Download PDF

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Snowball Earth wikipedia , lookup

Transcript
Journal of African Earth Sciences 44 (2006) 1–20
www.elsevier.com/locate/jafrearsci
Geological Society of Africa Presidential Review, No. 10
Evidence for the Snowball Earth hypothesis in the Arabian-Nubian
Shield and the East African Orogen
R.J. Stern
a
a,*
, D. Avigad b, N.R. Miller a, M. Beyth
c
Geosciences Department, University of Texas at Dallas, Box 830688, Richardson, TX 75083-0688, USA
b
Institute of Earth Sciences, Hebrew University of Jerusalem, Jerusalem 91904, Israel
c
Geological Survey of Israel, 30 Malkhei Yisrael Street, Jerusalem 95501, Israel
Received 7 April 2005; received in revised form 26 September 2005; accepted 12 October 2005
Available online 27 December 2005
Abstract
Formation of the Arabian-Nubian Shield (ANS) and the East African Orogen (EAO) occurred between 870 Ma and the end of the
Precambrian (542 Ma). ANS crustal growth encompassed a time of dramatic climatic change, articulated as the Snowball Earth
hypothesis (SEH). SEH identifies tremendous paleoclimatic oscillations during Neoproterozoic time. Earth’s climate shifted wildly, from
times when much of our planet’s surface was frozen to unusually warm episodes and back again. There is evidence for four principal
icehouse episodes: 585–582 Ma (Gaskiers), 660–635 Ma (Marinoan), 680–715 Ma (Sturtian), and 735–770 Ma (Kaigas). Evidence
consistent with the SEH has been found at many locations around the globe but is rarely reported from the ANS, in spite of the fact that
this may be the largest tract of Neoproterozoic juvenile crust on the planet, and in spite of the fact that Huqf Group sediments in Oman,
flanking the ANS, record evidence for Sturtian and Marinoan low-latitude glaciations. This review identifies the most important evidence
preserved in sedimentary rocks elsewhere for SEH: diamictites, dropstones, cap carbonates, and banded iron formation (BIF). Expected
manifestations of SEH are integrated into our understanding of ANS and EAO tectonic evolution. If Kaigas and Sturtian events were
global, sedimentary evidence should be preserved in ANS sequences, because these occurred during an embryonic stage of ANS evolution, when crustal components (island arcs, back-arc basins, and sedimentary basins) were mostly below sea level. Previous SEH investigations have been largely reconnaissance in scope, but potentially diagnostic sedimentary units such as diamictites, marine carbonates
with d13C excursions and banded iron formations are reported from the ANS and are worthy of further investigation. Collision and
uplift to form the EAO destroyed most marine sedimentary basins about 630 Ma ago, so evidence of Marinoan and Gaskiers glaciations
will be more difficult to identify. Several post-accretionary Neoproterozoic sedimentary basins in Arabia may preserve sedimentary evidence but such evidence has not been documented yet. The Huqf Group of Oman contains sedimentary evidence for the Marinoan glaciation but no evidence that the Gaskiers glaciation was significant in this part of the world. Deep erosion at 600 Ma throughout the
northern ANS and EAO may be related to Marinoan continental glaciation, which may have accomplished much of the cutting of the
ANS peneplain, but final shaping of the peneplain took place over the next 60 million years.
African geoscientists can contribute to our understanding of Neoproterozoic climate change through careful field studies, and the
international geoscientific community interested in Neoproterozoic climate change should pay attention to evidence from the ANS.
Future investigations should include knowledge of the SEH and its controversial aspects, in addition to the greater plate tectonic setting
of the ANS.
Ó 2005 Elsevier Ltd. All rights reserved.
Keywords: Neoproterozoic; Snowball Earth; Arabian-Nubian Shield; East African Orogen
*
Corresponding author.
E-mail address: [email protected] (R.J. Stern).
1464-343X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jafrearsci.2005.10.003
2
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
Contents
1.
2.
3.
4.
5.
6.
7.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
The Snowball Earth hypothesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
The Arabian-Nubian Shield and the East African Orogen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Diagnostic evidence for Neoproterozoic glaciation and post-glacial warming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4.1. Dropstones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4.2. Diamictites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.3. Cap carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.4. Banded iron formations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.5. Paleomagnetic evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Expected manifestations of Neoproterozoic glaciations in the ANS and EAO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
The record of glaciation in the ANS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
6.1. Evidence for Kaigas (735–770 Ma) glaciation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
6.2. Evidence for Sturtian (680–715 Ma) glaciation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
6.3. Evidence for Marinoan (635–660 Ma) and Gaskiers (582–585 Ma) glaciations . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1. Introduction
There are three grand and intertwined Neoproterozoic
(1000–554 Ma) themes. First is the evolution of increasingly complex life, whereby a biosphere characterized by
single-celled organisms at the beginning of Neoproterozoic
time evolved to be dominated by more complex multicellular organisms at its end (Knoll, 2003). The second grand
theme is a great Wilson/Supercontinent Cycle, which began
with the rupture of the Rodinian supercontinent and formation of new oceanic realms. As these fragments dispersed, oceanic realms closed and a new supercontinent
was generated from the shards of Rodinia. The closing
ocean generated great fringing arcs and oceanic plateaus,
and these were swept up in front of the advancing continental fragments and incorporated into the new Gondwana
supercontinent. The third grand Neoproterozoic theme
concerns the tremendous paleoclimatic oscillations that
have become the focus of the ‘‘Snowball Earth hypothesis’’. Earth’s climate seems to have shifted wildly, from
times when perhaps the entire planet’s surface was frozen,
quickly turning to sweltering greenhouses, and back again
(Hoffman et al., 1998; Hoffman and Schrag, 2002).
These biological, tectonic, and climatic themes are
related and present a spectacular example of global change
on the maturing Earth. It is a wonderfully interdisciplinary
effort that seeks to understand how Neoproterozoic tectonics and life affected climate, and how Neoproterozoic tectonics and climate influenced biological evolution. A large
part of the research focuses on isotopic proxies of the Ccycle to discern how and why atmospheric concentrations
of greenhouse gasses CO2 and perhaps CH4 varied. Atmospheric concentrations of these gasses were important con-
trols of Phanerozoic climate (Royer et al., 2004), and
should also have been important for the Neoproterozoic.
Hypotheses of how Neoproterozoic life and climate
interacted are developing rapidly. One possibility is that
proliferating photosynthetic life increased atmospheric
oxygen as it drew down atmospheric CO2, leading to cooling and allowed the development of a protective ozone
layer. In turn, global cooling and warming cycles may have
stimulated evolution by alternately stressing the biosphere
and providing warm, shallow water ecosystems when ice
melted and sea level rose. Continental dispersal allowed
multiple ecological environments on different continental
shelves to develop in isolation, and so to stimulate evolution (Valentine and Moores, 1974).
It is less clear how Neoproterozoic tectonics affected climate because so many explanations are possible. Continental configurations during Phanerozoic time exert important
controls on climate, with harsher climates during times of
supercontinent assembly and warmer, more humid climates
during times when continents were dispersed and sea level
is high (Worsley et al., 1986; Veevers, 1990). Similar controls must also have been important during Neoproterozoic time, although we are not yet confident that we
understand continental configurations during this time.
The supercontinent cycle exerted other controls on climate
as well. Continental fragments produced from the breakup
of Rodinia clustered at low latitudes, where in theory
intense chemical weathering associated with a very rainy
tropical climate absorbed atmospheric CO2, while organic
matter was buried near river deltas (Evans, 2000; Donnadieu et al., 2004). Weathering of flood basalts that were
erupted in association with the break-up of Rodinia may
have drawn down atmospheric CO2 to start the first
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
glaciation (Goddéris et al., 2003). An increase in explosive
volcanism is linked to Pleistocene glaciation in the northern
hemisphere (Prueher and Rea, 2001), and an increase in arc
volcanism possibly triggered Neoproterozoic glaciations
(Stern, 2005). Massive release of methane as a result of
destabilizing methane hydrates in the continental shelves
has been suggested as an important contribution to rapid
warming following glaciation (Martin et al., 2001).
The spectacular nature of global climate variations interpreted from Neoproterozoic lithostratigraphic sequences
and attendant isotope records has resulted in catastrophic
explanations. It is suggested that climatic oscillations may
have been forced by radical changes in Earth’s pole of rotation (Williams, 1975; Hoffman, 1999; Evans, 2003) or as a
result of fundamental changes in Earth’s tectonic style
(Stern, 2005). It may be some time before all of the possible
explanations are advanced and tested, but it is clear that
understanding Neoproterozoic tectonics and paleogeography will be essential for understanding interactions
between Neoproterozoic life, climate, and tectonics. The
interdisciplinary nature of the effort to understand interactions between the solid earth, hydrosphere, and biosphere
will surely provide us with a more robust understanding
of how the Earth system operates.
This essay explores the extent to which evidence of Neoproterozoic Snowball Earth events, especially evidence for
marine ice cover and continental glaciation, are preserved
in the Arabian-Nubian Shield (ANS) of NE Africa and
western Arabia. The ANS formed during Neoproterozoic
time, and the early part of its evolution was associated with
volcanism and sedimentation below sea level, where sedimentary evidence of Snowball Earth episodes should be
preserved. A range of mostly marine environments characterized the embryonic ANS, from shallow-water shelves to
the abyssal seafloor, and some deposits should record the
extreme climatic variations observed for this time period
elsewhere around the globe. There are extensive tracts of
ophiolites in the ANS, crustal relicts of the Neoproterozoic
deep ocean. Sediments deposited on ANS ophiolites should
record how the deep ocean behaved during and between
Snowball Earth events. In spite of these opportunities,
there has been litttle effort to use the ANS to investigate
Neoproterozoic climate. We know of only two reports that
explicitly identify rock sequences potentially pertaining to
Snowball Earth episodes in the region (Beyth et al., 2003;
Miller et al., 2003).
This review is intended to stimulate research of the SEH
in the ANS in three ways. First, we hope to inform geoscientists studying Neoproterozoic rocks in the ANS, so that
they can help look for the evidence. Second, we hope to
draw the attention of geology students to this exciting area
of cross-disciplinary and international research, especially
students in those nations that have ANS outcrops where
careful field studies allow important contributions to be
made at relatively low cost. Finally, we hope to make the
international community aware that the ANS is a promis-
3
ing area for understanding Neoproterozoic global change,
and to encourage this community to extend their studies
to the ANS.
The organization of this paper is designed for each of
these target audiences. In the following sections, we outline
the Snowball Earth hypothesis and the evolution of the
Arabian-Nubian Shield. Then, we discuss what sorts of evidence should be sought in sedimentary rocks. Finally, we
use what we know about the timing of Neoproterozoic glaciations and tectonic evolution of the ANS to discuss what
is already recognized and what is likely or unlikely to be
preserved if the extreme climatic events inferred for the rest
of the world affected the evolving ANS.
A final caveat to the reader: many aspects of the Snowball Earth hypothesis are controversial and there is a developing array of competing models to best account for the
vital physical and chemical evidence. Most scientists agree
that Neoproterozoic time was characterized by remarkable
climate variations but details of this are still being resolved,
for example whether or not the oceans were completely icecovered and whether or not glaciations were globally synchronous (see for example Young, 2004; Williams, 2004).
Those studying ANS exposures should keep an open mind
about what is observed and how this is best interpreted.
2. The Snowball Earth hypothesis
The Snowball Earth hypothesis (SEH) focuses on evidence that Earth experienced several cycles of unparalleled
climatic fluctuations during Neoproterozoic time and understanding why this happened. Conditions alternated rapidly
between ‘icehouse’ (intense, perhaps global ice cover) and
‘greenhouse’ (globally warm) conditions (Evans, 2000).
Hot and cold climatic swings may have been brief, perhaps
a few millions of years long, and these separated by much
longer intervals of more temperate climate. It is controversial whether or not the entire Earth ever became ice-covered,
but it is accepted that Neoproterozoic glaciations were more
extensive than late Cenozoic ‘‘ice ages’’. Paleomagnetic evidence indicates that much glacial debris was deposited in
low-latitude settings (Harland, 1964; Evans, 2000; Hoffman
and Schrag, 2000, 2002; Kilner et al., 2005). Glacial episodes
were followed by rapid warming, as evidenced by deposition
of thick sequences of ‘cap-carbonates’ above diamictites
deposited by ice-rafting or by other modes of periglacial sedimentation. These limestones and dolomites may have been
deposited very rapidly, as the warming ocean became supersaturated in carbonate.
Kirschvink (1992) identified three principal ways to test
the hypothesis. First, glacial units around the globe should
be more or less synchronous. Efforts continue to determine
the ages of glacial beds, and it may be several years before
we know how many glacial episodes there were and the
extent to which these were globally synchronous. It is often
difficult to determine the age of these deposits because units
containing datable materials, such as interbedded ash beds
4
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
with zircons, are often not present. In this regard, the
persistence of Neoproterozoic volcanic activity in the
Arabian-Nubian Shield should be advantageous for determining the ages of pertinent units as these are identified in
the ANS. The best modern evaluations indicate four principal icehouse episodes (Hoffman and Schrag, 2000, 2002;
Condon et al., 2005) and we use the glacial episode terminology and numeric age constraints compiled by MacGabhann (2005): 582–585 Ma (Gaskiers, also called
Varanger), 635–660 Ma (Marinoan), 680–715 Ma
(Sturtian) and 735–770 Ma (Kaigas). The formally
defined base of the Ediacaran Period (630–542 Ma) is
located at the contact of Marinoan glacial rocks and overlying Ediacaran cap carbonates in Enorama Creek, Australia (Knoll et al., 2004), thus three of the four glacial events
happened during the Cryogenian Period (850–630 Ma) and
one during the Ediacaran Period.
Geochronological studies are rapidly refining our understanding of when major glaciations occurred, although
presently the Gaskiers and Marinoan glaciations are more
tightly constrained than are the older the Sturtian and Kaigas glaciations. Marinoan diamictites in Namibia are dated
by U–Pb zircon techniques (ash interbedded at the top of
the Ghaub diamictite) at 635.5 ± 1.2 Ma (Hoffmann
et al., 2004). This age for the Marinoan event is supported
by two U–Pb zircon ages for ash beds from just above the
Nantuo Tillite (2.3 m above: 635.2 ± 0.6 Ma; 9.5 m above:
632.5 ± 0.5 Ma; Condon et al., 2005). In contrast, the age
of the Sturtian glaciation may have taken longer or consisted of multiple episodes, from 670 to 725 Ma. Glacial
deposits in Idaho, USA, are constrained using SHRIMP
U–Pb zircon techniques to have occurred between
709 ± 5 Ma and 667 ± 5 Ma (Fanning and Link, 2004),
whereas in Oman ash beds within the Ghubrah diamictite
yielded a U–Pb zircon age of 711.8 ± 1.6 Ma (Allen
et al., 2002).
Kirschvink (1992) also suggested that if the SEH was
broadly correct, then global icehouse/greenhouse events
should have produced similar deposits around the globe.
This is often found, in particular, where unusual carbonate
units abruptly overlie glacial successions. These are the
‘‘cap carbonates’’ discussed later.
Carbon-isotopic compositions of especially carbonate
rocks are crucial for characterizing and correlating these
deposits, particularly in locales where there is no zircon
geochronology. Carbon-isotope stratigraphy is uniquely
powerful for correlating Neoproterozoic calcareous sediments because the greatest changes in C-isotopic variations
in Earth history occurred in Neoproterozoic time (Fig. 1),
and icehouse–greenhouse deposits are accompanied by
Fig. 1. Secular variation in carbon (A) and strontium (B) isotopic composition of shallow marine carbonates, showing the relative timing of ArabianNubian Shield (ANS) basement rocks (shaded) to Neoproterozoic ‘‘Snowball Events’’. Carbon-isotope excursions correspond to Gaskiers, Marinoan, and
Sturtian glaciations. Figure modified from Miller et al. (2003).
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
unusual swings in the carbon-isotopic composition of carbonate sediments. Carbonate sediments proxy for the isotopic composition of seawater, and provide a robust
record of the changing C-isotopic composition of the Neoproterozoic atmosphere and hydrosphere. Neoproterozoic
carbon-isotope cycles are thought to mark when earth’s climate changed from ‘icehouse’ to ‘greenhouse’ conditions,
and the number of Neoproterozoic C-isotope excursions
suggest that there may be six Snowball Earth events
(Fig. 1).
The basic principles that relate C-isotope excursions and
Snowball Earth events reflect differing amounts of effectively buried dead biosphere and different contributions
from a homogeneous mantle. The isotopic composition
of carbon in the atmosphere and the ocean in which carbonate forms is controlled by equilibrium between reservoirs of inorganic and organic C, and how much of the
latter is effectively buried. The carbon-isotope balance is
preserved as relative variations between 13C and 12C
retained in carbonate sedimentary rocks, measured as
d13C relative to an isotopic standard; typically PDB Cretaceous belemnite carbonate; an arbitrary seawater proxy is
set at 0&). Metabolic processes most effectively integrate
light carbon into biomass, so proliferating life depletes
the ocean in 12C (and has negative d13C). This enriches
the CO2 and bicarbonate in seawater in heavier 13C (with
positive d13C). Fractionation of C-isotopes in the atmosphere and hydrosphere (and thus in carbonate rocks)
can become extreme if a significant proportion of this
organic matter is buried as organisms die and are removed
from the C-cycle, similar to what has been noted on a smaller scale for Cretaceous ‘black-shale’ events (e.g., Kuypers
et al., 2002). Extreme fractionation is also possible for
stratified oceans (mentioned later) or changes in amounts
of biomass produced via photosynthetic versus chemoautotropic metabolic pathways (Hayes et al., 1999).
There is very active research to resolve the Neoproterozoic carbon isotope record and whether or not fluctuations
correspond to glacial episodes (e.g., Halverson et al., 2005).
There are several explanations for the causes of these variations, particularly how these may be related to changes in
the biosphere and hydrosphere. A detailed review of these
hypotheses is beyond the scope of this paper but the interested reader will find a good overview of carbon and other
isotopic systematics in Ohmoto (2004).
According to the most popular accounts (Kirschvink,
1992; Hoffman et al., 1998; Hoffman and Schrag, 2002) global glaciation is thought to have decimated biological
activity, freeing much light carbon. This would have been
recorded as lower d13C in carbonate during times of diminished biological activity. Weakening of the biosphere is
thus manifested as lighter C-isotopic composition of the
atmosphere and hydrosphere, approaching the isotopic
carbon of carbon escaping from Earth’s mantle due to volcanic activity (d13C 6&). Global glaciation is thought
to have ended when atmospheric CO2 increased sufficiently
that warming due to this ‘greenhouse gas’ overcame the
5
effect of cooling due to the high albedo of an ice-covered
world. Destabilization of methane hydrates held in shelf
sediments and massive release of methane—another greenhouse gas—may have also been important for ending Neoproterozoic glacial episodes (Kennedy et al., 2001). Global
warming led to rapid deglaciation, accelerated biological
activity and renewed burial of isotopically light carbon,
and seawater returned to normal, heavier carbon-isotopic
compositions.
Finally, Kirschvink (1992) suggested that deepwater
deposits of the Neoproterozoic Ocean should also record
the extreme climatic events, particularly in the form of
banded iron formations (BIFs). Neoproterozoic BIFs may
reflect re-oxygenation of the oceans following anoxia caused
by a global ice sheet, as discussed in Section 4.4 below. Neoproterozoic carbonate sediments were likely deposited only
in shallow-water shelf environments because calcareous
plankton did not diversify until the middle Mesozoic
(Ridgewell et al., 2003). Thus, snowball event carbonates tell
us little about the deep ocean. BIF may better record how
Neoproterozoic climate change affected the deep ocean.
3. The Arabian-Nubian Shield and the East African
Orogen
The Arabian-Nubian Shield (ANS) outcrops around the
Red Sea in NE Africa and W. Arabia as a result of uplift
and erosion on the flanks of the Red Sea in Oligocene
and younger times (Fig. 2A). The ANS may be the largest
tract of juvenile continental crust of Neoproterozoic age on
Earth (Patchett and Chase, 2002). ANS evolution can be
simplified into four stages, as shown in Fig. 3. This accompanied a supercontinent cycle that defined Neoproterozoic
tectonics, beginning with the breakup of the end-Mesoproterozoic supercontinent Rodinia in the early Neoproterozoic (Hoffman, 1999). ANS juvenile crust was generated
around and within the Mozambique Ocean (Stern, 1994).
Arcs and oceanic plateaux were swept up as the Mozambique Ocean closed. The tectonic cycle culminated in a
protracted collision between what has come to be known
as East and West Gondwana (each of which may have
been only partially consolidated, e.g. Alkmim et al.,
2001; Collins and Pisaversky, 2005), resulting in the East
African Orogen (EAO) and a supercontinent ‘Greater
Gondwana’ or ‘Pannotia’ at the end of Neoproterozoic
time (Fig. 2B). In reconstructed Gondwana, the EAO
extends from the Mediterranean (Tethys) southward along
the eastern margin of Africa and across East Antarctica
(Stern, 1994; Jacobs et al., 2003).
Exposed crust of the EAO changes dramatically along
its length. The northern EAO, the ANS, is dominated by
exposures of juvenile Neoproterozoic crust, especially
greenschist-facies supracrustal and abundant intrusive
rocks. Fig. 4 presents the Nd-model age summary of
Stern (2002) for the EAO and ANS. This provides an
isotopic proxy for Neoproterozoic paleogeography, with
those regions characterized by Neoproterozoic model ages
6
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
Fig. 2. (A) The Arabian-Nubian Shield. Stars denote regions with evidence for Snowball Earth deposits. Modified after Miller et al. (2003). Location of
Fig. 10 is shown as stars labeled BIF. (B) and (C) Paleogeographic reconstructions of the Arabian-Nubian Shield as part of the End-Neoproterozoic
supercontinent at 580 Ma, from Meert and Torsvik (2004). (B) The high-latitude Laurentia option places the present-day eastern margin of Laurentia at
the south pole adjacent to the Amazonian and Rio Plata cratons at 580 Ma. Baltica has rifted from NE-Laurentia opening the Iapetus Ocean. (C)
Configuration in (B) is rotated to show an alternative configuration for the final stages of Gondwana assembly and closure of the Mawson Sea between
Australo-Antarctica and the rest of Gondwana.
largely forming in oceanic realms, with sedimentation and
volcanism in shallow to abyssal submarine environments,
compared to pre-Neoproterozoic Nd-model age regions
characterized by continental environments. This is an oversimplification, but Fig. 4 does emphasize the point that
the ANS is largely juvenile Neoproterozoic crust whereas
the southern EAO mostly formed from older continental
crust. The ANS largely escaped high-grade metamorphism because terminal collision occurred in the south,
allowing the EAO to escape northward (Bonavia and
Chorowicz, 1992; Abdelsalam and Stern, 1996). In con-
trast, the southern EAO (Tanzania and Madagascar) was
more intensely deformed and metamorphosed and contains
abundant granulite-facies rocks, many with pre-Neoproterozoic protolith ages (Kröner et al., 2003). These rocks represent the intensely overprinted margins of the colliding
continents and testify to greater thickening of the crust in
the south and correspondingly deeper erosion. The ANS
and EAO evolved together, especially in their later stages,
but because its early development mostly took place below
sea level, the ANS should contain a better sedimentary
record of events predating terminal collision.
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
7
Fig. 3. Stages in the tectonic evolution of the Arabian-Nubian Shield and the East African Orogen, modified after Stern (1994).
Formation of the ANS began as Rodinia began to disintegrate between 900 and 800 Ma, as inferred from the oldest
(870 Ma) juvenile Neoproterozoic rocks in the ANS
(Stern, 1994) and from events in eastern Gondwana
(Cawood, 2005). ANS crust was largely generated at circum-Mozambique Ocean intraoceanic arc systems (Tadesse
et al., 1999; Woldehaimanot, 2000). Oceanic plateaux may
also have formed above mantle plumes within the great
ocean; these would have been accreted and added to the
mix of juvenile crust (Stein, 2003). Juvenile arc and plateau
terranes collided and were welded into larger tracts of juvenile crust as the Mozambique Ocean closed, forming arc–arc
sutures, composite terranes, and, ultimately, the ANS
(Johnson and Woldehaimanot, 2003). ANS juvenile crust
was trapped as the ocean closed between fragments of East
and West Gondwanaland, ultimately nestling within the
630 Ma terminal collision zone of the EAO (Meert,
2003). Convergence between fragments of E and W Gondwana continued and the EAO was further deformed during the
last 80 million years of the Precambrian (Veevers, 2003).
Deformation included strike-slip shear zones and tectonic
collapse structures in the northern EAO (Egypt, Sudan,
and northern Arabia), formation of N-trending upright tight
folds and shear zones in the central EAO (Ethiopia, Eritrea,
and southern Arabia), and formation and uplift of highgrade gneisses and granulites in the southern EAO (Abdelsalam and Stern, 1996). The most intense collision was in the
southern EAO, which must have had the thickest crust, highest mountains, and the deepest erosion. Pannotia or Greater
Gondwana began to break up almost as soon as it formed at
the end of Neoproterozoic time, The supercontinent continued to shed microcontinents into especially Asia all
through Paleozoic and early Mesozoic time, with the core
of Gondwana finally rupturing in Late Jurassic time.
4. Diagnostic evidence for Neoproterozoic glaciation and
post-glacial warming
4.1. Dropstones
A dropstone is an isolated, oversized clast in laminated
sediments that depresses the underlying laminae. The
dropstone may be draped by sediments. Most dropstones
form by debris falling from ice rafts (Fig. 5A), although
some Phanerozoic dropstones are transported as kelp
holdfasts or by floating tree roots. Such explanations cannot explain Neoproterozoic dropstones. Dropstones could
also potentially result from coarse material produced by a
meteorite impact falling back to Earth, but such clasts are
likely to be highly shocked. Volcanic bombs could be
thrown out a few kilometers from a violent eruption
and land on laminated sediments. Other than these caveats, Neoproterozoic dropstones indicate glaciation, and
that the ice mass carrying coarse rocks floated on water
(Fig. 5A). When they are not deformed, dropstones are
usually angular and marked by glacial scratches and
grooves, and they deform the underlying sediment
whereas the overlying sediment drapes the dropstone
(Fig. 5B).
Recognition of dropstones in sedimentary successions is
most convincing where there has been little deformation.
Deformation in the ANS renders the identification of dropstones much more difficult, and we know of no convincing
occurrences here. Dropstones are reported from the Huqf
Group in NE Oman, where Allen et al. (2004) note that
the Fiq Formation contains four horizons of proximal
and distal marine glacial deposits with dropstones and
other evidence of ‘rainout’ of debris from icebergs in a marine environment.
8
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
Fig. 4. Pre-Jurassic configuration of the East African Orogen in Africa
and surrounding regions, modified from Stern (2002). Regions referred to
in text: Egypt (Eg); Sudan (Su); Sinai–Israel–Jordan (SIJ); Afif terrane,
Arabia (Aa); Rest of Arabian Shield (Ar); Eritrea and northern Ethiopia
(En); Southern Ethiopia (Es); Eastern Ethiopia, Somalia, and Yemen (Ee);
Kenya (K); Tanzania (T); Madagascar (M). Numbers in italics beneath
each region letter are the Nd-model ages. Regions of juvenile crust have
Nd model ages of 1.0 Ga or less; these regions likely existed below sea
level during Kaigas (770–740 Ma) and Sturtian (750–700 Ma) glaciations. Regions with Nd model ages >1.0 Ga may or may not have been
below sea level at these times. The entire EAO was topographically
elevated following collision, beginning about 630 Ma. The EAO was
probably subject to extreme continental glaciation during Marinoan
(630–600 Ma) and glaciation. Suture labeled ‘BN’ is the Bir Umq–
Nakasib suture.
4.2. Diamictites
The correct identification and interpretation of diamictites is critical for evaluating the record of ancient glaciations. Flint et al. (1960) introduced the term diamictite
for lithified, poorly-sorted, non-calcareous terrigenous sedimentary rocks, from the Greek diamignymi meaning ‘to
mingle thoroughly’. Diamictites are poorly sorted polymict
conglomerates and breccias and contain a wide range of
clast sizes and shapes. These can form in many ways—
for example as debris flows and as ejecta blankets from
meteorite impacts, as well as due to the actions of glaciers
(Eyles and Januszcak, 2004). Tectonic and volcanic activity
that formed the ANS provided many opportunities to produce diamictites without glaciers. Thus, the identification
of a sedimentary unit as a diamictite does not require the
interpretation of glacial activity, but it does focus attention
on and hopefully result in more careful scrutiny of the unit.
Diamictites that result from glacial activity encompass a
variety of peri- and subglacial environments, including terminal and lateral moraines, deposited in both marine and
subaerial environments. An abundance of angular clast
shapes supports an interpretation of glacial origin, but
these may be common only in marine sedimentary environments. Rounded cobbles can also result from terrestrial
glaciation, because rock fragments deposited by glaciers
must be transported by melt streams from upland moraines
to lowstanding basins. Diamictites deposited in marine
environments are more likely to be preserved, simply
because these can be more deeply buried and protected
from erosion than those deposited above sea level. The
most unequivocal evidence for a glacial origin of diamictite
is the identification of scratch marks or striations on clasts
or recognition of dropstones, but these criteria are complicated in the ANS because deformation has obscured many
primary sedimentary structures. Clast lithology and age
may be better ways of determining whether or not ANS
diamicties are glaciogenic. Those that may be glacial in origin should contain a large diversity of lithologies, shapes,
and sizes (polymict conglomerate or heterolithologic
breccias, depending on clast shape). Because of uncertainty
surrounding the origin of ANS diamictites and the obliteration of delicate sedimentary structures by deformation,
breccias that contain only a single clast type (volcanic, plutonic, or sedimentary) are difficult to demonstrate to have
formed glacial origin. Some monomict diamictites may
have been deposited by glacial action, but the recognition
of limited provenance makes it more likely that such deposits formed by non-glacial debris flows. ANS diamictites
with mostly volcanic clasts in particular are not convincing
evidence of glacial activity. Identification of clasts that
have no local provenance is a strong argument for long distance transport by floating ice. A potential example is the
Atud Conglomerate of E. Egypt, where some granitic cobbles yield pre-Neoproterozoic U–Pb zircon ages and are
inferred to have been transported hundreds of kilometers
from sources that now lie west of the Nile (Dixon, 1981).
No breccias inferred to have formed by meteorite impact
are yet reported from Neoproterozoic units of the ANS.
Neoproterozoic sediments of the Huqf Supergroup in N.
Oman (J. Akhdar area) provide a valuable lesson in glacial
activity and diamictite sedimentation in a region that is
now near the ANS. The Abu Maarah Group contains
are two important diamictite horizons that are associated
with distinct glacial episodes (Le Guerroué et al., 2005a):
the Ghubrah Formation (with a diamictite ash date of
711 ± 6 Ma; Allen et al., 2002)), and the overlying Fiq
Formation (constrained between 712 and 544 Ma but not
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
9
Fig. 5. Formation of dropstones in a near-glacial aqueous environment, modified after Hladil (1991). (A) Continental glacier erodes rock and transports
this to a lake or to the sea. Icebergs form as glacier calves and drifts with the current, carrying embedded rocks. Melting iceberg eventually releases
embedded rocks, which fall to lake or sea floor, impacting sediments and forming dropstones. (B) Results of dropstone experiments where the
consolidation of the substrate is varied, from poorly to very lithified. Note that the intensity and depth of deformation of substrate increases with
decreasing substrate lithification, as a function of the diameter of the dropstone (D). In general, the best diagnostic feature to identify a dropstone is the
penetration of underlying lamina by the clast and simultaneous lack or insignificance of compactional deflection of lamina above the clast.
yet radiometrically dated; Kilner et al., 2005). Ghubrah
and Fiq formations are separated by an angular unconformity and the older Ghubrah Formation is highly deformed.
The Ghubrah Formation is broadly Sturtian and the Fiq
Formation is widely regarded as Marinoan in age (Burns
and Matter, 1993; Brasier et al., 2000; Le Guerroué et al.,
2005a) , although it could be older (Kilner et al., 2005).
The Ghubrah Formation is dominated by several hundred
meters of diamictite characterized by: (1) poor stratification; (2) unsorted randomly dispersed clasts of diverse size
(up to 1 m; usually <10 cm) and lithology (crystalline and
metamorphic rocks, mafic and felsic volcanics and sedimentary rocks) that comprise about 15% of the diamictite;
and (3) unsorted silty-shaly or sandy matrix. The Ghubrah
is not marked by cyclical sedimentation, and lacks the shallow water elements and thick turbidites characteristic of
the younger Fiq (P. Allen, pers. comm., 2005). Some clasts
are striated and some are interpreted as dropstones. The
diamictite is interpreted by Le Guerroué et al. (2005a,b)
as due to ice-rafting. Siltstone units (up to 10 m thick)
are thought to have been deposited in a marine environ-
ment during a time of reduced influence from ice-rafting
(Le Guerroué et al., 2005a).
The Fiq Formation is 1.5 km thick. Diamictites probably comprise 20% or less of the succession. The rest of the
Fiq Formation is turbidites, shales, slumped silts and
shales, debris-flows, and wave-ripped sandstones. The Fiq
Formation contains diamictite horizons up to 30 m thick,
which are somewhat stratified (clast concentrations or
matrix grain size) when proximal. The Fiq Formation is
divided into two facies associations (Leather et al., 2002;
Allen et al., 2004): (1) proximal and distal glaciomarine;
and (2) non-glacial gravity flow and shallow marine. Clast
sizes range from 1 cm to 2 m, and some are facetted or striated. Leather et al. (2002) identified seven stratigraphic
cycles in the Fiq Formation and interpreted these to indicate when ice sheets advanced and retreated. Recently identified magnetic reversals in the Fiq (and overlying Hadash
cap dolomite) indicate that the glacial to interglacial
climatic transition took place over an extended time
period (perhaps >105–106 yr; Kilner et al., 2005). Of further climatic significance, paleolatitude estimates from
10
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
paleomagnetic measurements place the Oman region at
13°, indicating tropical glaciers at sea level.
The Shuram Formation of the overlying Nafun Group
has a major negative carbon isotope excursion that is
correlated with the Gaskiers (Varanger) glaciation (Le
Guerroué et al., 2005a; Halverson et al., 2005), although
no glacial sediments or evidence of glacial erosion are recognized. Le Guerroué et al. (2005b) now think there is no
correlation at all between the Shuram C isotope shift and
the Gaskiers glaciation. They think that the Shuram shift
lasted much longer than any glaciation, starting at about
600 Ma and ending around 550 Ma. Regardless of the
duration of the Shuram C-isotope shift, the apparent
absence of an unconformity or glacial sediments suggests that the Gaskiers glaciation was relatively mild near
Oman.
The two main glacial deposits in Oman seem to correspond to Sturtian and Marinoan global episodes. The
Fiq (Marinoan) glaciation in Oman was characterized by
repeated advances and recessions, whereas Sturtian deposits were less variable (P. Allen, pers. comm. 2005). If the
relative thickness of diamictite beds correlates with the
intensity of the glaciation that formed them, it may be that
the Ghubrah (Sturtian) glaciation was more intense than
the Fiq (Marinoan) glaciation in the region around Oman.
At present, however, diamictite thickness is not related simply to glacial intensity or duration.
4.3. Cap carbonates
Many Neoproterozoic glacial deposits are capped by
layers of pure dolostone and limestone, known as ‘cap carbonates’ (Hoffman and Schrag, 2002). Cap carbonates are
typically thicker than underlying glacial beds, and these
differences may reflect different sedimentation rates. Glacial sediments may have been deposited over millions of
years. Cap carbonates are traditionally thought to have
been deposited in only a few tens of thousands of years,
but recent discovery of paleomagnetic reversals in some
cap carbonate sequences (Oman, Kilner et al., 2005;
Amazon craton, Trindade et al., 2003) suggests a longer
duration (perhaps >105–106 yr ). Cap carbonates are especially paradoxical because they indicate an abrupt change
from glacial to apparently tropical conditions, and there
is a general perception that cap carbonates reflect a global
greenhouse climate. Cap carbonates are typically depleted
in 13C in the lower beds and rebound in younger cap carbonate beds to positive d13C values indicating biological
fractionation becomes increasingly important.
Several physical characteristics of cap carbonates are
commonly noted. As noted by Shields (2005), cap carbonates are usually thin (typically <5 m, but up to 27 m thick)
and uniform deposits of pale pink to buff microcrystalline
dolomite, with minor siliciclastic content. They are often
laminated on a cm-scale, seldom preserving primary calcitic textures, and may show graded (reverse and normal)
bedding. Sheet cracks, doming and brecciation associated
with isopachous dolomite cementation are common, as
are associations with high energy deposits (i.e., hummocky
cross-stratification and giant wave ripples; formerly interpreted as ‘‘pseudo-tepee’’ structures). This dolostone base
is usually laterally extensive, but sometimes discontinuous,
and may be overlain by transgressive shales, siltstones or
thick post-cap limestones. Many localities preserve evidence of post-cap dolostone seafloor precipitation as seafloor aragonite fans and barite. Stromatolitic carbonates
are often noted in the post-cap dolostone sequence. It
should be noted that Marinoan cap carbonates are particularly renowned for these attributes and there is a substantially smaller body of data (Prave, 1999; Hoffman and
Schrag, 2002) concerning Kaigas and Sturtian cap
carbonates.
The origin and environmental conditions regulating
physical and chemical characteristics of cap carbonates
are among the more contentious aspects of competing
SEH models. All models associate deglaciation with
extreme increases in the alkalinity of seawater. Currently
there are four competing models: (1) overturn of a redoxstratified ocean (Knoll et al., 1986; Grotzinger and Knoll,
1995; Canfield, 1998; James et al., 2001); (2) extreme
chemical weathering due to supergreenhouse conditions
(Kirschvink, 1992; Hoffman et al., 1998; Hoffman and
Schrag, 2002; Higgins and Schrag, 2003); (3) massive oxidation of destabilized methane hydrates (Kennedy et al.,
2001); and (4) sudden formation and gradual dissipation
of a global meltwater plume that stimulated microbial
mediation of carbonate precipitation (plumeworld hypothesis; Shields, 2005). It is beyond the scope of this review to
systematically evaluate these, but each varies significantly in
the interpretation of carbon cycle dynamics from C-isotope
data.
The diagnostic association of cap carbonates above
glaciogenic sediments has not yet been reported from the
ANS, but there have been few deliberate searches thus
far. In Oman, where several examples of glacial diamictites
are documented, the older Ghubrah diamictites (Sturtian)
do not have a cap carbonate, but this may have been
removed by erosion (P. Allen, pers. comm.). As mentioned
earlier, only the uppermost diamictite of the possibly
Marinoan-aged Fiq Formation has a cap carbonate. This
is the dolostone of the <15 m thick Hadash Formation
(Allen et al., 2004). Another candidate is a thick sequence
of laminated dolomites and stromatolitic limestones of
the Tambien Group, N. Ethiopia (Beyth et al., 2003; Miller
et al., 2003). These have clear petrographic, textural, and
isotopic affinities with cap carbonates but have not been
found to overly glaciogenic diamictites, as discussed in a
later section.
Carbonate sedimentary rocks are not common in the
ANS, so when these are found investigators should examine the basal contact interval for previously recognized
cap carbonate features (e.g., Shields, 2005) as well as
underlying strata for evidence of glaciation. However,
investigators should also recognize the comparative lack
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
of cap carbonate documentation for pre-Marinoan
sequences that may be more prominent in the ANS.
4.4. Banded iron formations
BIF was commonly deposited during Late Archean and
Paleoproterozoic time but disappeared as oxygen concentrations in the atmosphere and oceans increased and dissolved Fe was removed from seawater (Rouxel et al.,
2005). Rising oxygen concentrations oxidized Fe2+ dissolved in seawater into insoluble Fe3+, which precipitated
and accumulated on the seafloor to form BIF. The oxidation of seawater Fe2+ was completed in Paleoproterozoic
time, so that BIFs are missing from the Mesoproterozoic
record. BIF reappeared in Neoproterozoic times in association with Snowball Earth events. Most Neoproterozoic
occurrences formed during the Sturtian ice age; only one
case is documented from a possible Marinoan-age glaciogenic sequence (Shields, 2005; Proust and Deynoux, 1994).
There are two general categories of BIF: Superior-type
and Algoma-type. Superior-type BIF is associated with
shelf sediments (quartzite, marble, etc.) whereas Algomatype BIF is associated with volcanic rocks and immature
sediments. Superior-type BIF require a global or at least
regional change in water chemistry to precipitate Fe,
whereas Algoma-type BIF may reflect more local oceanographic conditions and sources of Fe. Archean BIFs are
mostly Algoma-type, whereas Paleoproterozoic BIFs are
mostly Superior-type. Neoproterozoic BIF can have affinities to either Algoma- or Superior-type. A more detailed
review of BIF can be found in Trendall (2002).
Neoproterozoic BIF are an important argument for
SEH but are controversial. Kirschvink (1992) suggested
11
that covering the oceans with ice could isolate the deep
oceans from the atmosphere and thus lead to anoxia in
the deep ocean. Deep waters would become reducing, so
that Fe supplied from seafloor hydrothermal vents
remained as Fe2+ in solution. When the ice sheets melted,
the supply of oxygenated water to the deep sea resumed
and Fe2+ in solution oxidized to insoluble Fe3+, which precipitated out as BIF. This model is shown schematically in
Fig. 6. A second explanation for Neoproterozoic BIF calls
on glaciation of Red Sea rift-type basins. Deep, Fe-charged
anoxic brines in such basins would have precipitated Fe
oxides on being mixed with ‘‘normal’’ seawater as a result
of glacially driven thermal overturn (Young, 2002). This
kind of argument is especially convincing for Neoproterozoic Superior-type BIF and is less convincing for Neoproterozoic Algoma-type BIF, which is the kind of BIF most
likely found in the ANS. Nevertheless, a link with largescale glaciation seems required if evidence for glaciation
(e.g., diamictite, dropstone) is found in association with
Neoproterozoic BIF of either category.
Trendall and Blockley (2004, p. 421) warn: ‘‘The Snowball Earth hypothesis is at an early stage of testing, and the
emphasis placed by some authors. . .on the relationship
between rift-related mafic volcanism and some Neoproterozoic [BIFs] indicates that the evidence for a purely climatic control of their deposition is not yet definitive.’’
4.5. Paleomagnetic evidence
For the Snowball Earth hypothesis it is not only crucial
to demonstrate which deposits are best interpreted as glacial, but also to show evidence that such deposits were
formed at low to equatorial latitudes (Evans, 2000). This
Fig. 6. Model for formation of Neoproterozoic banded iron formations (BIF). (A) Snowball Earth: anoxic ocean. Ice covering ocean surface isolates
seawater from mixing with atmosphere, cutting off the source of oxygen. Oxidation of organic matter consumes oxygen dissolved in seawater, with the
result that seawater becomes anoxic and reducing. Iron introduced as Fe2+ at mid-ocean ridge hydrothermal vents remains in solution, causing a buildup
of Fe2+ in seawater. (B) Deglaciation: ocean ventilation. Melting of ice allows mixing of atmosphere and seawater, re-oxygenating seawater. Increased
oxygen concentrations in seawater oxidizes Fe2+ dissolved in seawater to Fe3+. This forms insoluble iron-oxides and is deposited as BIF on the seafloor.
Modified from Fig. 8 in http://www-eps.harvard.edu/people/faculty/hoffman/snowball_paper.html.
12
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
is best demonstrated with careful paleomagnetic measurements. Unfortunately, there have not yet been any paleomagnetic studies of potential Snowball Earth deposits in
the ANS.
5. Expected manifestations of Neoproterozoic glaciations in
the ANS and EAO
Our understanding of the tectonic evolution of the EAO
and ANS (Fig. 3) has implications for how the waxing and
waning of Neoproterozoic Snowball Earth episodes should
be preserved. Uncertainty about the number and timing of
Snowball Earth episodes has already been noted, but there
is nevertheless agreement that two major episodes occurred
prior to 700 Ma and two after 630 Ma. The older episodes (Kaigas and Sturtian) occurred while ANS crust
was still forming as arcs, back-arc basins, and oceanic plateaus around and within the Mozambique Ocean (Fig. 7).
There should have been many submarine sedimentary
basins available to collect the distinctive debris produced
by Kaigas and Sturtian glaciations and subsequent warming and reoxygenation episodes. Some of these may have
formed in shallow water of a few hundred meters depth,
on continental shelves, atop oceanic plateaus, and around
island arcs. Other deposits should have formed at abyssal
depths, as deep as the 2500–5000 m characteristic of the
modern seafloor, associated with backarc basins and intraoceanic forearcs, and on the floor of the Mozambique
Ocean itself. If the Snowball Earth episodes prior to
700 Ma affected the Mozambique Ocean and its periphery, evidence should be preserved in the ANS. These deposits would have been deformed during later accretion and
collision events, but distinctive ‘Snowball Earth’ sedimentary deposits should still be recognizable in parts of the
ANS.
It is less likely to find sedimentary deposits of Marinoan
and Gaskiers Snowball Earth events. The period after
650 Ma very likely witnessed increasing relief in the
ANS and EAO, as collisions between various fragments
in the Mozambique Ocean occurred, culminating in terminal collision between E. and W. Gondwana (Fig. 7). Most
of the ANS was probably above sea level by 630 Ma, with
the highest relief in the southern EAO. At the end of the
Neoproterozoic, the EAO may have rested near the south
pole (Fig. 2B), so if there were global glaciations, the region
is likely to have been covered with a thick continental ice
sheet. It is possible that glacial deposits of the Gaskiers
and Marinoan episodes could be preserved in deep graben
around the margins of the ANS, such as that preserving the
Huqf Supergroup in Oman. Marinoan and Gaskiers deposits could also be preserved in continental shelf deposits on
the northern flank of the End-Neoproterozoic supercontinent, such as may exist beneath Israel.
6. The record of glaciation in the ANS
Diamictites and other likely examples of Snowball
Earth deposits preserved in the ANS are discussed below.
These are presented in terms of the four episodes identified by MacGabhann (2005): 735–770 Ma (Kaigas),
680–715 Ma (Sturtian), 635–660 Ma (Marinoan), and
582–585 Ma (Gaskiers). This grouping and the age assignments are likely to change as studies advance around
the globe, but these episodes provide a useful framework
for the following observations. This discussion draws
heavily our understanding of the Huqf Group in Oman
for indications of how glacial episodes are likely to be
manifested in the ANS. There is also a rich record in
NW Africa, summarized by Evans (2000) that is also
instructive.
Fig. 7. Expected interactions of Snowball Earth episodes with evolving ANS and EAO. Column on right refers to Snowball Earth episodes generalized
from Hoffman and Schrag (2002). Column on left refers to events discussed in the text.
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
6.1. Evidence for Kaigas (735–770 Ma) glaciation
The best evidence for the earliest glacial sedimentation
in the ANS is found in deposits on the southern margin
of the Bi’r Umq–Nakasib Suture Zone of Sudan and Arabia (Fig. 4). This suture zone can be traced for more than
600 km, from the central Arabian Shield almost to the Nile
and is the major suture separating the northern and southern ANS (Johnson et al., 2003). Diamictites are found on
the southern flank of the suture at two widely separated
locations. In Arabia, Johnson et al. (2003) report that the
base of the 770 Ma Mahd Group unconformably overlies
the 816 ± 3 Ma Dhukhr batholith, indicating a significant
13
episode of erosion between 770 and 816 Ma (Fig. 8A).
This is the oldest unconformity documented within the
ANS. The Mahd Group rests directly on this unconformity
and while dominated by volcanic rocks, its base is defined
by a 1–5 m thick diamictite. The diamictite is matrix
supported, with a dark-grey, immature, arkosic matrix
that contains abundant, angular to sub-angular clasts (up
to 30 cm wide) of granitic and felsic volcanic rocks
(Fig. 8B). Johnson et al. (2002) noted that this diamictite
was ‘‘. . .conceivably deposited during a Neoproterozoic
glacial event.’’
A diamictite of similar age is found in the Meritri Group
in the Sudanese sector of the Bi’r Umq–Nakasib Suture, a
Fig. 8. Photographs of outcrops in the ANS with evidence for Kaigas (A,B) and Sturtian (C–E) Snowball Earth events. (A) Unconformity of 770 Ma
Mahd Group basal tillites on 806 Ma Dhukhar batholith, Saudi Arabia. Finger points to unconformity. (B) Mahd Group basal tillite, note angular clasts
of granitic rocks in dark matrix. (C) Tambien Group tillite, Negash synform, N. Ethiopia. (D) Atud conglomerate dropstone, Wadi Khuda area (SE
Egypt). Angular dropstone is composed of quartz porphyry. (E) Atud diamictite at Wadi Kareim, Egypt. The diamictite sits stratigraphically below
immature clastics and BIF and consists of blocks up to 2 m of quartzite, granite, granodiorite, metavolcanics, and pebble conglomerate in schistose matrix.
Scientists point at three of these blocks. (F) BIF in Wadi Dabbagh, Egypt (hand lens for scale). Dark, hematite-rich layers are interbedded with thinner
carbonate bands.
14
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
few km west of Port Sudan. Abdelsalam and Stern (1993)
infer that sediments along the SE side of the suture manifest a deformed passive margin sequence. This sequence
begins with the Arbaat volcanic Group, which is succeeded
stratigraphically and structurally sediments of the Salatib
and Meritri groups. The Arbaat volcanic Group yields a
U–Pb zircon age of 790 ± 2 Ma. Following suture-related
deformation, the Arbaat, Salatib, and Meritri groups were
intruded by granitic plutons as old as 754 ± 3 Ma (Stern
and Abdelsalam, 1998). This constrains the age of the
Salatib and Meritri sediments to younger than 790 Ma
and older than 754 Ma.
The Salatib Group consists of intercalated rhyolite, conglomerate, mudstone, wacke, quartzite, and carbonate sediments. The Meritri Group consists of (from oldest to
youngest): conglomerate, lithic wacke, and interbedded
limestone, red sandstone, and felsic tuff. Abdelsalam and
Stern (1993) infer an original thickness of 2 km. The conglomerate is polymict and matrix supported, and is better
identified as diamictite. Clasts vary greatly in size, from a
few cm up to a meter or more, and in composition. About
50% of the clasts are plutonic (granite, granodiorite, diorite), 35% are volcanic (rhyolite and ignimbrite), and
15% are sedimentary (carbonates and subordinate clastic
rocks). This diamictite is succeeded by a lithic wacke with
sedimentary structures indicating transport from SE to
NW. Given the similarity of ages of Meritri and Mahd
Group diamictites, these may be correlatable and provide
evidence for the Kaigas glaciation in the ANS. The Meritri
Group diamictite should be investigated by a sedimentologist with appropriate expertise.
6.2. Evidence for Sturtian (680–715 Ma) glaciation
Evidence for Sturtian glaciation relatively close to the
ANS is found in the Huqf Supergroup of SE Oman, where
the basal Ghubrah Formation contains thick glaciogenic
diamictite. Tuffaceous wackes interbedded with the diamictite yielded a U–Pb zircon age of 723 +16/10 Ma (Brasier
et al., 2000). Less well-dated evidence of broadly Sturtian
‘Snowball Earth’ deposits are found in Ethiopia, Eritrea,
Egypt, and northern Arabia. In Ethiopia, a deformed
metasedimentary unit known as the Tambien Group
contains evidence of a Sturtian glaciation (Fig. 9). The
Tambien Group is mostly carbonate, but in the Negash
synform it consists of a thick section of carbonates capped
by a distinctive polymict diamictite interpreted to be glacial
in origin (Miller et al., 2003; Beyth et al., 2003; Fig. 8c). In
thin section, the diamictite contains clasts of felsic volcanics, fine-grained carbonates, low-grade semipelitic metasediments. In the field we also saw red granite, black
limestone, pegmatite quartz and chert clasts up to 5 cm
Fig. 9. Location of Tambien Group exposures in Ethiopia and Eritrea (modified after Beyth et al., 2003).
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
in greatest dimension. The age of the Tambien Group is
constrained by the fact that it overlies 800 Ma metavolcanics and syntectonic granitoids and is intruded by
610 Ma granites. This age constraint is consistent with
the inference that Tambien Group sediments were
deformed by the 630 Ma terminal collision to form the
EAO. Within the 800–610 Ma window, Miller et al.
(2003) found that Sr- and C-isotopic compositions of
Tambien Group carbonates are most consistent with an
age range of 720–750 Ma, broadly corresponding to the
Sturtian glaciation. A cap-carbonate has not been found
above the diamictite. Stratigraphic equivalents of the
Tambien Group can be expected to exist in southern
Arabia but these have not yet been identified.
Sedimentary units in the Eastern Desert of Egypt and
NW Saudi Arabia may record evidence of the Sturtian
Snowball Earth event in the northernmost ANS. The evidence from this region consists of diamictite and BIF.
The diamictite is known as the Atud conglomerate in Egypt
and as the Nuwaybah Formation (Zaam Group) in Arabia.
BIF is distributed throughout the Central Eastern Desert
of Egypt and is also found in the Silasia Formation in
NW Arabia (Fig. 10). Atud conglomerate and BIF are part
of a metasedimentary succession associated with ophiolites
(Stern, 1981; Stern et al., 2004). The ophiolite at Wadi
Ghadir has been dated by zircon evaporation techniques
at 746 ± 19 Ma (Kröner et al., 1992). Ophiolite, BIF, and
diamictite represent an oceanic assemblage that may
preserve evidence of deep marine conditions during the
Sturtian glaciation. Ophiolite and overlying metasediments
were similarly deformed and then intruded by syntectonic
granodiorites dated by Rb-Sr whole rock techniques at
15
674 ± 13 Ma (Stern and Hedge, 1985), and these ages constrain the Atud conglomerate and BIF to broadly belong to
the Sturtian episode.
The Atud conglomerate is only recognized in eastern
Egypt, where it can be found between 26°N and 22°N.
Its clasts are poorly sorted, polymict, and matrix supported. Clasts are generally subrounded and range in size
up to a meter across. It is a distinctive unit because its clasts
are quite different than the ensimatic assemblages that
characterize the Eastern Desert, and include grey quartzite,
arkose, felsic metavolcanics, granodiorite, and minor dark
grey marble. This is not a formal stratigraphic name, and
we propose that the unit is better referred to as the ‘Atud
diamictite’. Geochronologic data support the inference that
Atud diamictite clasts sample much older rocks than are
exposed in the Eastern Desert of Egypt and so must have
been transported some distance. Two granitic cobbles from
the NW of Marsa Alum (also referred to as the Wadi
Mobarak medisedimentary unit) yielded highly discordant
U–Pb zircon upper intercept ages of 1120 and 2060 Ma
(Dixon, 1981). Dixon (1979) obtained a discordant U–Pb
zircon upper intercept of 2.3 Ga for a granitic cobble from
Atud conglomerate outcrops west of Quesir. Pre-Neoproterozoic basement is unknown in Egypt east of the Nile, and
Dixon (1981) concluded that these clasts were derived from
older crust to the west or south, perhaps from the Saharan
Metacraton (Abdelsalam et al., 2002). Dixon (1979) suggested this material was transported such great distances
by ice rafting, a conclusion that is consistent with occasional dropstones (Fig. 8D).
Diamictite is also found in the Nuwaybah Formation
(Zaam Group) of NW Arabia (Davies, 1985). This unit
Fig. 10. (A) Location of Neoproterozoic BIF in the Central Eastern Desert of Egypt; X’s mark approximate locations of major deposits (Sims and James,
1984); (B) BIFs of the Sawawin District, N. Saudi Arabia. Location and extents of major deposits are shown as dark lines (modified after Goldring, 1990).
Location of maps shown as stars labelled ‘BIF’ in Fig. 2.
16
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
has not been studied in detail but appears similar in the
field to and is probably correlative with the Atud
diamictite.
BIF of broadly Sturtian age is found in the Central Eastern Desert of Egypt and in NW Saudi Arabia (Fig. 8E and
F). Arabian and Egyptian BIF formed in a single basin and
were separated by the Cenozoic opening of the Red Sea.
BIF in Egypt is found in the Central Eastern Desert,
between latitudes 25°15 0 and 26°30 0 N, where 15 occurrences are known (Fig. 10A). Egyptian BIF occurs as fairly
regular bands interbedded with metasediments and metavolcanics in a zone that originally had a stratigraphic thickness of 100–200 m, within which the aggregate thickness of
BIF is about 10–20 m (Sims and James, 1984). The BIFbearing sediments are associated with metavolcanics rocks
and are intruded by metadiabase sills. There is controversy
regarding how the Egyptian BIFs formed, although these
ideas were mostly developed prior to the Snowball Earth
hypothesis. Kamel et al. (1977) advocated an effusivemarine sedimentary mode of formation for Wadi Kareim
iron ores. Sims and James (1984) suggested that BIF
formed as chemical precipitates during lulls in dominantly
subaqueous, calc-alkaline volcanism, apparently within an
intraoceanic island-arc environment.
BIFS in the Midian region of NW Saudi Arabia occupy
a smaller region than do their Egyptian counterparts
(Fig. 10B). Arabian BIF occurs within the Silasia Formation, which, like the Egyptian section, is associated with
metavolcanic rocks. The exposed thickness of the Silasia
Formation is estimated to be about 1160 m in the reference
area of Wadi Sawawin. Also similar to the Egyptian section, the Silasia Formation is intruded by metadiabase sills.
It is also intruded by plutonic rocks of the Muwalylih suite,
dated by U–Pb zircon techniques at 710–725 Ma (Hedge,
1984). Johnson (2004) suggested on this basis that the Silasia Formation BIF could have been deposited in association with Sturtian glaciation.
Both Egyptian and Arabian BIFs are strongly deformed
and metamorphosed to the greenschist facies. These ores
are similar, mostly oxide facies, interbedded hematite and
jasper, and contain 40–46% Fe (Sims and James, 1984;
Goldring, 1990).
We infer that ANS BIFs formed about the same time as
the Sturtian glaciation, or shortly afterwards. Goldring
(1990) agreed with Sims and James (1984) that the Midian
BIFs were Algoma-type deposits, but also suggested that
the iron was precipitated as a result of oxidation of ferrous
iron in water by oxygen evolved during photosynthesis by
algae. The identification of algal fossils in the Egyptian
BIFs (El-Habaak and Mahmoud, 1995) supports the interpretation that biological activity may have been important
for multiple episodes where a marine environment that was
rich in Fe2+ was converted to an oxygenated environment
precipitating Fe3+.
BIF has only been reported from the northernmost part
of the ANS, as discussed above. Other components of the
ANS should have been deep basins below sea level during
the Sturtian glacial episode, so it is puzzling why BIF is not
more common in the ANS if formation was a synchronous,
deep sea expression of a ‘‘hard’’ Neoproterozoic Snowball
Earth.
6.3. Evidence for Marinoan (635–660 Ma) and Gaskiers
(582–585 Ma) glaciations
By about 630 Ma, collision had advanced sufficiently
that much of the ANS had probably risen above sea level.
Marinoan and Gaskiers Snowball Earth episodes, if present, are likely to have been manifested as continental glaciations, perhaps continental ice sheets. These would have
been powerful agents of erosion and could have rapidly
reduced relief as the EAO mountains grew.
Garfunkel (1999) identified the ‘Main Erosion Phase’ in
the northern ANS, which he suggested cut 8–14 km deep at
about 600 Ma. This is also consistent with evidence from
40
Ar/39Ar studies of micas for rapid cooling (and uplift)
at 600 Ma (Cosca et al., 1999). Similarly, a major phase
of erosion identified in the NE part of the Arabian Shield
during the interval 615–585 Ma was inferred to result from
epeirogenic uplift (Cole, 1988). Most explanations for
600 Ma exhumation focus on tectonic unroofing (AlHusseini, 2000; Blasband et al., 2000). We suggest that
Marinoan glaciation may have also been responsible for
much of this deep erosion. Unroofing farther south, in
Sudan and southern Egypt and Sudan, may have occurred
about 570 Ma, perhaps related to Gaskiers glaciation
(Bailo et al., 2003).
Certainly it is possible that some beveling occurred at
the base of thick continental ice sheets, but this possibility
has not been widely explored in the literature, largely
because unequivocal evidence for glaciation of the appropriate age has not been found. The greatest erosion is
expected to have occurred where relief was highest, in the
southern EAO, but we do not yet understand when and
how this region was beveled.
Evidence that pertains to the unroofing puzzle may also
be preserved in post-amalgamation basins of the northern
Arabian Shield (Johnson, 2003). Rocks of the Jurdhawiyah
Group and Hibshi Formation were deposited between 640
and 620 Ma in partly fault-controlled basins, which could
have existed at the time of Marinoan glaciation, but there
is no reported evidence for glacial deposits in these basins.
The Jurdhawiyah and Hibshi basins closed and inverted
during subsequent north–south shortening and northand south-vergent reverse faulting. The Jibalah Group
was deposited in isolated, pull-apart basins caused by
strike- and dip-slip movements of the Najd fault system.
A study of 3 km thick section of conglomerate, limestone,
sandstone, and shale in the Jifn Basin (NE Arabian Shield)
was reported by Kusky and Matsah (2003). They constrained its age (by U–Pb zircon techniques) to lie between
625 ± 4 Ma and 577 ± 5 Ma, so the Jibalah Group in the
Jin Basin could have been deposited during Gaskiers glaciation. Kusky and Matsah (2003) show a photograph of a
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
‘‘possible dropstone’’ (their Fig. 7D) but do not explore the
significance of this in detail. Further studies are needed to
establish whether or not evidence for Snowball Earth
events is preserved in Arabian post-accretionary basins,
and especially to constrain their ages as tightly as possible.
There are also sediments in the northern ANS that could
represent glacial deposits, or at least fluvial reworking of
sediments from Marinoan glaciers. The Saramuj conglomerate of Jordan and the Hammamat Group of NE Egypt
are both about 600 Ma old (Jarrar et al., 1993; Wilde and
Youssef, 2002) and perhaps were deposited as Marinoan
glaciation waned. Jarrar et al. (1991) interpreted a high
velocity, braided stream/alluvial fan system. Jibalah Group
sediments could have a similar origin. We speculate that all
of these coarse sediments could be periglacial tillites of
Marinoan age, reworked by meltwater streams as Marinoan glaciers receded but further investigation is needed
to confirm or refute this suggestion.
The Zenifim Formation, found only in boreholes from
the subsurface of Israel, Jordan, and Sinai, may be another
manifestation of Marinoan or Gaskiers glaciation. It is
>2500 m thick and consists of immature arkose-dominated
clastics and conglomerates associated with alkaline volcanics (Weissbrod and Sneh, 2002). Recanati (1986) reported
a K/Ar age of 606 ± 9 Ma for an igneous intrusion into
Zenifim sediments, implying that the Zenifim Formation is
older than this and supporting an interpretation that it
was mostly deposited during Marinoan time. It will be difficult to prove on the basis of drill core that these deposits are
or are not sedimentary deposits associated with a Marinoan
glacial episode, but the possibility should be considered.
There is less evidence to support an important role for
Gaskiers (600–570 Ma) glaciation in the ANS. The ‘Main
Erosion Phase’ happened before Gaskiers time, and continental sediments deposited 600 Ma, such as the Hammamat Formation of Egypt and Saramuj conglomerate of
Jordan, have not been removed. This is also consistent with
the record of glaciations preserved in the Huqf Supergroup
of Oman, discussed above. Garfunkel (1999) infers modest
(1–2 km) erosion of the northern ANS between 600 Ma
and the beginning of Cambrian time. The 560–540 Ma
Elat Conglomerate of southern Israel may have been
deposited during Gaskiers time, on a deeply dissected relief
that suggests sea level was quite low, perhaps as a result of
Gaskiers glaciation elsewhere. The Elat conglomerate contains clasts as large as 1.5 m (Weissbrod and Sneh, 2002),
and the possibility of a glacial or periglacial origin is worthy of further study.
One important observation is the vast peneplain that
North Africa and Arabia, which formed in multiple stages
over 100 million years following terminal collision
between E. and W. Gondwana and prior to deposition of
Cambrian marine sediments. This represents a continentscale erosional unconformity, which can be traced from
Morocco in the west to Oman in the east (Avigad et al.,
2003). There appears to be a sharply beveled surface below
the oldest Phanerozoic sediments all across North Africa
17
and Arabia, except for local monadnocks and where tectonism has occurred.
This extraordinary surface must have an extraordinary
origin. The wearing down of orogenic relief of ArabianNubian Shield had to be sufficient to yield a uniform, Nsloping surface that permitted Cambro-Ordovician streams
to flow north across it, as indicated by north-flowing paleocurrent directions from the Wajid sandstone in southern
Arabia (Dabbagh and Rogers, 1983). In order to explain
1.1–1.2 Ga detrital zircons in Cambrian quartz arenites in
Israel (Avigad et al., 2003), the headwaters of CambroOrdovician drainage may have reached as far south as
modern Tanzania, the northernmost limit of crust of this
age (Kröner et al., 2003). Alternatively, till could have been
glacially transported from southern Africa at least part
ways to the north and later reworked by streams.
Deep erosion and peneplanation involved at least two
episodes of erosion. The final cutting of the peneplain
occurred in early Cambrian time, because the peneplain
in southern Israel truncates dikes as young as 532 Ma
(Beyth and Heimann, 1999). The final cutting of the peneplain during early Cambrian time was not glacial, but associated with a warm and humid climate, as indicated by
thick laterite immediately below the peneplain.
In conclusion, the geological record may be taken to
support, if indirectly, an important role for Marinoan glacial erosion of the Arabian-Nubian Shield but much less
evidence in support of Gaskiers glaciation. This is consistent with the record preserved in Oman sediments, discussed above. How the basal Cambrian peneplain formed
has not been well studied, and the possible role of glaciation in its formation needs to be considered further. N.
African and Arabian geologists could contribute by initiating field research programs to characterize this unconformity in their regions.
7. Conclusions
The Snowball Earth hypothesis provides a valuable new
perspective on the evolution of the Arabian-Nubian Shield,
and new opportunities for African scientists to contribute
to our understanding of the Earth system. The effects of
Neoproterozoic glaciations can also provide age-diagnostic
stratigraphic markers, which are rare in ANS supracrustal
sequences. Tectonic evolution must be considered when
considering these effects, because—like most orogenic
belts—the ANS evolved as Neoproterozoic time progressed
from a mostly marine realm to a mountainous continental
environment. The transition occurred 630 Ma, about the
time of Marinoan glaciation. Sedimentary deposits of glacial episodes older than 630 Ma—Kaigas and Sturtian—are recognized in the ANS. There is evidence of
diamictites and BIF, and possible, but yet unsubstantiated,
cap carbonates. Kaigas-aged diamictites of apparent glacial origin are preserved on the south flank of the Bir
Umq–Nakasib suture, both in Arabia and in Sudan. Sedimentary evidence for Sturtian glaciation appears to be
18
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
widespread and is found in diamictites of the Tambien
Group in Ethiopia, Atud diamictites in Egypt, Nuwaybah
diamictites in Arabia, and banded iron formations in Egypt
and Arabia. This sedimentary evidence for a strong episode
of Sturtian glaciation is also consistent with the presence of
thick glaciogenic diamictites of the Ghubrah Formation in
Oman.
Evidence for Marinoan and Gaskiers glaciations is less
clear for the ANS, partly because the ANS was above sea
level during this time. The sedimentary record in Oman
suggests that Marinoan glaciation was important regionally but that the Gaskiers glaciation was not. Marinoan
glaciation may have caused the deep erosion of the ANS
that occurred about 600 Ma ago, providing a surface that
evolved over the next 60 million years or so into the continental-scale peneplain beneath basal Cambrian strata.
The ANS also provides opportunities for studying the
effects of Neoproterozoic climate change on the deep, open
ocean. Analogy with modern seafloor indicates that ANS
ophiolites formed at depths of 2–3 km below sea level
and the overlying pelagic sediments should record chemical
products of these deep waters. There is opportunity for
geochronologists and sedimentary geochemists to work
together to date the ophiolites and interpret the chemical
message preserved in overlying sediments.
Acknowledgment
We are grateful for assistance in the field from scientists
at Mekelle University and Ezana Minerals Corporation in
Mekelle, Ethiopia, especially Solomon Gebresilassie,
Kurkura Kabeto, Dirk Küster, and Kiros Mehari. We also
appreciate comments from Erwan Le Guerroué and Phillip
Allen (ETH, Zurich), Peter Johnson (SGS, Jeddah), and
Joe Meert (U Florida). The comments of two anonymous
referees and editor Eriksson are greatly appreciated as well.
This work is supported by USA–Israel Binational Science
Foundation (BSF) grant no. 2002337. This is UTD Geosciences contribution number 1068.
References
Abdelsalam, M.G., Liegeois, J.-P., Stern, R.J., 2002. The Saharan
metacraton. J. African Earth Sci. 34, 119–136.
Abdelsalam, M.G., Stern, R.J., 1993. Tectonic evolution of the Nakasib
suture, Red Sea Hills, Sudan: evidence for a late Precambrian Wilson
Cycle. J. Geol. Soc. London 150, 393–404.
Abdelsalam, M.G., Stern, R.J., 1996. Sutures and Shear Zones in the
Arabian-Nubian Shield. Journal of African Earth Sciences 23, 289–
310.
Al-Husseini, M.I., 2000. Origin of the Arabian Plate Structures: Amar
Collision and Najd Rift. GeoArabia 5, 527–542.
Alkmim, F.F., Marshak, S., Fonseca, M.A., 2001. Assembling West
Gondwana in the Neoproterozoic: clues from the Sao Francisco craton
region, Brazil. Geology 29 (4), 319–322.
Allen, P.A., Bowring, S., Leather, J., Brasier, M.D., Cozzi, A., Grotzinger,
J.P., McCarron, G., Amthor, J.E., 2002. Chronology of Neoproterozoic glaciations: new insights from Oman. In: The 16th International
Sedimenetological Congress, Abstract Volume, Johannesburg, South
Africa, pp. 7–8.
Allen, P.A., Leather, J., Brasier, M.D., 2004. Anatomy of a Neoproterozoic glacial epoch: the Fiq glaciation and its aftermath, Huqf
Supergroup of Oman. Basin Research 16, 507–534.
Avigad, D., Kolodner, K., McWilliams, M., Persing, H., Weissbrod, T.,
2003. Origin of northern Gondwana Cambrian sandstone revealed by
detrital zircon SHRIMP dating. Geology 31, 227–230.
Bailo, E., Schandelmeier, H., Franz, G., Sun, C.-H., Stern, R.J., 2003.
Plutonic and Metamorphic Rocks from the Keraf Suture (NE Sudan):
a Glimpse of the tectonic evolution of the NE margin of W. Gondwana
during Neoproterozoic time. Precambrian Res. 123, 67–80.
Beyth, M., Avigad, D., Wetzel, H.-U., Matthews, A., Berhe, S.M., 2003.
Crustal exhumation and indications for Snowball Earth in the East
African Orogen: north Ethiopia and east Eritrea. Precambrian Res.
123, 187–201.
Beyth, M., Heimann, A., 1999. The youngest igneous event in the
crystalline basement of the Arabian-Nubian Shield, Timna Igneous
Complex. Isr. J. Earth Sci. 48, 113–120.
Blasband, B., White, S., Brooijmans, P., de Brooder, H., Viser, W., 2000.
Late Proterozoic extensional collapse in Arabian-Nubian Shield. J.
Geol. Soc. London 157, 615–628.
Bonavia, F.F., Chorowicz, J., 1992. Northward expulsion of the PanAfrican of Northeast Africa guided by a reentrant zone of the
Tanzania Craton. Geology (Boulder) 20, 1023–1026.
Brasier, M., McCarron, G., Tucker, T., Leather, J., Allen, P., Shields, G.,
2000. New U–Pb zircon dates for the Neoproterozoic Ghubrah
glaciation and for the top of the Huqf Supergroup, Oman. Geology 28,
175–178.
Burns, S.J., Matter, A., 1993. Carbon isotopic record of the latest
Proterozoic from Oman. Eclogae Geologicae Helvetiae 86, 595–607.
Canfield, D.E., 1998. A new model for Proterozoic ocean chemistry.
Nature 396, 450–453.
Cawood, P.A., 2005. Terra Australis Orogen: Rodinia breakup and
development of the Pacific and Iapetus margins of Gondwana during
the Neoproterozoic and Paleozoic. Earth Sci. Rev. 69, 249–279.
Cole, J.C., 1988. Geology of the Aban Al Ahmar Quadrangle, Sheet 25F,
Kingdom of Saudi Arabia (explanatory notes). Deputy Ministry for
Mineral Resources Map GM-105A,C.
Collins, A.S., Pisaversky, S.A., 2005. Amalgamating eastern Gondwana:
the evolution of the Circum-Indian Orogens. Earth Sci. Rev. 71, 229–
270.
Condon, D., Maoyan, Z., Bowring, S., Wang, W., Yang, A., Jin, Y., 2005.
U–Pb ages from the Neoproterozoic Doushantuo Formation, China.
Science 308 (5718), 95–98.
Cosca, M.A., Shimron, A., Caby, R., 1999. Late Precambrian metamorphism and cooling in the Arabian-Nubian Shield: Petrology and
40
Ar/39Ar geochronology of metamorphic rocks of the Elat area
(southern Israel). Precambrian Res. 98, 107–127.
Dabbagh, M.E., Rogers, J.J.W., 1983. Depositional environments and
tectonic significance of the Wajid Sandstone of southern Saudi Arabia.
Journal of African Earth Sciences 1, 47–57.
Davies, F.B., 1985. Geologic map of the Al Wajh quadrangle, sheet 26B,
Kingdom of Saudi Arabia. Saudi Arabian Deputy Ministry for
Mineral Resources Geoscience Map GM-83, scale 1:250,000, 27 p.
Dixon, T.H., 1979. The evolution of continental crust in the Late
Precambrian Egyptian Shield. Ph.D. Thesis, UC San Diego, 231 p.
Dixon, T.H., 1981. Age and chemical characteristics of some pre-PanAfrican rocks in the Egyptian Shield. Precambrian Res. 14, 119–
133.
Donnadieu, Y., Goddéris, Y., Ramstein, G., Nédélec, A., Neert, J., 2004.
A ‘snowball Earth’ climate triggered by continental break-up through
changes in runoff. Nature 428, 303–306.
El-Habaak, G.H., Mahmoud, M.S., 1995. Carbonaceous bodies of
debatable organic provenance in the Banded Iron Formation of the
Wadi Kareim area, Eastern Desert, Egypt. J. African Earth Sciences
19, 125–133.
Evans, D., 2000. Stratigraphic, geochronological, and paleomagnetic
constraints upon the Neoproterozoic climatic paradox. American
Journal of Science 300, 347–433.
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
Evans, D., 2003. True polar wander and supercontinents. Tectonophysics
362, 303–320.
Eyles, N., Januszcak, N., 2004. Interpreting the Neoproterozoic glacial
record: the importance of Tectonics. In: Jenkins, G.S., McMenamin,
M.A.S., McKay, C.P. (Eds.). The extreme Proterozoic: geology,
geochemistry, and climate. AGU Geophysical Monograph 146, pp.
125–144.
Fanning, C.M., Link, P.K., 2004. U–Pb SHRIMP ages of Neoproterozoic
(Sturtian) glaciogenic Pocatello Formation, southeastern Idaho.
Geology 32 (10), 881–884.
Flint, R.F., Sanders, J.E., Rodgers, J., 1960. Diamictite, a substitute term
for symmictite. Geol. Soc. Am. Bull. 71, 1809–1810.
Garfunkel, Z., 1999. History and paleogeography during the Pan-African
orogen to stable platform transition: reappraisal of the evidence from
the Elat area and the northern Arabian-Nubian Shield. Israel J. Earth
Sci. 48, 135–157.
Goddéris, T., Donnadieu, Y., Nédélec, A., Dupré, B., Dessert, C., Grard,
A., Ramstein, G., François, L.M., 2003. The Sturtian ‘snowball’
glaciation: fire and ice. Earth Planet. Sci. Lett. 211, 1–12.
Goldring, D.C., 1990. Banded iron formation of Wadi Sawawin district,
Kingdom of Saudi Arabia. Trans. Instn. Min. Metall. (Sect B: Appl.
Earth Sci.) 99, B1–B14.
Grotzinger, J.P., Knoll, A.H., 1995. Anomalous carbonate precipitates: Is
the Precambrian the key to the Permian? Palaios 10, 578–596.
Halverson, G.P., Hoffman, P.F., Schrag, D.P., Maloof, A.C., Rice,
A.H.N., 2005. Towards a Neoproterozoic composite carbon isotope
record. Geol. Soc. Amer. Bull. 117, 1181–1207.
Harland, W.B., 1964. Critical evidence for a great Infra-Cambrian
glaciation. Geol. Rundschau 54, 45–61.
Hayes, J.M., Strauss, H., Kaufman, A.J., 1999. The abundance of 13C in
marine organic matter and isotopic fractionation in the global
biogeochemical cycle of carbon during the past 800 Ma. Chem. Geol.
161, 103–125.
Hedge, C.E., 1984. Precambrian geochronology of part of northwestern
Saudi Arabia, Kingdom of Saudi Arabia. US Geological Survey Open
File Report 83-381, 12 pp.
Higgins, J.A., Schrag, D.P., 2003. Aftermath of a snowball Earth.
Geochem. Geophs. Geosyst. 4, 1028. doi:10.1029/2002GC000403.
Hladil, J., 1991. The Upper Ordovician dropstones of Central Bohemia
and their paleogravity significance. Vest. Ustr. Ust. Geol. 66, 65–74.
Hoffmann, K.H., Condon, D.J., Bowring, S.A., Crowley, J.L., 2004. U–Pb
zircon date from the Neoproterozoic Ghaub Formation, Namibia;
constraints on Marinoan glaciation. Geology 32, 817–820.
Hoffman, P.F., 1999. The break-up of Rodinia, birth of Gondwana, true
polar wander and the snowball Earth. J. African Earth Sci. 29, 17–33.
Hoffman, P.F., Kaufman, A.J., Halverson, G.P., Schrag, D.P., 1998. A
Neoproterozoic Snowball Earth. Science 281, 1342–1346.
Hoffman, P.F., Schrag, D.P., 2000. Snowball Earth. Sci. Amer. 282 (1),
50–57.
Hoffman, P.F., Schrag, D.P., 2002. The Snowball Earth hypothesis:
testing the limits of global change. Terra Nova 14, 129–155.
Jacobs, J., Bauer, W., Fanning, C.M., 2003. Late Neoproterozoic/Early
Palaeozoic events in central Dronning Maud Land and significance for
the southern extension of the East African Orogen into East
Antarctica. Precambrian Res. 126, 27–53.
James, N.P., Narbonne, G.M., Kyser, T.K., 2001. Late Neoproterozoic
cap carbonates: Mackenzie Mountains, northwestern Canada: precipitation and global glacial meltdown. Can. J. Earth Sci. 38, 1229–
1262.
Jarrar, G.H., Wachendorf, H., Zellmer, D., 1991. The Saramuj Conglomerate: evolution of a Pan-African molasse sequence from southwest
Jordan. N. Jb. Geol. Palaontol. Mh. 6, 335–356.
Jarrar, G.H., Wachendorf, H., Zachmann, D., 1993. A Pan-African
alkaline pluton intruding the Saramuj Conglomerate, South-west
Jordan. Geol. Rundschau 82, 121–135.
Johnson, P.R., 2003. Post-amalgamation basins of the NE Arabian shield
and implications for Neoproterozoic III tectonism in the northern East
African Orogen. Precambrian Res. 123, 321–337.
19
Johnson, P.R., 2004. Proterozoic geology of western Saudi Arabia:
Northwestern sheet: Saudi Geological Survey Open-File Report SGSOF-2004-4, 29 p.
Johnson, P.R., Abdelsalam, M.G., Stern, R.J., 2002. The Bi’r Umq–
Nakasib Shear zone: geology and structure of a Neoproterozoic suture
in the northern East African Orogen, Saudi Arabia and Sudan. Saudi
Geological Survey Technical Report SGS-TR-2002-1, 33 p.
Johnson, P.R., Abdelsalam, M.G., Stern, R.J., 2003. The Bi’r Umq–
Nakasib Suture Zone in the Arabian-Nubian Shield: a key to
understanding crustal growth in the East African Orogen. Gondwana
Res. 6, 523–530.
Johnson, P.R., Woldehaimanot, B., 2003. Development of the ArabianNubian Shield: perspectives on accretion and deformation in the
northern East African Orogen and the assembly of Gondwana. In:
Yoshida, M., Windley, B.F., Dasgupta, S. (Eds.), Proterozoic East
Gondwana: Supercontinent Assembly and Breakup, Geological Society, London, Special Publication 206, 289–325.
Kamel, O.A., Hilmy, E.M., Niazy, E.A., 1977. Origin of Precambrian iron
ore deposits from Wadi Kareim, Eastern Desert, Egypt. Bull. NRC
Egypt, 401–413.
Kennedy, M.J., Christie-Blick, N., Sohl, L.E., 2001. Are Proterozoic cap
carbonates and isotopic excursions a record of gas hydrate destabilization following Earth’s coldest intervals? Geology 29, 443–446.
Kilner, B., Mac Niocaill, C., Brasier, M., 2005. Low-latitude glaciation in
the Neoproterozoic of Oman. Geology 33, 413–416.
Kirschvink, J.L., 1992. Late Proterozoic low-latitude global glaciation: the
snowball earth. In: Schopf, J.W., Klein, C. (Eds.), The Proterozoic
Biosphere. Cambridge University Press, New York, pp. 51–52.
Knoll, A.H., Hayes, J.M., Kaufman, A.J., Swett, K., Lambert, I.B., 1986.
Secular variation in carbon isotope ratios from Upper Proterozoic
successions of Svalbard and East Greenland. Nature 321, 832–838.
Knoll, A., 2003. Life on a Young Planet. Princeton University Press,
276 p.
Knoll, A., Walter, M.R., Narbonne, G.M., Christie-Blick, N., 2004. A
new period for the geologic time scale. Science 305, 621–622.
Kröner, A., Muhongo, S., Hegner, E., Wingate, M.T.D., 2003. Singlezircon geochronology and Nd isotopic systematics of Proterozoic highgrade rocks from the Mozambique belt of southern Tanzania (Masasi
area): implications for Gondwana assembly. J. Geol. Soc. London 160,
645–757.
Kröner, A., Todt, W., Hussein, I.M., Mansour, M., Rashwan, A.A., 1992.
Dating of late Proterozoic ophiolites in Egypt and the Sudan using the
single grain zircon evaporation technique. Precambrian Res. 59, 15–32.
Kusky, T.M., Matsah, M.I., 2003. Neoproterozoic dextral faulting on the
Najd Fault System, Saudi Arabia, preceeded sinistral faulting and
escape tectonics related to closure of the Mozambique Ocean. In:
Yoshida, M., Windley, B.R.F., Dasgupta, S. (Eds.), Proterozoic East
Gondwana: supercontinent assembly and breakup. Geological Society
of London, Special Publication 206, 327–361.
Kuypers, M.M.M., Pancost, R.D., Nijenhuis, I.A., Sinninghe Damsté,
J.S., 2002. Enhanced productivity led to increased organic carbon
burial in the euxinic North Atlantic basin during the late Cenomanian
oceanic anoxic event. Paleoceanography 17 (4), 1051. doi:10.1029/
2000PA000569.
Le Guerroué, E., Allen, P.A., Cozzi, A., 2005a. Two distinct glacial
successions in the Neoproterozoic of Oman. GeoArabia 10, 17–34.
Le Guerroué, E., Allen, P.A., Cozzi, A., 2005b. The largest d13C excursion
of Earth History: the late Neoproterozoic Khufai-Shuram boundary of
Oman. Abstract, European Union of Geosciences, Vienna.
Leather, J., Allen, P.A., Brazier, M.D., Cozzi, A., 2002. Neoproterozoic
snowball Earth under scrutiny: evidence from the Fiq glaciation of
Oman. Geology 30, 891–894.
Martin, M.J., Christie-Blick, N., Sohl, L.E., 2001. Are Proterozoic cap
carbonates and isotopic excursions a record of gas hydrate destabilization following Earth’s coldest intervals? Geology 29, 443–446.
MacGabhann, B.A., 2005. Age constraints on Precambrian glaciations
and the subdivision of Neoproterozoic time. IUGS Ediacaran
Subcommission Circular, August 21, 2005. List archive: <https://
20
R.J. Stern et al. / Journal of African Earth Sciences 44 (2006) 1–20
www.jcu.edu.au/pipermail/iugsediacaransubcommision/2005-August/
>.
Meert, J.G., 2003. A synopsis of events related to the assembly of eastern
Gondwana. Tectonophysics 362, 1–40.
Meert, J.G., Torsvik, T.H., 2004. Paleomagnetic constraints on Neoproterozoic ‘Snowball Earth’ continental reconstructions. In: Jenkins,
G.S., McMenamin, M.A.S., McKay, C., Sohl, L. (Eds.), The extreme
Proterozoic: geology, geochemistry, and climate. AGU Geophysical
Monograph Series, vol. 146, pp. 5–11.
Miller, N.R., Alene, M., Sacchi, R., Stern, R.J., Conti, A., Kröner, A.,
Zuppi, G., 2003. Significance of the Tambien Group (Tigrai, N.
Ethiopia) for Snowball Earth events in the Arabian-Nubian Shield.
Precambrian Res. 121, 263–283.
Ohmoto, H., 2004. 5.2 The Archean Atmosphere, Hydrosphere and
Biosphere. In: Eriksson, P.G., Altermann, W., Nelson, D.R., Mueller,
W.U., Catuneanu, O. (Eds.), The Precambrian Earth: Tempos and
Events, Developments in Precambrian Geology, vol. 12. Elsevier, pp.
361–388.
Patchett, P.J., Chase, C.G., 2002. Role of transform continental margins
in major crustal growth episodes. Geology 30, 39–42.
Prave, A.R., 1999. Two diamictites, two cap carbonates, two d13C
excursions, two rifts: the Neoproterozoic Kingston Peak Formation,
Death Valley, California. Geology 27, 339–342.
Proust, J.-N., Deynoux, M., 1994. Marine to non-marine sequence
architecture of an intracratonic glacially related basin. Late Proterozoic of the west Africa platform in western Mali. In: Deynoux, M.,
Miller, J.M.G., Domack, E.W., Eyles, N., Farichild, I.J., Young, G.M.
(Eds.), Earth’s Glacial Record. Cambridge University Press, Cambridge, pp. 121–145.
Prueher, L.M., Rea, D.K., 2001. Volcanic triggering of late Pliocene
glaciation; evidence from the flux of volcanic glass and ice-rafted
debris to the North Pacific Ocean. Palaeogeography, Palaeoclimatology, Palaeoecology 173, 215–230.
Recanati, P., 1986. The K/Ar and Rb/Sr systems in magmatic rocks from
the subsurface of the NE Negev. M.Sc. Thesis, Hebrew University, pp.
79.
Ridgewell, A.J., Kennedy, M.J., Caldeira, K., 2003. Carbonate deposition,
climate stability, and Neoproterozoic Ice ages. Science 302, 859–862.
Rouxel, O.J., Bekker, A., Edwards, K.J., 2005. Iron isotope constraints on
the Archean and Paleoproterozoic Ocean Redox state. Science 307,
1088–1091.
Royer, D.L., Berner, R.A., Montanez, I.P., Tabor, N.J., Beerling, D.J.,
2004. CO2 as a primary driver of Phanerozoic climate. GSA Today,
vol. 14; no. 3, doi: 10.1130/1052-5173(2004)014h4:CAAPDOi2.0.CO;2.
Shields, G.A., 2005. Neoproterozoic cap carbonates: a critical appraisal of
existing models and the plumeworld hypothesis. Tera Nova 17, 299–310.
Sims, P.K., James, H.L., 1984. Banded Iron-formations of late Proterozoic Age in the Central Eastern desert of Egypt: geology and tectonic
setting. Econom. Geol. 79, 1777–1784.
Stein, M., 2003. Tracing the plume material in the Arabian-Nubian Shield.
Precambrian Res. 123, 223–234.
Stern, R.J., 1981. Petrogenesis and tectonic setting of late Precambrian
ensimatic volcanic rocks, Central Eastern Desert of Egypt. Precambrian Res. 16, 195–230.
Stern, R.J., 1994. Arc assembly and continental collision in the Neoproterozoic East African Orogen: implications for the consolidation of
Gondwanaland. Ann. Rev. Earth Planet. Sci. 22, 319–351.
Stern, R.J., 2002. Crustal evolution in the East African Orogen: a
Neodymium isotopic perspective. J. African Earth Sci. 34, 109–117.
Stern, R.J., 2005. Evidence from ophiolites, blueschists, and ultrahighpressure metamorphic terranes that the modern episode of subduction
tectonics began in Neoproterozoic time. Geology 33, 557–560.
Stern, R.J., Abdelsalam, M.G., 1998. Formation of juvenile continental
crust in the Arabian-Nubian Shield: evidence from granitic rocks of
the Nakasib suture, NE Sudan. Geol. Rundschau 87, 150–160.
Stern, R.J., Hedge, C.E., 1985. Geochronologic constraints on late
Precambrian crustal evolution in the Eastern Desert of Egypt. Amer. J.
Sci. 285, 97–127.
Stern, R.J., Johnson, P.J., Kröner, A., Yibas, B., 2004. ‘Neoproterozoic
Ophiolites of the Arabian-Nubian Shield. In: Kusky, T. (Ed.),
Precambrian Ophiolites. Elsevier, pp. 95–128.
Tadesse, T., Hoshino, M., Sawada, Y., 1999. Geochemistry of low-grade
metavolcanic rocks from the Pan-African of the Axum area, northern
Ethiopia. Precambrian Res. 96, 101–124.
Trendall, A.F., 2002. The significance of iron-formation in the Precambrian stratigraphic record. Spec. Publs. Int. Ass. Sediment. 33, 33–
66.
Trendall, A.F., Blockley, J.G., 2004. 5.4 Precambrian Iron-Formations.
In: Eriksson, P.G., Altermann, W., Nelson, D.R., Mueller, W.U.,
Catuneanu, O. (Eds.), The Precambrian Earth: Tempos and Events,
Developments in Precambrian Geology, vol. 12. Elsevier, pp. 403–
421.
Trindade, R.I.F., Font, E., D’Agrella-Filho, M.S., Nogueira, A.C.R.,
Riccomini, C., 2003. Low-latitude and multiple geomagnetic reversals
in the Neoproterozoic Puga cap carbonate, Amazon craton. Terra
Nova 15, 441–446.
Valentine, J.W., Moores, E.M., 1974. Plate Tectonics and the History of
Life in the Oceans. Scientific American 230 (4), 80–89.
Veevers, J.J., 1990. Tectonic–climatic supercycle in the billion-year platetectonic eon; Permian Pangean icehouse alternates with Cretaceous
dispersed-continents greenhouse. Sediment. Geol. 68, 1–16.
Veevers, J.J., 2003. Pan-African is Pan-Gondwanaland: oblique convergence drives rotation during 650–500 Ma assembly. Geology 31, 501–
504.
Weissbrod, T., Sneh, A., 2002. Sedimentology and paleogeography of the
Late Precambrian–Early Cambrian arkosic and conglomeratic facies in
the northern margins of the Arabo-Nubian Shield Bulletin. Geol.
Survey Israel 87, 44.
Wilde, S.A., Youssef, K., 2002. A re-evaluation of the origin and setting of
the late Precambrian Hammamat Group based on SHRIMP U–Pb
dating of detrital zircons from Gebel Umm Tawat, North Eastern
Desert, Egypt. J. Geol. Soc. London 159, 595–604.
Williams, G.E., 1975. Late Precambrian glacial climate and the Earth’s
obliquity. Geol. Mag. 112, 441–544.
Williams, G.E., 2004. 5.7. The Paradox of Proterozoic Glaciomarine
deposition, Open Seas and strong seasonality near the Palaeo-Equator:
global Implications. In: Eriksson, P.G., Altermann, W., Nelson, D.R.,
Mueller, W.U., Catuneanu, O. (Eds.), The Precambrian Earth:
Tempos and Events, Developments in Precambrian Geology, vol. 12.
Elsevier, pp. 448–459.
Woldehaimanot, B., 2000. Tectonic setting and geochemical characterisation of Neoproterozoic volcanics and granitoids from the Adobha
Belt, northern Eritrea. J. African Earth Sci. 30, 817–831.
Worsley, T.R., Nance, R.D., Moody, J.B., 1986. Tectonic cycles and the
history of the Earth’s biogeochemical and paleoceanographic record.
Paleoceanography 1, 233–263.
Young, G.M., 2002. Stratigraphic and tectonic settings of Proterozoic
glaciogenic rocks and banded iron-formations: relevance to the
snowball Earth debate. J. African Earth Sci. 35, 451–466.
Young, G.M., 2004. 5.6 Earth’s two great Precambrian Glaciations:
aftermath of the ‘‘Snowball Earth’’ hypothesis. In: Eriksson, P.G.,
Altermann, W., Nelson, D.R., Mueller, W.U., Catuneanu, O. (Eds.),
The Precambrian Earth: Tempos and Events, Developments in
Precambrian Geology, vol. 12. Elsevier, pp. 440–448.