages of APD may be up to or
even locally well in excess of 100 m thick with a downlapping (foresets
and toesets), clinoformal aspect.
Proximal Outburst Sheet deposits (APD-pOSD): are characterised
by dm-scale (<20 m thick), weakly coarsening-upwards or aggrada­
tional facies packages comprising, thick to very thickly bedded, very fine
to medium grained, mostly clean (Tmud <5%), poor to moderately
sorted sandstones, locally pebbly with angular to rounded extraclasts
and/or mudclasts. Package bases are sharp, erosive or loaded, some­
times on a metre-scale (“Megaloads”). Internally, beds in the subsurface
are principally characterised by: (a) sets of both low and high angle cross
stratification interpreted, by analogy with outcrops (Annex VIB-1,2,4
and 6), as possible climbing dune cross stratification (CDCS) identical
in almost all respects to those described for Facies Association PSB-fCH
(Ghienne et al., 2010) and/or (b) extensive, often large-scale dewatering
structures (Fig. 9, a, A) including loads, flames, load balls, convolute
lamination and dish & pillar structures.
A synthesis of both subsurface and outcrop data suggests that ge­
ometries, are broadly lenticular, packages mostly comprising stacked,
high width/depth (W/D) ratio multistorey/multilateral channel bodies.
Outcrops and seismic stratigraphy suggest these bodies tend to amal­
gamate forming areally extensive sheet to lobe-like complexes with a
low angle basinwards dip of (<5–10◦ ) associated with the upper and
middle foresets of clinoformal complexes. Less common, are isolated,
dm-scale, lower W/D ratio channel bodies, encased within slope de­
posits of Associations APD-pSD and dSD.
Distal Outburst Sheet Deposits (APD-dOSD): these occur as dm-
scale (<10 m thick), fining upwards or aggradational facies packages
comprising slightly argillaceous, very fine to fine grained, moderately
sorted sandstones and sandy heterolithics (5< Tmud<40%), locally
displaying highly deformed mudclasts (Fig. 9, iiib). Bedding is typically
medium to thick (mostly 20–50 cm) with sharp, non-erosive, bases,
planar lamination, less common ripple lamination, and common dew­
atering structures (Fig. 9, i). Structures interpreted as possible climbing
dune cross stratification are notably absent.
Interpretation: Both associations pOSD and dOSD are interpreted as
the deposits of both high and low density turbidity currents, probable
hyperpycnal flows (an assemblage of depositional processes we could
15
Marine and Petroleum Geology 134 (2021) 105335
informally group under the term “Glacioturbidites”) and, as suggested
by often abundant low to high angle cross-stratification interpreted as
CDCS, powerful, sediment-laden tractional flows characterised by high
rates of fall-out from suspension (Ghienne et al., 2013).
Facies and sequence stratigraphic context, together with observa­
tions from analogous outcrops, suggest these deposits can best be
interpreted by reference to both a continuum of depositional processes,
as outlined above, and closely-linked glaciomarine environments. As­
sociation pOSD represents one end-member, reflecting deposition from
glacial outburst events or Jokhulhaups on a ramp-like braid-delta or
braidplain delta front. Frequent multiple high energy events spread
across the braid-delta front, and sourced by rapid melting at the ice
front, generated the distinctive sheet- or lobe-like bodies characterised
internally by a multistorey/multilateral, high W/D channel system
architecture.
Less common, isolated channel bodies encased within slope deposits
are, in contrast, interpreted to reflect lower rates of sediment discharge,
localised to specific entry points, such a distributary channels and mouth
bars on the delta front. Most were subsequently abandoned whilst others
may represent an initial phase of growth in some progradational sys­
tems, with the isolated channels giving way to sheet- or lobe-like bodies
as both sand and supply and outburst event frequency increased with ice
melting.
In either of the end-member cases described above, Association
dOSD is interpreted as the distal expression of these deposits, reflecting
deposition downslope or laterally from the powerful waning sediment-
laden flows. As such, deposition most likely occurred from low density
turbidity currents and the facies association might be considered as
transitional into either slope deposits laterally or lobe fringe deposits in
a downslope direction.
Proximal Sandy Slope Deposits (APD-pSD): occur as dm-scale
(<10 m thick), often coarsening up in outcrops passing from dSD and
into OSD (also laterally) or aggradational packages. Comprising slightly
to highly argillaceous (5–30% Tmud) very fine to fine, locally medium
grained, poor to moderately sorted sandstones and heterolithic sand­
stones. These may also show floating coarser sand grains and are locally
intercalated with subordinate debrite layers (Fig. 9, J and D) charac­
terised by abundant, floating mudclasts and fine gravel in an argilla­
ceous sand or sandy mudstone matrix. Less common are thin packages of
cleaner sandstones. Bed bases are typically loaded, often, as observed in
analogous outcrops (Annex VIB-3), on a metre-scale. Internally, grading
is absent or weak and bedding mostly medium to thick or locally, poorly
defined. The latter is a direct reflection of a characteristically massive or
dewatered aspect (Fig. 9, E). Sedimentary structures include slumps,
slides, dewatering pipes, flame structures, injectites and convolute
lamination which, from observations of analogous outcrop sections, is
often associated with metre-scale ball & pillow structures. Rare cleaner
sandstone packages show parallel and ripple cross-lamination. Local
bioturbation is observed in the more argillaceous intervals (Fig. 9, iiia)
but is generally rare and of low diversity.
Distal Sandy Slope Deposits (APD-dSD): occur as m to dm-scale
(<5–10 m thick) packages which are often very poorly bedded or
unbedded, depending on the intensity of soft sediment disruption, ag­
gradational or weakly coarsening upwards, highly argillaceous (20-
>50% Tmud), typically gradational into Association pSD and transi­
tional in character between these and the basinal hemipelagic deposits
of Association BD-Hem. Packages are characterised by mixed to muddy
heterolithics (Fig. 9, af) comprising subordinate very fine to fine
grained, poor to moderately sorted, ungraded or weakly graded sand­
stones intercalated with silty to sandy mudstones. Beds often show
dewatering structures (Annex VIB-5) with soft sediment deformation
similar to that of Association pSD. Less common are parallel and ripple
cross-lamination associated with the cleaner sands. Local burrowing
might be present (Fig. 9,B).
Interpretation: The deposits of Facies Associations APd-pSD and -dSD
are interpreted to have accumulated by deposition from a spectrum of
F.J. Bataller et al.
sediment gravity flows (mostly low density turbidity currents and debris
flows) and hyperpycnal flows in a submarine slope setting. This is
generally associated with two major environmental complexes:
(a) the upper to lower foresets of the clinoformal bodies generated by
the progradation of proglacial braid-deltas or braidplain deltas.
Within this framework, the close relationship with the sands of
Associations APD-OSD and dOSD, observed in both subsurface
and outcrop, is a key element of the GDE in which the slope de­
posits, typically grading upwards into the sheet-like outburst
deposits, form stacked coarsening-/cleaning-upwards sequences
forming the lower part of individual clinothems. In each such
package, the deposits of Associations pSD and dSD correspond to
a period of normal to slowly increasing sediment input. This is
characterised by mass flows remobilizing sediment arriving at the
delta front on to an initially unstable, steeper gradient slope,
generating the distinctively pervasive dewatering and soft sedi­
ment deformation. Abrupt and major increases in sand supply
due to outburst events resulted in a rapid transition to Associa­
tions OSD and pOSD. These are typically marked by a significant
shift in lithology and large-scale loading, a response to the huge
volumes of sand entering the delta front during a short timeframe
and which may also tend to reduce slope gradients. Subsequently
the outburst complex is abandoned, and deposition resumes in
the lower energy, background slope setting of Associations pSD
and dSD (Annex III-A,B). Facies relationships also suggest that
slope deposits will have continued to accumulate in by-passed
areas of the delta front and are thus lateral equivalents and
gradational along depositional strike into the sand-prone
outburst complexes.
(b) deposition as locally very thick (>100 m) often mud-prone
packages interpreted as moraines or outwash fan complexes
associated with a marine-terminating, tidewater glacier. In such a
context, Associations pSD and dSD are likely to have been
deposited rapidly from mass flows and hyperpycnal flows exiting
directly from the grounding line, grading both downslope and
laterally into the hemipelagic basinal muds of Association BD-
Hem. Much of the lithostratigraphic units, known as the Melaz
Shuqran Formation and, to a lesser extent, the lowermost Middle
Mamuniyat and Bir Tlacsin are likely to have formed in this way.
Accordingly, deposits of this environment are volumetrically
significant at several levels of the Upper Ordovician succession in
different parts of the study area, most likely associated with
mounded chaotic seismic facies and the potential to act as
intraformational seals.
4.1.3. Inactive Proglacial Delta Front (IPD)
This GDE comprises 7 component facies associations, all volumetri­
cally less important in the study wells than those of GDEs APD and PSB
(but locally important elsewhere in outcrops and subsurface). Although
the GDE is mostly sand-prone, some facies are mud-prone in character.
Depositional environments include Tidal Channels (IPD-TDc) and Tidal
Flats (IPD-TDf), Shoreface deposits, both Upper (IPD-uSF) and Lower
(IPD-lSF), as well as storm or flood-generated related shelfal lobes (IPD-
SHl), Muddy Shelf below storm wave base (IPD-SHm) and the volu­
metrically insignificant but often stratigraphically important Condensed
Horizons (IPD-CH) (Table 2). In broad terms, all of these facies record
deposition during periods when direct glacially-derived sediment supply
was significantly reduced, either because of the shifting and abandon­
ment of the active pro-glacial sediment supply or as a response to late-
stage glacial retreat and a now distant ice margin.
Tidal Channel deposits (IPD-TDch): typically form relatively thin
(<5 m thick), fining-upwards (or less commonly aggradational) pack­
ages mostly capping and/or intercalated with Facies Tdf and Usf. The
Association comprises clean to moderately argillaceous (0–20% Tmud)
fine to coarse grained, moderately to well sorted sandstones. Typical
16
Marine and Petroleum Geology 134 (2021) 105335
sedimentary structures are parallel lamination, small to medium-scale
cross-bedding with local, rare bioturbation (Skolithos ichnofacies
dominated by vertical suspension-feeding tubes).
Tidal flat deposits (IPD-TDf): occur as thin (<5 m thick), fining
upwards or aggradational sequences intercalated with a range of
different facies associations. Facies packages range from very fine to
medium grained, moderately to well sorted, slightly argillaceous sand­
stones to common sandy and muddy heterolithics (5–60% Tmud).
Sedimentary structures, observed both in cores and outcrops (Annex
I–B), include small-scale cross-bedding, parallel lamination, wave and
current ripple cross-lamination, flat-topped ripples (Annex VIC-6),
shrinkage cracks and locally, restricted low diversity ichnofauna
including the crawling trace Didymaulichnus.
Interpretation: neither Association TDch or TDf are volumetrically
significant in the study wells but do occur locally both in the well sec­
tions of other Murzuq and Illizi Basin concessions and in outcrops
(Annex I-A,C). The former (TDch) is best interpreted, most easily in
terms of facies context, as the deposits of high to medium energy tidal
channels on barred shorelines, forming on the inactive sections of the
braid-delta front. In contrast, the notably more argillaceous, often het­
erolithic character and a distinctive suite of sedimentary structures of
Association, together with facies context, strongly suggests Association
TDf is best interpreted as very shallow water deposits corresponding to
sandy to muddy tidal flats most likely occurring in lagoons, abandoned
segments of non-barred shoreline or even in the distal reaches of dis­
tributary and fluvio-glacial channels.
Upper Shoreface Deposits (IPD-uSF): occurs mostly as dm-scale
(<10 m thick), coarsening-upwards (or aggradational) facies packages
often gradational from Association lSF. They comprise, clean to slightly
argillaceous (0–10% Tmud), fine to medium grained, moderately to well
sorted sandstones with locally common stringers of coarser sand
(McDougall and Gruenwald, 2011) and fine gravel. Common sedimen­
tary structures include flat (Fig. 10, G and Annex II) to low angle
(Fig. 10, K) parallel lamination (including possible SCS) as seen also in
outcrop (Annex VIC-1,2 and 5), small to medium-scale cross-bedding
(Annex VIC-3) and both current and wave ripple (Fig. 10, m)
cross-lamination. Bioturbation also occurs, most commonly Planolites,
Paleophycus and Thalassinoides near the base, and Skolithos in the upper
part of packages.
Lower Shoreface Deposits (IPD-lSF): also form dm-scale (<10 m
thick), coarsening-upwards (or aggradational) packages typically
gradational into Association uSF. They are characterised by slightly to
very argillaceous (10–50% Tmud), very fine to fine grained, moderately
to well sorted, sandstone and sandy heterolithics. Common sedimentary
structures include flat (Fig. 10, C) to low angle parallel lamination
(including possible HCS as seen in Annex III-B), combined flow ripple
cross-lamination and, locally, burrowing also occurs, dominated by
Planolites, Paleophycus and Thalassinoides (Fig. 10, al).
Interpretation: Associations uSF and lSF form part of a single coars­
ening-/cleaning-upwards depositional system which, given the assem­
blage of sedimentary structures, burrows and overall facies context
suggests they are best interpreted as the deposits of respectively, storm-
wave-dominated Upper and Lower Shorefaces. The former most likely
reflects the role of storms and fair-weather rip and longshore processes
whilst the latter is a markedly lower energy setting dominated by storm
and fair-weather wave processes. In both cases sediment supply from the
pro-glacial braidplain was significantly reduced, either by major ice
retreat or local abandonment, reworking and deposition dominated by
basinal processes; waves, storms and tides. The role of the latter seems to
have been relatively minor, in the study area. Elsewhere, however, in
nearby subsurface concessions and outcrops of the Ghat area there is
significant evidence for a lateral gradation from uSF to tidally-
influenced sandsheets (McDougall and Gruenwald, 2011) possibly
associated with estuaries or tide-dominated deltas.
Shelf Lobe deposits (IPD-SHl): these typically occur as thin (<5 m
thick) fining-upwards packages comprising clean to moderately
F.J. Bataller et al.
argillaceous (0–20% Tmud), very fine to fine, locally medium grained,
moderately to well sorted, thin to medium bedded, sharp or locally
erosively-based (Fig. 10, d). sandstones and sandy heterolithics, Sedi­
mentary structures include flat lamination (Fig. 10, s), low angle parallel
lamination (probably HCS) and combined flow ripple cross-lamination.
Muddy Shelf deposits (IPD-SHm): are defined as thin (<5 m thick)
aggradational packages, with a typically serrate GR profile, comprising
muddy heterolithics and silty mudstones, with very thin to thin inter­
bedded, very fine grained sandstones, and silty mudstones (50–70%
Tmud). Typical structures include parallel lamination and combined
flow ripple cross-lamination. Locally there may also be rare convolute
lamination (Fig. 10, b). Locally, rare bioturbation may also occur.
Interpretation: These facies are best interpreted as the end-members
of the GDE IPD depositional continuum, often gradational into Associ­
ations lSF and uSF. This is especially true of Association SHm, inter­
preted as the product of very low energy, slowly deposited background
muds on an inner shelf (below fair weather wave base or FWWB but
above storm wave base) swept by distal, waning storm events. In
contrast, Association SHl, is interpreted as packages of largely clean
sands delivered from the shoreline by moderate to high energy storm
events and deposited below FWWB within a shelfal setting. It is likely
that the sands form lobe-like bodies fed through offshore extensions of
rip and tidal channels. In neither case is there any evidence of direct
glaciomarine influence.
Condensed horizons (Lags) (IPD-CH): often stratigraphically sig­
nificant, thin (<1 m) packages occurring within mud-prone successions
or capping coarsening-upwards parasequences. They are marked in
conventional logs by elevated density values, or very high resistivities in
FMI images which, in outcrop or, rarely in cores, often correspond to Fe-
rich cemented lags, locally with abundant coated grains and/or intense
bioturbation.
Interpretation: these thin, but distinctive horizons are rare but of
stratigraphic importance. They undoubtedly record periods of very low
sediment supply in marine environments. As such they likely represent
phases of maximum glacial retreat and occur at several levels of the
Upper Ordovician succession, perhaps most notably, but not exclusively,
at the boundary between the glaciogenic package and the overlying
Silurian age Tanezzuft Formation.
4.1.4. Glaciomarine basin deposits (BD)
This GDE comprises two component facies associations. Locally,
these are volumetrically important, with intraformational seal potential,
but are not generally significant throughout the Upper Ordovician suc­
cession as a whole. Facies are mud-prone, depositional environments
comprising Fringe deposits (BD-FD) and Hemipelagic Basinal Deposits
(BD-Hem), Together they form the distal end of the glaciomarine
depositional continuum, most commonly gradational and intercalated
both with each other (Fig. 10q) and with the sediments of GDE APD.
Fringe Deposits (BD-FD): occur as m-scale (2-<10 m) aggradational
or weekly coarsening-upwards packages. These comprise thin to me­
dium bedded, very fine to fine grained, silty and moderately argilla­
ceous, poor to well sorted sandstones or more commonly, sandy to
muddy heterolithics (20–70% Tmud). Beds tend to show grading, local
dewatering structures, minor soft sediment deformation, or more
commonly planar and/or ripple cross-lamination associated with sharp,
non-erosive or loaded bases (Fig. 10, q). There is also some suggestion
that minor bioturbation may occur locally. These facies are the down­
slope termination of the Outburst sheets (APD-pOSD and dOSD) and
Lobe deposits (IPD-SHl).
Hemipelagic Basinal Deposits (BD-Hem): in subsurface these
occur as thin (<2 m thick) aggradational muddy packages characterised
by a blocky or serrate high GR response. These comprise both muddy
heterolithics and massive to poorly bedded sandy to silty mudstones
(50–90% Tmud). Thin sands and silty sands are very fine grained and
characterised by sharp or loaded bases, grading, parallel lamination,
ripple cross-lamination and minor dewatering structures. The dominant
17
Marine and Petroleum Geology 134 (2021) 105335
mudstones often show locally abundant floating coarse sand grains,
granule to pebble-size extraclasts (Annex VI C-4) and also intraforma­
tional, angular-subangular, often deformed mudclasts. These facies are
gradational to and often intercalated with the Muddy, and less
commonly Sandy Slope Deposits of Associations APD-dSD and pSD.
Interpretation: these facies record deposition from a range of low
energy processes ranging from waning low density turbidity currents or
hyperpycnal flows to dilute debris flows and hemipelagic suspension,
possibly supplemented locally by direct iceberg rain-out (“dropstones”).
They correspond to the distal bottom or toesets of the braid-delta front
clinoforms previously discussed with the APD depositional system.
Alternatively, they may be associated with the distal zones of moraines
(most notably in “Melaz Shuqran”-type packages) forming at a sub­
aqueous ice front where sediment rates would locally be high, in
contrast to the lower sediment supply rates associated with an active
braid-delta front disconnected from a direct glacial source due to initial
ice retreat. Fig. 10q shows an example, from FMI, of thinly interbedded
Hemipelagic and Fringe deposits.
4.1.5. Glaciotectonics and remobilization (GTR)
This GDE comprises two component facies associations although
neither represent any specific depositional environment. Rather they
should be considered as a distinctive assemblage of structures super­
imposed on both pre-glacial bedrock or other intraformational facies
associations. Within this framework we recognize Glaciotectonised
Sediments (GTR-gTEC), associated with ice advance and Slides and Mass
Transport Complexes (GTR-MTD) related to the remobilization of
consolidated or partly consolidated sediments as a response to glacial
erosion and/or unloading.
Glaciotectonised Sediments (GTR-gTEC): these are observed at,
and immediately below key horizons, ranging from the Main Glacial
Erosion Surface (MGS) truncating the Hawaz bedrock to several intra­
formational levels, within the Upper Ordovician as seen in Annex III.
The features observed range from NNW-SSE oriented striations (only
visible in outcrop, as for example, those published by Girard et al.
(2015) from outcrops found to the SW of the study area) and other ice
erosional features such as roche moutoneé (Beuf at el, 1971), wet
sediment striae, abundant cm-scale step fracturing and dewatering
structures which often occur for up to several metres below these sur­
faces. Sand injection structures are also observed on the flanks of pale­
ovalleys, presumably a response to glacial-induced overpressures. Most
striking though are thicker sections (5–10 m), below the ice erosional
surfaces, dominated by tilting and associated small to medium or locally
large-scale folding and even thrusting, generated in response to the
“bulldozer” effect of advancing ice on poorly to partly consolidated
surfaces.
Slides and Mass Transport Complexes (GTR-MTD): these are
defined as dm to hundred metre-scale packages, observed both in
seismic and outcrop, characterised by pervasive listric faulting, rota­
tional blocks or landslips and/or olistostromes of pre-glacial Cambro-
Ordovician substrate locally mixed with sheared & folded packages
derived from basal Upper Ordovician deposits. These most typically
occur on the margins of Hawaz paleohighs but also locally, associated
with intra-Upper Ordovician paleohighs. They most likely represent the
effects of slope instability and elevated pore pressures on the partly
consolidated sediments forming the margins of newly excavated pale­
ovalleys and are likely gradational into the Slope deposits of Active
Braid-Delta Front or moraines associated with subglacial to proglacial
processes.
4.2. Depositional models
Once the large-scale depositional complexes or GDEs and component
facies associations were defined, several conceptual models were built.
These were constructed with the aim of both capturing and summarizing
the range of potential spatial and stratigraphic variability in ice margin
F.J. Bataller et al.
position, sedimentation rates, accommodation space and the distribu­
tion of depositional environments (GDEs and facies associations) during
a typical cycle of ice advance and retreat.
Four scenarios are depicted, as shown in Fig. 11: (model A) sub-
glacial erosion, proglacial glaciotectonism and only limited deposition
(IPD), corresponding to a phase of glacial advance or ice front stabili­
zation, (model B) depositional systems characterised by increasing
sediment supply, evolving from a glaciomarine (model B*) setting after
initial ice retreat and general flooding, to a proglacial fluvial setting
(Girard et al., 2015). reflecting initial glacial retreat ultimately domi­
nated by fluvial channel, sheetflood and submarine density flow pro­
cesses (PSB & APD), (model C) depositional systems characterised by the
highest sediment supply rate, reflecting rapid ice retreat, possibly ice
sheet collapse (PSB & APD, both outburst-dominated) infilling accom­
modation space left by the immediate post-glacial/ice retreat sea level
rise (but quickly reduced due to glacio-isostatic rebound) and (model D)
low sediment supply systems (IPD), corresponding to the later stages of
ice retreat, mainly dominated by marine processes reworking the pro­
glacial sandur.
Each of these scenarios or models represents an end-member of a
depositional continuum and may be applied as a predictive guide to
interpreting different glaciogenic successions between and within well
sections, although it is important to keep in mind the significant lateral
variability typical along ice margins reflecting the coexistence of active
and inactive glacially-fed delta systems (Kurjański et al., 2021):
In summary, the models introduced here represent different patterns
of sedimentation during, respectively, the late advance/ice front stabi­
lization phase, the early glacial retreat, the stage of maximum sediment
supply and the late glacial retreat phase. Each model has different im­
plications for the facies successions encountered in wells and offers a
framework for improved interpretation.
4.3. Seismic stratigraphic analysis
In this section we present two key seismic lines from the NE and SW,
respectively (Figs. 12 and 13), of the study area. They are considered
representative of the Upper Ordovician architecture, as seen throughout
the area, and are interpreted in detail from the perspective of a seismic
stratigraphic framework. The aim is to link the depositional models
presented above, generated on the basis of both facies interpretation and
analysis of stacking patterns in the study wells, with the “ground truth”
of the seismic data thereby establishing a consistent 3D framework for
predictive interpretation of the Upper Ordovician reservoir.
However, whilst this approach has obvious and significant benefits
for interpreting the sedimentary architecture and reservoir potential of
the area it is important to note some of the limitations on this method­
ology in order to fully appreciate the degree of confidence that may be
placed on the conclusions. The key issue is seismic vertical resolution;
this is a hard limitation and is approximately 20–25 m in the study area.
Reflectors are not, in most cases, beds and instead largely aggregate
packages of differing impedance although locally surfaces correspond­
ing to major lithological changes may be captured.
This is obviously significantly different to both outcrops and wells. In
both cases, cm-scale beds and structures can be recognised. As a
consequence, there is a significant impact on our ability, in some or
indeed many cases, to distinguish facies associations when this depends
on correct identification of key structures and stacking which cannot be
resolved at seismic level. Nonetheless whilst a this is a major limitation,
what we lose with resolution we gain with the continuous 3D coverage
of a significant volume of rock, which can be easily manipulated to
generate any number of inlines or crosslines and, most importantly,
seismic lines along or parallel to depositional dip. Also, when tied to the
well data the seismic becomes an exceptionally powerful tool. Effec­
tively wells become control points and when used to define both
reflector character and terminations, with the seismic stratigraphic
techniques outlined previously, we are able to distinguish numerous
18
Marine and Petroleum Geology 134 (2021) 105335
glacial packages, define their geometries, lateral extent and construct a
depositional history.
The key architectural elements or packages of the Upper Ordovician
succession in the first seismic line (Fig. 12) are summarized, within the
context of Fig. 11, as follows:
Seismic Package A: overlies the main glacial unconformity trun­
cating the Hawaz Formation (Fig. 12-av) and is interpreted to record the
development of a single facies belt. Based on the well section, this
package is interpreted to comprise the following sub-packages: (a) de­
posits of Facies Associations IPD (Figs. 12–1,2) and (b) APD (Figs. 12–3),
overlain by (c) deposits of Association IPD (Figs. 12–4). In addition, the
basal element also shows potential evidence for glaciotectonism (frac­
tures in Fig. 12-as) whereas the youngest sub-package does not (Fig. 12-
at). In summary, Package A is interpreted to record low sediment supply
associated with an ice maximum followed by a significant increase in
sediment supply and, finally, reduced sediment supply, reflecting the
glacial retreat captured by this seismic package.
Seismic Package B: truncates Package A along a major unconfor­
mity, penetrated by Well H (Fig. 12-at), and forming an erosive, intra­
formational, paleohigh. The flanks of this erosive feature appears to
show chaotic seismic facies (onto which reflectors from younger pack­
ages onlap). These sub-packages (Figure 12–5 and 6) are undrilled but,
given the chaotic reflector pattern, are interpreted as muddy, slumped
and possibly glaciotectonised sediments. Onlapping these sub-packages,
we see probable high energy deposits (APD Figs. 12–7) capped by
possible IPD-type sediments (Figs. 12–8,9). Towards the SE, these IPD
deposits also appear to show intra-formational faults (Figs. 12–8). These
faults appear to be reverse in character and could be interpreted as
glaciotectonic in origin, developed perpendicular to the direction of ice
movement proposed by Bataller et al. (2019). Significantly, this un­
conformity appears to become conformable towards the NW or
basinwards.
Seismic Package C: is undrilled by Well H but clearly onlaps the
unconformity truncating Package B. However, it is not unclear if the sub-
package located towards the SE (Figs. 12–11) is related to that located
towards the NW (Fig. 12–), as the continuity of the reflectors is
frequently compromised by faults, incisions and paleohighs as seen in
Annexes IV and V. It is interpreted to comprise deposits of Associations
APD (Figs. 12–13) and IPD (Fig. 12– and possibly 11). This sequence
corresponds to the development of the third “facies belt”.
Seismic Package D: is incised into the previous seismic package and
is absent across the crest of the paleohigh drilled by Well H. It is thought,
by comparison to other wells in the study area, notably Well D (Figs. 5
and 13), to comprise deposits of Association BD (Figs. 12–14) and is
interpreted as the distal expression of a “facies” belt (fourth) becoming
more proximal towards the E-SE, away from the study area.
The second seismic line (Fig. 13), is perpendicular in orientation (dip
parallel) to that shown in Fig. 12, and largely corresponds to the infill of
a U-shaped valley. The key architectural elements or packages of the
Upper Ordovician succession are summarized as follows utilizing a
different sequence of letters than that used in Fig. 12 in order to avoid
any implication of direct correlation:
Seismic Package V: (Figs. 13–1,2), occurs in both lines A and B
(Fig. 13). Only partly drilled by Well D, it is composed of chaotic re­
flectors infilling the initial paleorelief associated with the main glacial
erosion surface (Figs. 13–1) and a wedge-shaped body, which shows
downlap to both SE and NW off a small Hawaz paleohigh (Figs. 13–2). It
is interpreted as an assemblage of series of mud-prone deposits (Facies
Associations APD-dSD & BD-Hem) and mass transport complexes (Facies
Association GTR), probably glaciotectonised and deeply eroded by
subsequent ice advances and retreats.
Seismic Package W: is undrilled, restricted to the NW of the line due
to truncation and downlapping onto a major unconformity truncating
Package V. It comprises two sub-packages: (1) a thick, basal unit,
possibly comprising the deposits of Association APD (Figs. 13–3) and a
notably thinner unit (Figs. 13–4) possibly comprising Association IPD,
F.J. Bataller et al.
although this interpretation is not based on well data.
Seismic Package X: is partly drilled by Well D but confined to the SE
of the line. Three sub-packages are identified: (1) a thin unit charac­
terised by chaotic reflectors (Figs. 13–5, undrilled but interpreted as
mud-prone, mainly comprising slumps and debrites; (2) a thick package,
drilled by Well D and characterised by well-developed horizontal re­
flectors; interpreted as deposits of Association APD (Figs. 13–6). This is
absent towards the NW and truncated by the third sub-package (3). This
is a wedge-shaped unit. restricted to the SE (Figs. 13–7), forming a cli­
noform. Downlapping and thinning towards the NW. The interval is also
drilled by Well D and is interpreted as IPD, probably with deposits of
Association PSB developed towards the SE and muddy, basinal BD facies
towards the NW, infilling paleorelief associated with the truncation of
Seismic Package V.
Seismic Package Y: is also drilled by Well D and formed (from base
to top) by Facies Association APD (Figs. 13–8) followed by the deposits
of Association IPD (Figs. 13–9). It is characterised by parallel reflectors
onlapping paleorelief and pinching-out towards the SE, most probably
related to later erosion and/or non-deposition over the paleohigh
formed by Package X. In contrast, towards the NW, the thinning appears
to be purely depositional in character, a reflection of the clinoformal
geometries.
Seismic Package Z: appears to gently erode older packages or to
infill remnant paleorelief (Figs. 13–10). Drilled by Well H it is mud
prone, interpreted to largely comprise fringe and hemipelagic deposits
from Facies Association BD most probably capped by a thin condensed
horizon as suggested by a relatively low GR response immediately below
the Tanezzuft. It is best developed to the NW down depositional dip and
to the NE along strike. However, to the SE, it appears to pinch-out but
may well continue some way to then SE as a sub-seismic-scale package.
Within this framework, in the section shown in Fig. 12, the four
19
Marine and Petroleum Geology 134 (2021) 105335
(caption on next page)
seismic packages are interpreted to correspond to 4 key “cycles” or
“sediment/facies belts” (In Fig. 12, interpreted intervals or sub-packages
1,2,3 & 4 for Cycle 1; Cycle 2: sub-packages 5,6,7,8 & 9; Cycle 3: sub-
packages 11,12 & 13 and Cycle 4: sub-package 14 only). In Fig. 13 the
section shown comprises, 5 key “cycles” or “sediment/facies belts” (in
Fig. 13, Cycle 1: interpreted intervals or sub-packages 1 & 2; Cycle 2:
sub-packages 3 & 4; Cycle 3: sub-packages 5,6 & 7; Cycle 4: sub-
packages 8 & 9 and Cycle 5: sub-package 10 only), although, in both
cases, not all the cycles are complete due to intraformational erosion
and/or non-deposition. These packages suggest a common pattern of
low sediment input characterised by chaotic reflectors (IPD) associated
with an ice maximum, high sediment input (APD) with parallel re­
flectors onlapping or downlapping into previously existing surfaces
reflecting high sediment inputs during early ice retreat, and low sedi­
ment input again (IPD) distinguished by less chaotic seismic facies and
commonly-developed clinoforms (Figs. 4 and 11).
Correlation between the seismic packages interpreted in Figs. 12 and
13 is complex given the common occurrence of faults, nested incisions at
various scales, “chaotic” facies and paleohighs, both intraformational
and at Top Hawaz, as shown in Annexes IV and V. However, despite
these inherent difficulties we suggest that seismic packages C and D of
Fig. 12 are correlatable with Packages Y and Z, respectively, of Fig. 13.
Beyond these suggestions it is apparent that more extensive, block scale
seismic correlations and mapping are likely to be associated with sig­
nificant uncertainties. Nevertheless this approach could most definitely
be applied in order to improve interpretation and reduce risk: (a) locally
in areas of interest within blocks, such as potential prospects or even
discoveries/fields for infill exploration and production, or (b) used to
guide high resolution correlations between wells which can be used as a
in input for a 3D reservoir model.
Despite this practical limitation on the regional correlation of
F.J. Bataller et al. Marine and Petroleum Geology 134 (2021) 105335

Fig. 11. Simplified end-member depositional models summarizing the possible variations in ice position, sedimentation rate, accommodation space and the
consequent patterns in depositional environments (GDEs and facies associations).

(a) This model represents the probable patterns of sedimentation during a phase of ice front stabilization after a phase of advance and before any major retreat,
during which sedimentation rates are low, as most of the sediment is entrained within or at the base of the ice sheet. If the ice front has stabilized on land, then
pro-glacial sandur and delta front facies are likely to be narrow and poorly developed or with low preservation potential along with associated sub- or pro-
glacial tillites. If stabilization occurred in a marine-contact or tidewater setting then sandur facies will be absent (Fig. 11, model A*), reworked by glacial
erosion, and sub-glacial sediment will exit at the base of the floating ice front as dense efflux jets forming a sub-aqueous moraines or grounding line fan complex
(Deschamps et al., 2013). In the former case, diagnostic features would be the abundance of slope (dSD and pSD) and basinal (BD-Hem) facies, but the key are
the presence of erosive features and glaciotectonism (e.g. sheath folds, step fractures), abundant soft-sediment deformation, dropstones and a significantly
greater abundance of rotational slides and mass transport complexes. The common occurrence of abundant iceberg rainout dropstones and/or dilute pebbly
debris flows (MTCs) both appear to suggest a marine-terminating (“tidewater”) icefront, at least in the study area.
(b) This model represents the initial stage of ice retreat in the case of land-based glaciers, reflecting a significant early rise in sediment supply (after an initial
flooding due to early ice retreat and an initial relative delay in glacioisostatic rebound) delivered to a rapidly expanding proglacial braidplain or sandur and
braidplain delta front (GDE PSB), the whole forming a large-scale clinoformal body. Deposition on the sandur, or topset of the clinothems, mainly occurred in
braided channel systems and associated sheetfloods with occasional but increasingly more significant outburst events or Jokhulhaups (which may be associated
with the transition from inter-outburst to outburst stages as in the model proposed by Girard et al., 2012a, 2012b). At this stage the braidplain delta front was
still characterised by a relatively high gradient distinguished by rapidly prograding mouth bar complexes forming the uppermost section of the clinothem
foresets, grading laterally and downslope into slope deposits (APD-pSD & -dSD). With increasingly common outburst events, discrete but major channels began
to incise the slope deposits building locally significant, but areally restricted sand-prone channels and lobes, as the toe set of the clinoform (APD-pOSD and
dOSD) otherwise encased within the notably more argillaceous slope deposits.
(c) As ice sheet retreat continues to accelerate, sediment volumes increase significantly. The braidplain or sandur environment (GDE PSB) is similar in many
respects to model B but with a potentially marked increase in the relative volumetric importance of outburst or Jokhulhaup deposits. Such high sediment
volumes, which initially infilled the additional accommodation space left by ice retreat prior to the sea level drop associated with rebound resulted in the rapid
progradation and reduction in gradient of both the braidplain (topset) and braidplain delta front (foreset) systems. As such, fluvial-dominated mouth bars
typical of Model B are rare during this phase; instead, huge volumes of sand reach the delta front through rapidly migrating high W/D, multistorey/multilateral
channel complexes (Associations APD-pOSD and -dOSD). These amalgamate to form areally extensive sheets and lobes, dipping gently basinwards and grading
into the glaciomarine basin deposits (GDE BD) forming the toesets of the braidplain delta system. Between outburst-dominated phases of sedimentation and,
laterally, in by-passed zones, argillaceous sands and sandy mudstones accumulated from sediment flows remobilizing shallow water deposits to form slope
systems (Associations APD-pSD and -dSD).
(d) This model corresponds to the final part of the advance-retreat cycle when the ice margin has retreated to a significant distance from the shoreline and sediment
supply rates have begun to decline significantly. Relative sea level will also rise after dropping during post-glacial rebound (once this rebound has lost in­
tensity), generating accommodation space. However, the latter will have been reduced by the increasing effects of glacioisostatic rebound, generating addi­
tional sediment supply and the high sediment volumes characteristic of the previous stages. In contrast to the previous stage (model C), the bulk of the sediment
released from the melting ice (which will be in a relatively distant location, if any ice is remaining) will have been trapped in an increasingly extensive, low
gradient, proglacial sandur and braid-delta front. It is assumed that the sandur was dominated by braided streams and sheetfloods but periodically affected by
violent, episodic but probably less common, relative to Model C, glacial outburst (Jokhulhaups) events. In response to the marked reduction in meltwater-
derived sediment supply, the braidplain delta front was dominated by basinal processes. Sediments accumulated in distributary channels (Association PSB-
dCH), tidal channels (Association IPD-TDch), mixed fluvial/wave-dominated mouth bars (Associations PSB-pMB & -dMB), wave-dominated shorefaces (As­
sociations IPD-uSF & -lSF). The latter are also likely to have fed shelfal storm lobes (Association IPD-SHl) and possibly even tidal sand sheets or bars passing into
distal pro-delta shelfal muds (Association IPD-SHm), in what was effectively a non-glacial shallow marine setting characterised by storm, wave and tidal-
reworking of an inactive, subsiding braidplain.

packages it is nonetheless apparent that, from the foregoing examples, 5. Discussion

the degree of architectural complexity of the Late Ordovician glacial
succession is very significant and supports the conclusions of other au­
thors (Ghienne et al., 2003 and Moreau et al., 2015; amongst others). On
the basis of this key observation, and driven by the proposed paleo­
geographic shifts between the facies belts generated during different ice
advance-retreat cycles or facies belts (Bataller et al., 2019), the original
four or five-fold stratigraphic frameworks comprising Melaz Shuqran,
Lower Mamuniyat, Middle Mamuniyat, Upper Mamuniyat and Bir
Tlacsin (McDougall and Martin, 2000) and its equivalents proposed by
Ghienne et al. (2013) and Le Heron et al. (2009), amongst others cannot
be treated as basin-scale correlatable lithostratigraphic units, a concept
which is clearly no longer sustainable. As suggested by previous authors,
notably Ghienne et al. (2003), Moreau et al. (2015), Ghienne et al.,
2007a, 2007b, Girard et al. (2012a, 2012b), based on outcrop studies,
the focus should rather be on allostratigraphy and identification of
glacial advance-retreat cycles as the key element in understanding the
spatial distribution of these sediments and ultimately leading to more
realistic potential correlations. In this paper results clearly support these
previous conclusions and provide an example of how a detailed inte­
gration of subsurface data sets can synthetically support outcrop data
leading to enhanced correlations and modelling of these glacial depo­
sitional systems, with a special emphasis placed on seismic stratigraphy
to provide the three-dimensional allostratigraphic perspective.
5.1. Proposed model for the stratigraphic evolution of the upper
Ordovician
The integration of seismic stratigraphy, as exemplified by the two
seismic lines interpreted in Figs. 12 and 13, calibrated with well (core/
FMI) interpreted facies associations, together with the depositional
models previously presented and the results of seismic spectral decom­
position, as demonstrated by Bataller et al. (2019), has allowed the
reconstruction of specific sequences based on reflector termination and
seismic facies. As a summary of these observations and with the aim of
developing a chronostratigraphic depositional framework, we propose
the following idealized succession for a glacial advance-retreat cycle
based on this study area:
• Sub-glacial erosion surface generated by ice advance or a surface of
pro-glacial fluvial downcutting linked to glacioeustatic sea level fall
or, in some cases, post-glacial rebound. In the former case, this is
often associated with glaciotectonic deformation through several
metres of the poorly to moderately consolidated substrate, be it
Hawaz bedrock or intraformational in character
• In the hypothetical case of cold-based ice sheets, where subglacial
erosion is limited, sediment supply rates would generally be low to
very low. This outcome, although superficially similar to what may
be expected from an advancing ice sheet loaded with entrained

20
Another random document with
no related content on Scribd:
 —Grâce à Dieu! le Généreux! le Bienveillant! Nous nous sommes
rassasiés de couscous et de poulets.
Le lettré est un homme pauvre, et les quelques sous versés toutes les
semaines par ses élèves lui permettent à peine de subsister. Ce n’est qu’aux
fêtes, où chaque enfant apporte sa part du festin, que le maître peut calmer
la faim qui le tenaille sans répit.
Cependant le lettré se félicite d’un métier qui l’honore. Sa connaissance
impeccable du Livre lui procure des joies innocentes. Il aime à dérouler
l’interminable ruban des versets, selon les sept modes différents. Même, il
est capable, nous révéla-t-il un jour avec fierté, d’en réciter plusieurs
chapitres à l’envers, en commençant par le dernier mot.
—Comment est ton état, ô lettré? Es-tu content de Saïd?
—Allah! O mon Maître! soupire le pauvre homme. Il m’a tué!... De ma
vie je n’ai connu un enfant pareil... Pardonne-moi, c’est pour cela que je
suis venu.
—Tu as honoré notre maison!... Qu’a donc fait Saïd, ce fils de péché?
—Par malice, il abîma sa planchette. Je lui ordonnai de descendre dans
la cour afin de la blanchir, et, comme il s’obstinait à ne pas bouger...
Le lettré s’arrête et paraît fort gêné.
—O lettré! il fallait le battre.
—Allah!... J’ai voulu lui donner quelques petits coups de baguette sur la
plante des pieds, mais aussitôt,—hachek!—il a fait voler son eau sur moi!...
Puis il s’est roulé à terre en poussant des cris affreux, comme si on le sciait
en deux.
—Où est-il à présent, ce vaurien?
—J’ai fermé l’école avec ma clé, laissant les élèves sous la surveillance
de mon fils, et je suis venu prévenir le hakem... Cet enfant l’emporte sur
moi!
Pour un peu, le lettré se mettrait à pleurer; ses mains tremblent...
d’indignation peut-être... de crainte aussi. Évidemment il a peur que mon
mari ne donne raison à l’exécrable Saïd. Je dois le rassurer, et partir moi-
même avec lui afin que cette affaire se dénoue, sans plus d’atteinte à son
prestige.
Nous avons choisi son école parce que les notables de la ville y envoient
leurs enfants. Selon la coutume, elle dépend d’une mosquée, ainsi qu’un
hammam où les fidèles se purifient, et la fontaine. Cette mosquée étant
parmi les plus anciennes de Meknès, le hammam est noir de crasse, la
fontaine a perdu toutes ses mosaïques, et les poutres sculptées, qui
soutiennent l’école, fléchissent, près de s’effondrer. Mais, comme toujours
en pays musulman, du milieu des ruines surgit une intense vie joyeuse. La
vieille école s’emplit de la cadence ardente sur laquelle cinquante petites
voix récitent le Coran. Dès la ruelle, j’en perçois les modulations, les coups
de baguette scandant la mesure, et il me revient à l’esprit l’histoire de cette
sultane qui faisait élever cent jeunes vierges à l’exercice perpétuel du
Coran, «si bien que leur bourdonnement surpassait en douceur celui des
abeilles, et que leurs paroles étaient plus savoureuses que le miel».
Le lettré introduit, dans une serrure ingénieuse et primitive, sa clé en
bois hérissée de clous.
Nous montons un lamentable escalier, étroit et raide, dont les générations
ont fait sauter les mosaïques et usé les poutrelles.
Tranquillement accroupi près du seuil, au milieu de cinquante petites
paires de babouches, Saïd se plaît à les mélanger, avec un air de malicieuse
satisfaction. Mais, dès qu’il m’entend, le petit scélérat se met à pleurer et à
pousser mille cris effrayants. A lui seul il couvre la voix de tous ses
compagnons qui égrènent les pieux versets.
—Sellal Qlouba! Sellal Qlouba! vocifère-t-il.
—Que dis-tu, Saïd?
—Sellal Qlouba est dans la rue! J’ai peur de Sellal Qlouba! O ma mère!
protège-moi! O ma mère! Je suis réfugié en toi! sanglote le petit en se
prosternant à mes pieds pour embrasser ma robe.
Le lettré m’explique, d’une humble voix effrayée, qu’un bruit s’est
répandu depuis quelques jours: un homme, venu de loin, Sellal Qlouba,—
l’arracheur de cœurs,—parcourt la ville avec un fusil et une sacoche où il
enferme les entrailles de ses victimes... La voix du lettré s’éteint, de plus en
plus basse. On dirait qu’il craint d’être entendu par Sellal Qlouba. L’effroi
le paralyse autant que ses écoliers dont les visages se contractent depuis que
la malice de Saïd réveilla leurs alarmes.
J’ai grand’peine à emmener l’enfant qui, par méchanceté, refuse de
descendre l’escalier et se laisse à moitié rouler sur les marches disjointes.
Arrivé dans la rue, il change d’attitude. Nous devons traverser les souks,
et il escompte déjà les pois chiches grillés qu’il pourra s’acheter si je lui
donne un sou. La face comique de ouistiti s’exerce au sourire.
Mais nous passons devant le marchand de pois chiches sans nous arrêter.
—N’as-tu pas honte, ai-je répondu à sa demande, c’est du bâton que tu
devrais manger!
—Je n’ai pas voulu descendre pour blanchir ma planchette, à cause de
Sellal Qlouba, reprend-il. J’avais peur.
—Allons, Saïd! Il n’y a pas de Sellal Qlouba, tu le savais bien quand tu
as crié tout à l’heure. Et, du reste, il ne faut craindre qu’Allah.
—Il ne faut craindre qu’Allah! répète docilement la petite voix.
Il trottine auprès de moi, rasséréné, mais tout à coup je sens sa main
trembler dans la mienne.
Une troupe de gamins remonte la rue avec des cris épouvantables.
—Sellal Qlouba! hurlent-ils, Sellal Qlouba...
Les boutiquiers inquiets rabattent en hâte les volets de leurs échoppes;
les fillettes qui allaient à la fontaine, chargées de leur cruche, se sauvent en
pleurant; des femmes affolées s’empêtrent dans leurs haïks; quelques
hommes se précipitent vers la mosquée...
Dès qu’il est à la maison, Saïd, encore tout ému, terrorise les petites
filles par ses descriptions.
—Il est plus grand qu’un minaret, il a un ventre comme une outre. Sa
bouche! ô mes sœurs! se bouche est semblable à Bab Mansour[65]. Vous
pouvez demander à ma mère. Elle l’a vu.
Qui donc oserait nier l’existence d’un être qui met toute la ville en
panique?
Sellal Qlouba!
L’arracheur de cœurs!

3 janvier 1917.
Une suite d’événements palpitants a secoué l’indolence habituelle des
jours, en la demeure de Si Larbi el Mekki.
Ce fut d’abord le mariage de Fathma, sa fille cadette, et, le soir même du
départ pour la maison nuptiale, l’accouchement imprévu de sa tante Drissia.
Elle était là, en grand costume, un éblouissant caftan jaune fleuri de
bouquets multicolores, et elle prenait sa part des réjouissances, lorsque tout
à coup elle poussa un cri, puis un autre, le visage crispé de souffrance...
mais bientôt ce fut fini, deux jumeaux venaient de naître au son des
instruments.
Le fête ne fut interrompue que fort peu d’instants. Dès que l’accouchée
eut été installée sur les matelas au fond de la salle, les yous-yous et les
chants reprirent avec une nouvelle vigueur. Les invitées commentaient, sans
se lasser, l’inattendu de cet incident et répétaient:
—Grâce à Dieu! Quelle chose étonnante! Elle n’a poussé que deux cris!
Le cortège nuptial étant parti, je quittai l’assemblée, malgré les instances
pour me retenir, car on allait mettre le henné à cinq petits garçons, dont la
circoncision aurait lieu le lendemain.—Si Larbi ayant sans doute estimé que
les frais et l’embarras des noces serviraient ainsi à double fin. Il n’avait
point prévu qu’Allah en ajouterait une troisième, et même une quatrième,
car un des jumeaux mourut pendant la nuit, et son cercueil fut emporté dès
l’aube,—bien avant que n’arrivât le siroual[66] de la mariée.
Je suis revenue ce matin. Les joueurs de hautbois et de tumbal s’exercent
déjà devant la porte, les joues démesurément gonflées ou les baguettes
rageuses. La maison bourdonne comme une ruche. Si Larbi piétine en son
vestibule, impatient de diriger toutes choses, mais ne pouvant, à cause des
invitées, pénétrer dans sa demeure.
Les négresses se bousculent à travers le patio, elles installent les sofas,
versent des bols de fumante harira, préparent les plateaux à thé, les coupes
pleines de henné, de sel et de cumin qui serviront tout à l’heure.
Les petits héros de la fête sortent un à un dans la cour, superbement
vêtus. Leurs caftans de drap aux vives couleurs traînent à terre,—car ils
n’ont pas de ceinture aujourd’hui.—Ils sont recouverts d’une courte tunique
ramagée d’argent: des bandelettes blanches criblées de taches roses
ceignent leurs fronts, et leurs burnous d’un vert aigre blessent les regards.
On les installe sur une estrade, autour de laquelle les femmes en toilette
viennent s’accroupir. Il y a le fils aîné de Si Larbi, un grand garçon mince
qui a peut-être douze ans, puis trois de ses cousins beaucoup plus jeunes, et
enfin le minuscule négrillon Messaoud, enseveli dans l’ampleur de ses
vêtements.
Il doit être bien étonné, le pauvre gosse, de se trouver ainsi paré! Certes,
il n’échappe à personne qu’il est un esclave, dont les caftans trop longs, le
burnous défraîchi, furent prêtés pour la circonstance, alors que ses
compagnons arborent fièrement leurs draperies neuves. Mais il domine
l’assistance, il est assis sur des coussins, il n’a rien à faire et les femmes ont
poussé des yous-yous à son apparition! Ses yeux ronds s’écarquillent plus
que d’habitude avec une naïve expression de stupeur.
A côté de lui, un bambin ne cesse de pleurer, affolé par la perspective de
l’opération. Ses mains, agrippées à la robe de sa mère, la retiennent, près de
lui, droite devant l’estrade, troublant ainsi l’ordonnance de la fête. Et le
petit lève vers la jeune femme de pitoyables regards suppliants.
—O mon malheur! gémit-il sans relâche. O mon malheur!
Les autres sont dignes, un peu émus sans doute au fond du cœur, mais ils
s’étudient à rester impassibles et raides, ainsi qu’il convient. Quelques
propos des invitées doivent parvenir jusqu’à eux et les troubler davantage.
Car elles parlent sans aucune retenue de la prochaine cérémonie; elles en
décrivent complaisamment les détails aux tout petits, vautrés auprès d’elles,
et dont ce sera le tour dans quelques années.
Ma voisine, qui étale un étourdissant caftan violet à fleurs géranium, fait
même, avec deux doigts écartés en ciseaux, des gestes d’une trop explicite
impudeur... Le bébé, vers qui elle se penche, n’en paraît point ému et
continue à sucer son pouce en toute sérénité.
Le soleil descend peu à peu dans le patio, il ajoute aux toilettes un éclat
superflu. Des roses faux heurtent les bleus trop vifs, les oranges, les jaunes
ardents, les ramages d’or qui fulgurent en éclairs à travers les satins.
Et puis l’acidité agaçante des cinq petits burnous verts...
Mais les musiciennes, tapant à tour de bras sur les tambours de formes
diverses, et chantant avec fureur, dominent le tumulte des gens, des voix et
des couleurs.
... Drissia l’accouchée, halette sur des matelas, le visage rouge et les
mains brûlantes.
Tout près d’elle des invitées, très splendides, causent avec une animation
qui m’étonne. Je saisis le nom, cent fois répété depuis quelques jours, de
Sellal Qlouba.
L’arracheur de cœurs, personne ne l’a vu, mais chacun le décrit et en
propage l’épouvante.
«Les gens le disent.» Cela suffit. Des paniques se multiplient à travers la
ville et les écoles demeurent désertes, car les mères n’osent plus laisser
sortir leurs petits.
Lella Lbatoul, parente de Si Larbi, est ici. Je vais m’asseoir auprès d’elle
et l’interroge:
—O docte et prudente! toi qui ne prononces point de paroles au hasard,
explique-moi cette étonnante histoire de Sellal Qlouba. Y crois-tu
vraiment?
—J’ai appris, me répond-elle, à me défier des choses qui passent de
bouche en bouche, et sont racontées par les enfants ou les esclaves.
Cependant il me semble qu’on ne parlerait pas ainsi de Sellal Qlouba s’il
n’existait pas... Les gens disent que c’est un homme de la tribu des
Mzadem[67], très loin, dans le sud, au delà de Marrakech. Or, par une
malédiction d’Allah, tous ceux de cette tribu sont affligés d’un chancre qui
leur ronge le nez. Et ce mal ne saurait guérir qu’au moyen d’un remède
composé par un taleb[68], avec des cœurs arrachés aux petits enfants. C’est
pourquoi Sellal Qlouba partit en chasse à travers le pays.
—Connais-tu, dans ton entourage, un seul enfant qui ait été sa victime?
—Non, grâce à Dieu!... Aussi ne suis-je pas très assurée que Sellal
Qlouba soit à Meknès. Cependant, par précaution, je n’ai point envoyé mon
fils à la mosquée tous ces jours-ci...
Tandis que nous causions, un ordre est arrivé du vestibule, et l’excitation
s’exagère. De robustes négresses se placent devant les jeunes garçons
qu’elles chargent à califourchon sur leur dos. Le petit éploré jette des cris
aigus et tend désespérément les mains vers sa mère:
—Je ne veux pas! Oh! je ne veux pas!... Laissez-moi!...
On l’emporte de force avec les autres, dont le calme commence à se
démentir.
Aussitôt les mamans sont conduites vers l’estrade et installées à leur tour
parmi les coussins. Elles sont quatre, puisque, bien entendu, le négrillon n’a
pas la sienne ici, mais seulement une mère très lointaine, en Mauritanie ou
au Tchad, dans un des pays sauvages où l’on va voler des enfants afin de les
vendre ensuite aux habitants civilisés des villes marocaines.
Personne donc n’occupe la place de Messaoud, personne, en songeant à
lui, ne sent battre son cœur à trop grands coups. Les mamans semblent un
peu émues. Heureusement elles ont à remplir des rites très absorbants: bien
étaler les plis de leurs robes; tenir leur pied droit dans un bassin de cuivre
rempli d’eau, en y foulant le mors d’un cheval, dont les rênes, relevées
d’une main, sont mordues entre les incisives; et enfin se regarder, sans
distraction, en un petit miroir que l’on a placé dans leur autre main.
Ces gestes compliqués ont pour but, prétendent les lettrés, de fixer leur
attention de telle sorte qu’elles n’éprouvent pas un trop vif émoi durant la
circoncision. Mais elles, les femmes, gardiennes des traditions, savent bien
qu’elles accomplissent des rites très graves qui assureront le bonheur et la
santé de leurs fils.
La mère du petit éploré y met une conscience admirable; rigide,
immobile, les sourcils contractés par l’effort, elle cligne à peine des yeux,
absorbée en sa propre image. Les autres s’exécutent plus mollement et la
neggafa les en réprimande:
—O honte! dit-elle à l’une des étourdies, tu n’a pas mis de rouge sur tes
joues et tu effleures à peine les rênes de tes lèvres pendant que l’on
circoncit ton enfant! Prends garde qu’Allah ne fasse retomber sur lui son
mécontentement.
De l’autre patio, où sont réunis les hommes, on entend les sons aigres et
sourds des instruments. Un petit esclave arrive en courant, il porte sur sa
tête un plateau où l’on a déposé les caftans, les tuniques et les burnous vert
acide. Aussitôt après reviennent les négresses chargées de leurs fardeaux.
Ils ne sont plus enveloppés que d’un drap blanc, comme un suaire, et leurs
têtes ballottent à droite et à gauche, affreusement contractées par la
souffrance.
Ils crient! Ils crient! la bouche grande ouverte, les lèvres tordues. Ils
hurlent! mais on ne les entend pas, car les musiciennes hurlent plus fort
qu’eux en maltraitant leurs tambourins et les yous-yous des invitées
s’excitent à couvrir les voix douloureuses.
Les mamans maîtrisent avec peine leur émotion. Celle qui mordait si
négligemment ses rênes pleure à présent de toutes ses larmes. On dépose les
cinq petits sur un matelas et les négresses s’en vont, le dos de leurs
vêtements tout ensanglanté... Ils crient, les pauvres circoncis! Ils crient! Ils
lassent les chants et les yous-yous. Bientôt on distingue leurs gémissements.
Chaque mère console son fils, l’embrasse, lui promet «que c’est fini, qu’on
ne recommencera jamais».
Le négrillon reste tout seul, mais lui, il ne pleure pas du tout. Peut-être
comprend-il que ce serait inutile, qu’il n’y a personne pour le cajoler, ni
l’apaiser... A quatre ou cinq ans, déjà, il doit avoir sa philosophie de
l’existence... Un peu de sueur mouille ses tempes, une larme sèche au coin
de ses yeux, il a l’air encore plus ébahi que tout à l’heure. Sa petite patte
noire, crispée sur l’étoile, l’écarte de la cuisante blessure. Il attend
patiemment que se calme la souffrance et il regarde, sans mépris, ses
compagnons, tous plus âgés que lui, qui savent si mal supporter leurs
tourments. Ce sont les petits maîtres, les enfants riches et libres, ils ont des
parents pour les gâter... Lui, Messaoud le négrillon, n’en est pas à sa
première expérience douloureuse; depuis longtemps il sait accepter
silencieusement tous les maux, car les cris ne servent à rien et importunent
les gens. En sorte qu’aujourd’hui c’est lui le privilégié. Il souffre moins que
les autres.
Le grand Sadik oublie toute espèce de dignité et secoue sa tête en
sanglotant:
—Oh! Oh! Oh! le barbier! il m’a coupé!... Oh! Oh! le barbier!...
Et les autres, adoptant ce thème lamentable, hurlent en chœur:
—Oh! le barbier!... Oh! le barbier!...
Leurs cris montent, se dépassent, s’apaisent exténués, puis repartent avec
une nouvelle frénésie. Les musiciennes redoublent leurs efforts; les invitées
bavardent et changent de toilette, les esclaves s’affairent à préparer le festin,
dont treize plats déjà sont alignés dans la cour.
Et, au fond de la salle, Drissia l’accouchée agite ses bras en prononçant
des paroles incohérentes, tandis que le bébé vagit comme un jeune cabri.

5 janvier.
Lorsque j’entrai dans le harem de Mouley El Kébir, deux Juives
proposaient aux Cherifat des étoffes et des passementeries.
L’une était fort vieille, d’un âge indicible, avec un profil crochu, de petits
yeux ternes perdus au fond des orbites, une bouche édentée aux lèvres
minces, une flasque peau ridée pendillant sous le menton comme une
barbiche de chèvre, et des poignets sillonnés de veines, ainsi que ces troncs
d’arbres morts où s’incrustent les racines des lierres.
L’autre, toute jeune, jolie, potelée, rose et blanche. De larges yeux
inexpressifs éclairaient son doux visage innocent.
 Pourtant il y avait une ressemblance entre ces deux femmes et l’on
devinait qu’un jour, plus tard, il sortirait une affreuse vieille pointue, de tant
de grâce et de fraîcheur...
Elles se tenaient discrètement près de la porte, humbles, déférentes, avec
des sourires craintifs. Et elles se prosternèrent, sur le seuil, en quittant leurs
nobles clientes.
Le maître étant absent, des ordres furent donnés pour que s’éloignassent
les serviteurs mâles, et nous allâmes dans l’arsa soigneusement close. Les
Chérifat, nonchalantes, firent quelques pas dans les allées et, tout de suite
lasses, s’affalèrent sur des sofas que les esclaves disposaient le long d’un
mur. Je passai plusieurs heures avec elles.
Je partis vers le moghreb, et m’étonnai, au sortir du fantastique chemin
entre les ruines, de retrouver les deux Juives blotties l’une contre l’autre,
frissonnantes comme des poules durant un orage...
Elles se précipitent vers moi, baisent le bas de ma jupe, mon épaule, mes
mains.
—Nous nous mettons sous ta protection! Ne nous abandonne pas!
implorent-elles.
—Sans doute, mais qu’y a-t-il?
—Écoute! disent-elles avec un visage de terreur. Les Aïssaouas!...
Au delà de Bab Mansour, je perçois, en effet, la rumeur caractéristique,
le rythme précipité du nom d’Allah...
Les Juives continuent leurs jérémiades:
—Nous n’osons passer, et voici que le moghreb approche!... Ah!
Seigneur! Les Aïssaouas nous tueront certainement... ils égorgent et
dépècent les Juifs qu’ils rencontrent, c’est leur coutume... Azar Tobi rentra
l’autre jour, échappé de leurs mains, avec un visage en sang, et des
vêtements tout déchirés!... Qu’allons-nous devenir? Prends nous sous ta
garde! Auprès de toi, sans doute, ils n’oseront nous toucher.
Des larmes brillent dans les petits yeux desséchés de la sorcière, elles
ruissellent sur les joues roses de sa fille. J’arrive péniblement à me libérer
de leurs bras et je traverse Bab Mansour entre les tremblantes Juives.
A l’autre extrémité de la place El Hédim, un groupe d’Aïssaouas se livre
aux pieuses contorsions d’usage. Ils sont loin et fort préoccupés de leurs
danses, ils ne nous aperçoivent même pas. Les femmes se rassurent et me
remercient.
—Rentrez chez vous par les souks, leur dis-je, vous n’avez plus rien à
craindre.
Mais, aussitôt le péril écarté, elles ont repris leurs préoccupations
mercantiles.
—Non, me répond la vieille, nous n’allons point encore au Mellah, mais
du côté de ta demeure, chez le Chérif Mouley Hassan, afin de proposer des
tentures pour la chambre nuptiale qu’il prépare.

22 janvier.
Depuis hier, Saïd est malade, de sa maladie habituelle, une effroyable
indigestion. Car Saïd, parmi tous ses défauts, ne «rétrécit» pas quant à la
gourmandise, mais ses intestins délabrés ne peuvent supporter les choses
bizarres dont il est si friand et qu’il parvient à se procurer malgré notre
défense: halaoua[69] qu’un marchand déroule d’un bâton, figues de
Barbarie, millet agglutiné dans de la mélasse, et, surtout, pois chiches secs
et croquants. Les petites amulettes d’argent, que nous avions suspendues à
sa mèche d’Aïssaoui, ont disparu mystérieusement. Saïd prétend que des
camarades les lui dérobèrent à l’école. Je croirais plutôt que Saïd les a
vendues, ou échangées contre des gâteaux.
Mais voici bien des jours qu’il ne lui reste plus rien à monnayer, et je
comprends mal comment il put acheter cette provision de beignets et de
glands-doux rôtis que je viens de découvrir derrière son lit. A toutes mes
questions, il répond par de nouveaux cris scandés de gémissements
lamentables:
—O mon malheur! ô ma petite mère... Mes os sont cassés!... O mon
foie!... Mon cœur éclate!
—Tu es encore une fois retourné chez tes sœurs! Ce sont elles qui t’ont
donné ces beignets?
—O ma mère! Par le serment je ne les ai pas vues! Je n’ai pas quitté la
mosquée avant l’aser. Demande au lettré... Comment aurais-je été chez mes
sœurs?... O mon petit ventre. Qu’il me fait mal!
Saïd a toujours les accents de l’innocence. Je renonce à savoir et vais
retrouver mon mari dans le salon. Kaddour l’avertit, justement, qu’un
indigène attend à la porte.
 —Qui est-ce?
—Je ne le connais pas. Il dit qu’il veut te parler, à toi-même... Sur lui,
pas de mal, ajoute le mokhazni pour exprimer que l’autre semble riche.
—Fais-le monter...
Kaddour accompagne un Marocain bien vêtu, à la figure blême et
bouffie, au regard fuyant. Sans doute un marchand de Fès dont il a le type.
Il nous salue avec des formules obséquieuses que mon mari doit arrêter.
—Est-ce pour une affaire? Pourquoi ne pas être venu me parler au
bureau?
Après des explications compliquées, le Marocain finit par solliciter un
permis pour sortir du sucre. Il veut l’envoyer à Fès, où le bénéfice est plus
fort, évidemment.
—Tu sais bien que chaque ville reçoit sa part de sucre. Si j’en laissais
sortir, j’en priverais les gens d’ici.
—Ta parole est la plus grande, ô hakem!... Je te demande cinquante
petits sacs, pas davantage. Il y en a tant d’autres à Meknès!
—Excuse-moi, c’est tout à fait impossible.
—Je me réfugie en ton enfant, ô hakem! Je sais que Saïd est cher à ton
cœur. Allah protège tes jours et les siens!... Quarante petits sacs seulement?
—Assez de paroles. Je ne peux t’en laisser sortir même la moitié d’un.
Le gros marchand comprend que l’insistance est inutile. Cependant, il
semble sur le point d’ajouter quelque chose... il hésite... puis se ressaisit et
s’éloigne lentement.
Mais, après un instant, Kaddour revient.
—Qu’est-ce encore?
—Cet homme, il demande l’argent.
—Comment l’argent?... Quel argent?
—Il dit: les cinq réaux qu’il a donnés hier au petit pour qu’il te parle de
cette affaire.
... L’acquisition des beignets et des glands ne m’étonne plus, ni même la
vénalité de Saïd qui trafique à présent de son influence!
Dès nos premières questions il se remet à pleurer pitoyablement; des cris
affreux couvrent nos reproches. Saïd paraît soumis à tous les tourments des
djinns.
 —Allons, Kaddour! c’est clair. Le marchand a dit vrai. Rends-lui ses
réaux, et conseille-lui de ne plus heurter à notre porte.
Saïd se tord et gémit. L’effroi contracte sa petite figure simiesque. Il est
tout à fait affolé.
Le battre?... A quoi cela servirait-il? Aucune punition ne peut le corriger,
il est mauvais jusqu’aux moelles... Et puis, aujourd’hui sa maladie n’est pas
feinte. Demain il aura perdu le souvenir de sa faute.
Mon mari se contente de le menacer des plus épouvantables châtiments
s’il reçoit, à nouveau, les cadeaux des gens.
—O mon père! répète l’enfant tout contrit, obéissant à Dieu[70]... De ma
vie je ne recommencerai!... Obéissant à Dieu! Obéissant à Dieu!

6 février.

«C’est entre lys, cassies, roses, odeurs suaves,

Chansons, amis tendres, boissons et musiciennes
Que l’âme s’épanouit dans la joie [71]...»

La voix du chanteur, pleine et sonore, alanguit notre indolence.

Étendus sur les sofas gonflés de laine souple, nous possédons tout ce qui
enchante l’être délicieusement: la félicité du repos, la quiétude, l’ivresse
engourdissante des parfums, et ce riadh irréel, bleu, glacé de lune, qui
s’étend devant la belle salle où nous sommes réunis.
Jouissons de l’heure et de ses plaisirs! Comme les peintures du plafond,
la musique enlace mille arabesques plaisantes sur un thème simple. L’esprit
s’amuse à en suivre les détours un instant, puis, lassé par cet effort,
s’abandonne à sa béatitude...
Des esclaves au corps parfait passent dans l’allée miroitante, derrière les
rames des bananiers. Les paons se sont perchés très haut dans les branches.
Au sommet du jet d’eau, dansent les reflets de lune... Le jardin, plein de
senteurs, dort, étrangement verdi par la froide lumière. Bleuâtres et mauves
comme des fleurs perverses, les roses défaillent sous les orangers.
Afin de mieux goûter ces délices nocturnes, Si Ahmed Jebli, notre hôte,
a fait venir de Fès le chanteur célèbre, le maître El Fathi. Les amis de choix,
rassemblés, lui savent gré de ces jouissances délicates, mais en témoignent
discrètement. Mouley Hassan qui, parfois, a recours au riche marchand pour
des emprunts, daignera, ce soir, honorer notre réunion...
Le Chérif se fait attendre longtemps... Un mouvement parmi les esclaves
nous avertit de son arrivée. Majestueux et trop fier, il entre en saluant d’un
signe de tête imperceptible, et, conduit par le maître de maison, il s’installe
au milieu du divan, à la place d’honneur, juste devant la porte et le magique
jardin sous la lune...
Il a le visage grave d’un prince observé par la foule.
Presque aussitôt, El Fathi prélude. Jusqu’alors il laissait aux autres
musiciens le soin d’occuper l’assistance. Sa voix emplit la vaste salle. Une
voix souple et savante, au timbre inattendu, très haute, gutturale et belle
cependant. Il domine l’orchestre qui épie ses moindres gestes, il lui impose
son rythme personnel et ses variations. D’une main il frappe
impérieusement le divan pour marquer la cadence. Lorsque El Fathi finit un
thème, les musiciens le reprennent en sourdine, avec des modulations
imperceptibles. Les chants adoucis du chœur laissent mieux percevoir
l’accompagnement du luth, et celui du rbab qui gémit comme une
tourterelle.
A des motifs larges, de plain-chant, succèdent les phrases d’une
mélancolie raffinée. La poésie désuète de leurs paroles accentue cette
impression poignante dont nous étreint l’œuvre des civilisations très
anciennes. A travers les chansons, l’amour s’exalte, rit et pleure, mais
parfois aussi une plainte évoque les temps révolus:

«O mon regret pour les jours passés

Dans les plaisirs, dans la joie,
Jours favorables et paisibles!

»O séparation des demeures de l’Andalousie,

Donne-moi du répit!

»O Allah! par ta grâce et ton assistance,

Par ton Prophète bien-aimé,
Apaise ma douleur incessante!

»O séparation des demeures de l’Andalousie,

Donne-moi du répit!»
 Grenade!... Terre qu’Allah fit enchanteresse! eaux murmurantes, vaste
plaine aux horizons infinis, incendiés de soleil, et les blanches sierras
glacées!... Divine Grenade où les Maures ajoutèrent de la beauté!
Ils savaient que les eaux doivent ruisseler des vasques et que les jardins
pleins de cyprès, de jasmins et de roses, s’encadrent de buis symétriques. Ils
savaient qu’aux sommets des plus merveilleuses collines, il faut des palais
de marbre où l’on enferme les sultanes...
Qu’avons-nous fait de Grenade après eux?
Qu’avons-nous su?...

«O séparation des demeures de l’Andalousie,

Donne-moi du répit!»

Devant ce riadh frémissant de feuillages et d’esclaves, je sens la détresse

de l’Alhambra, de ses cours désertes, mortes... Mais il ne sied pas d’attacher
trop d’importance à la musique profane. Ces lamentations n’ont ému que
moi, l’étrangère.
Nos compagnons, installés par petits groupes autour de la salle, écoutent,
impassibles. Si Ahmed Jebli et deux ou trois de ses amis, originaires de Fès
comme lui, et plus mélomanes que les Meknasis, battent la mesure de leur
orteil.
Lorsque le chant se termine, sur une sorte d’invocation lancée par El
Fathi, des négresses aux bras robustes apportent les plateaux, les aspersoirs,
les brûle-parfums. Notre hôte dispose lui-même, sur les braises, des
morceaux de bois odorant qu’il tire d’une cassette en argent.
Que la vie semble bien faite et suave en cette soirée! Le thé à la
citronnelle, les parfums, les chants, les belles draperies et les sofas
moelleux contentent les sens, tandis qu’une musique raffinée, de paisibles
entretiens occupent l’esprit sans le lasser...
Lorsque Mouley Hassan parle, chacun l’écoute avec déférence. Il revient
inlassablement à lui-même et aux siens.
—Certes, dit-il à mon mari, Mouley Ismaïl fut au Maroc l’unique sultan.
Il se faisait appeler le diadème des princes... Plus de cent mille soldats
nègres composaient ses armées; d’innombrables ouvriers travaillaient à ses
palais ou à des fortifications que des gens ont cru, depuis, être l’œuvre des
djinns. Tous les pays berbères, contre lesquels les Français luttent à présent,
lui étaient soumis. Et, pour les maintenir dans l’obéissance, il conçut dans
sa vieillesse, après cinquante ans de règne, le projet de relier Meknès à
Marrakech par des remparts ininterrompus.
«Les aveugles, disait-il, pourront se diriger à travers le pays, en suivant
ces murs de leurs bâtons.» Il l’eût fait, si son destin n’avait été enfin écrit.
»Nous, les Ifraniin, poursuivit Mouley Hassan avec orgueil, sommes
d’une autre lignée de Chorfa, plus proches du Prophète; mais après deux
siècles, en considération de Mouley Ismaïl, nous épousons encore ses
descendantes. Le sang du grand sultan, que me transmirent ma mère et mes
aïeules, était digne de s’allier à celui de mes ancêtres.
Nos compagnons, recueillis, approuvaient en hochant du turban. Et,
comme les musiciens préludaient à nouveau sur les luths, Mouley Hassan se
leva.
Sans doute, tenait-il à marquer ainsi qu’il était venu par condescendance,
et non pour le plaisir de la musique.
—J’ai des esclaves, avait-il dit avec négligence, qui frappent du luth, du
rbab, et du tambourin à la limite de la perfection; et d’autres qui chantent
tous nos vieux airs andalous ainsi que ceux du Caire, de Fès et d’Alger. Je
n’épargnai rien pour leur éducation et les fis initier à Fès, dans l’art des
instruments, par le maître Saouri...
Après son départ, les conversations devinrent plus familières. Les autres
invités, riches négociants et possesseurs de cultures, se sentaient mieux
entre eux.
—Mouley Hassan a omis de te parler du dernier sultan de Meknès, son
cousin, nous dit aussitôt le tajer Ben Melih; si Mouley Ismaïl a régné plus
de cinquante ans, celui-là ne régna pas cinquante jours... Encore ne régnait-
il que sur ses propres esclaves, car il n’osait quitter son palais. Il n’avait pas
un soldat et le trésor était vide... Son vizir, Si Allal Doukkali, cet
orgueilleux que tu connais, réunit une fois au Dar Maghzen tous les
négociants de Meknès. Il leur fit part de cette détresse. Et nous, d’une seule
voix, nous assurâmes ne pas avoir un liard pour donner à notre maître.
»Cependant je possédais mille sacs de sucre et ne pouvais les dissimuler
comme des réaux. Or le sultan me pria de les lui prêter pour en faire de
l’argent. Mon embarras fut extrême... J’acceptai, sous la condition que Si
Allal garantirait la dette de son maître... Mais le vizir s’y refusa. Il n’avait
pas plus confiance que moi-même, et je gardai mon sucre... Grâce à Dieu!
car, ayant appris que les Français approchaient de Meknès, le sultan
s’empressa d’abdiquer quelques jours plus tard...
—Nous nous divertissons encore en songeant à cette aventure, reprit Si
Ahmed Jebli; mais certes nous n’avons pas à dire contre ce sultan, le
pauvre!... Il ne fit de mal à personne et son cœur était blanc...
—Tel n’est pas celui d’un Chérif d’entre les Chorfa, dont on sait les
histoires curieuses, insinua Si Larbi, et qui s’enrichit avec les dépouilles,
non de ses ennemis, mais de ses épouses... Si le Coran excellent n’avait fixé
à quatre le nombre de nos femmes, il posséderait tout l’Empire fortuné... Il
portait son choix sur les plus riches orphelines, afin de les mieux spolier.
Quand un tuteur résistait, il le faisait destituer en payant le Cadi... On
raconte que ce Chérif admirable ne fut arrêté que par la résistance d’une
petite fille...
A ces paroles, nos compagnons sourirent discrètement, mais leurs
visages devinrent plus graves lorsque notre hôte déclara:
—Une petite fille ne saurait s’opposer longtemps aux desseins d’un
puissant... Sachez que celui-ci offrit au Sultan des présents si splendides,
que notre maître ordonna de célébrer le mariage sans tarder... Telle est
l’histoire du Chérif et de l’adolescente rébarbative, bien plus surprenante,
en vérité! que toutes celles que nous entendîmes aujourd’hui.
Ainsi j’appris comment est fixée la destinée de Lella Oum Keltoum...
Les grands murs sans fenêtres, aux portes toujours closes, ne suffisent
pas à garder leurs secrets. Et ces bourgeois si prudes, qui ne prononcent
point le nom d’une femme, songeaient tous à la jouvencelle dont la
fraîcheur et les richesses réjouiront les dernières années de Mouley Hassan,
tandis qu’El Fathi, de sa voix suraiguë, détaillait les charmes d’une belle.
 «O sourire de la bien-aimée, aussi clair que la rose
Mouillée par la rosée matinale!
O son allure quand elle marche et se pavane!
Comme une branche vêtue de ses feuilles!
O sa bouche, rayon de miel parfumé!
Autour d’elle, tournoient les abeilles...»

15 février.
—C’est un Juif, hachek! me dit Yasmine.
Hachek: formule de pudique restriction, dont la nôtre, «sauf ton respect»,
ne rend pas le pittoresque.
Yasmine est une fillette bien élevée. Elle n’ignore pas qu’il convient
d’ajouter «hachek!» après avoir nommé les choses et les animaux les plus
vils, du bitume, du charbon, un âne, un chien, un Juif...
Quelques-uns poussent la décence plus loin encore.
—Une femme! hachek! ne manque pas de dire notre correct serviteur
Hadj Messaoud, même lorsqu’il s’adresse à moi.
Donc c’est un Juif, sauf mon respect! Que veut ce Juif? Il se présente,
humble et noirâtre, fouille en sa vieille sacoche et me tend une bague
ancienne ornée de rubis.
—Elle est à toi, me dit-il.
Je repousse le bijou, indignée, mais non surprise, car il est habituel de
vouloir corrompre la femme du hakem.
—Pardonne-moi, insiste le Juif, elle t’appartient. Tu l’as achetée, il y a
un an, au fils du rabbin qui est mon neveu. Je l’ai reconnue quand on a
voulu me la vendre et c’est pourquoi je te la rapporte.
J’examine la bague. Ce Juif a raison. Quel voleur avisé l’a donc
soustraite à nos collections, sans que je m’en aperçoive?
—Un enfant, tout petit, me dit le Juif. Il me l’a proposée pour un
guirch[72]. Lorsque je l’interrogeai, il prit peur et se sauva. Mais je le
reconnaîtrais bien.
Moi aussi! Ce ne peut être que Saïd, le tourment de notre vie.
Je congédie le Juif avec des remerciements, car il refuse toute
récompense et multiplie les protestations de reconnaissance et de
dévouement.
—Que le Seigneur nous laisse le hakem, en fait de bénédiction! ne cesse-
t-il de répéter.
Maintenant il va falloir punir Saïd... Ah! je suis lasse!... Cet enfant a le
génie du mal!... L’autre jour il fit à Rabha des propositions indécentes...
Hier il débonda la fontaine, inondant ainsi le patio.
Saïd est fouetté... Hurlant, rageur, il se précipite vers le salon:
—O mon malheur! s’écrie Yasmine. Que va-t-il faire?
C’est vrai. Saïd a la coutume de se venger quand on le punit, et il conçoit
des vengeances ingénieusement détestables.
Je suis Yasmine, à sa recherche. Sur le seuil de la salle, nous nous
arrêtons, horrifiées: au milieu de notre plus beau tapis, un vieux Rabat,
velouté comme un tapis de Perse, Saïd vient de déposer... ce qu’il a
déposé!... Hachek!

24 février.
—Avoue-le, Saïd, tu es retourné chez tes sœurs aujourd’hui.
—O ma mère, tue-moi si je les ai vues!
—Tu mens! Kaddour vient de t’apercevoir sortant de chez elles.
—Par le Dieu Clément! profère l’enfant, je n’ai pas même passé dans le
vent de leur quartier!
—Et comment Kaddour t’y a-t-il reconnu?
—Fais attention, ô ma mère, que Kaddour a pu se tromper. N’y a-t-il pas
d’autres enfants de ma taille à Meknès?
Saïd a le raisonnement subtil et prompt. Plus tard, s’il devenait un lettré,
il excellerait aux discussions oiseuses et à la controverse.
—Prends garde surtout de ne point aller chez tes sœurs.
—O ma mère, ta parole est sur ma tête! Comment irais-je puisque tu me
l’as défendu? Et puis, qu’ai-je à faire avec ces chiennes? Se sont-elles
souvenues de moi quand mon père m’a chassé?
—Bien. Va jouer avec Rabha.
Saïd descend l’escalier en s’aidant de ses mains pour franchir les
marches hautes. Il est encore si petit! Puis il se dirige vers la cuisine.
 A cette heure il n’y a peut-être personne, et Saïd, seul à la cuisine, c’est
le prélude assuré d’une indigestion.
Je veux l’y chercher, Yasmine m’arrête un moment au passage, et, quand
j’arrive, Saïd est déjà grimpé sur le fourneau, parmi les casseroles. Il
examine leur contenu, tellement affairé qu’il ne m’entend pas. Du reste, j’ai
marché sans bruit afin de le surprendre dans son vol. Mais, à mon
étonnement, au lieu de pêcher un morceau, Saïd tire de sa petite sacoche un
papier et, dans la marmite élue, jette une sorte de poudre.
—Que fais-tu là? dis-je brusquement.
—O ma mère!... Avec ce temps froid, je me chauffais.
—Et cette poudre que tu as versée? Qu’est-ce que cette poudre?
Cette fois Saïd ne saurait nier, la moitié du paquet est encore dans sa
main. Il se met à trembler, tandis qu’une crainte passe en mon esprit...
—O ma mère! pardonne-moi. Je ne sais pas ce qu’est cette poudre... Mes
sœurs me l’ont donnée ce matin. Elles m’ont promis des oranges si je la
mettais, sans être vu, dans votre nourriture, là où il y aurait de la tomate... O
ma mère, je ne croyais pas mal faire, pardonne-moi!
Pour la première fois, Saïd a dit la vérité, car elle lui paraît moins
effrayante que le mensonge. Une angoisse me trouble tandis que les paroles
de Larfaoui reviennent à ma mémoire... Il n’est pas besoin que Kaddour
confirme ce que, déjà, j’ai deviné...
—O Puissant! s’écrie-t-il après avoir examiné la poudre que je lui tends,
c’est du rahj[73], ce maléfice que l’on vendait au souk avant l’arrivée des
Français!... Par le Prophète! est-ce possible? Ce fils de péché voulait vous
empoisonner!
Saïd a pris un air tellement candide que je ne sais même pas s’il
comprend l’action que ses sœurs ont voulu lui faire commettre... Mais que
ne commettrait-il pour une orange?
Kaddour est devenu bien jaune, et ses yeux noircissent à la limite des
ténèbres. Sans un mot, il saisit l’enfant et lui, toujours indulgent à ses
fautes, tendrement habile à leur trouver des excuses, il se met à le battre
avec rage.
Saïd pousse d’épouvantables rugissements. Kaddour a la main si dure!
—O mon père! crie l’enfant, ô mon père, secours-moi!... Je veux
retourner chez toi! Viens me prendre, ô mon père!... Ils veulent me tuer! ô
mon père!
Je parviens, toute tremblante, à arrêter Kaddour qui frémit.
—C’en est assez! Emmène-le à son père!... Et qu’on ne le revoie
jamais!... Ses sœurs, tu les conduiras au pacha. S’il plaît à Dieu, elles
expieront leurs méfaits... Ne touche plus à ce démon. Que le potier se
débrouille avec ce qu’il a engendré!
Kaddour s’éloigne, traînant Saïd en pleurs. La misérable petite chose qui
était entrée dans notre vie s’en détache...

Délivrée de Saïd, que l’existence paraît donc savoureuse et facile!

8 mars.
Un petit tas rutile au soleil sous les arcades. Les caftans accroupis
dépassent à peine une coudée au-dessus du sol. Le caftan jaune de Rabha se
penche vers les caftans roses et bleus de Yasmine et de Kenza.
Je sais qu’il n’est pas question de poupées, les fillettes marocaines ne
connaissent guère cette distraction, mais plutôt de quelque histoire
colportée par les terrasses.
Des phrases, parvenues jusqu’à moi, attirent mon attention:
—Elle était vierge, déclare Kenza.
—Les gens le disent!... Son visage est rond et brillant comme la lune.
Dada Fatouma l’a vue...
—Tous les hommes sont fils de péché, prononce Yasmine, avec une
mine avertie.
—L’autre se dessèche et jaunit de teint.
—De qui parlez-vous, petites filles? demandai-je.
—De Lella Meryem... O ma mère, l’ignores-tu? Cette gazelle a une
rivale dans sa demeure! Mouley Hassan vient d’offrir à son fils une belle
esclave blanche, et Mouley Abdallah est entré, chaque nuit, dans sa
chambre...
—Chose surprenante, en vérité! Qui te l’a rapportée?
—Une négresse de Lella Oum Keltoum. Toute la ville à présent le sait...
Les esclaves de Lella Meryem le racontèrent à des voisines.