Continental crust

Continental crust is a unique feature of the Earth and together with oceanic crust makes up the outermost layer of the planet. Its density ranges from 2.7-2.9g/cm³ and extends from the surface to the Mohorovicic discontinuity which marks the crust–mantle boundary (Rudnick and Gao, 2003). It covers ~41% of the Earth surface area and constitutes ~70% of Earth’s crustal volume (Cogley, 1984). The continental crust thickness varies between 20 and 70 km with averaging of 35-40 km. The lateral extent of the continent crust is marked by continental shelf, of which 31% is submerged beneath the oceans (Cogley, 1984; Mooney et al., 1998). The continental crust is ancient as it preserves a record of the Earth's physical and chemical evolution. Present-day continental crust mostly formed at the end of the Archean and the beginning of the Proterozoic, between 3.2 to 2.0 Ga (Taylor and McLennan, 1995). The oldest rock within the continental crust is the 4.0 Ga Acasta Gneisses, Slave Province of NW Canada (Taylor and McLennan, 1995). The oldest detrital zircons are found in the Jack Hills area in western Australia and reported to be 4.4 billion years old (Wilde et al., 2001). In contrast, the oldest oceanic crust is ~200 Ma (Cloos, 1993). The structure of the continental crust consists of an upper layer composed mainly of sedimentary rock and granite to granodiorite, a middle layer of amphibolite facies and granulite facies, and a lower layered consisting of granulite-facies country (Rudnick and Fountain, 1995; Wedepohl, 1995).

Continental crust formation is a continuous process with at least 80% produced at a convergent margin setting (Barth et al., 2000; Cawood, 2013). The formation of continental crust is more or less related to convection within the mantle that is caused by hot mantle material rising upwards, cooling, and followed by the subduction of cooler oceanic crust back into the mantle (Plank and Langmuir, 1998).

4 1.2. Super continental cycles

Generally, the crust of the Earth is comparable to an egg shell that has been repeatedly cracked where the cracks represent plate boundaries. The majority of crust recycling occurs at plate boundaries as well as the formation of new crust. Plate interactions occur in three principal ways: 1) collision (convergent), 2) transform, and 3) spreading (divergent). The current paradigm of plate tectonics is the modern version of continental drift theory (Wegener, 1912) that suggests the continents have moved to their current positions. The significant concept derived from plate tectonics theory is that continents have grown through time and have experienced a cycle or multiple cycles of assembly and break-up (Umbgrove, 1940). A supercontinent is regarded as an inevitable consequence of plate tectonics, and is defined as the assembly of most or all of Earth's continental blocks or cratons to form a single large landmass (Rogers and Santosh, 2004). Sutton (1963) was amongst the first to suggest there were multiple (four) supercontinent cycles during Earth’s history. One supercontinent cycle is said to take 300 to 500 million years (Worsley et al., 1985). Currently, it is thought there were five pre-Pangean supercontinents at supercontinents at ca. 2.6 Ga, 2.0 Ga, 1.8 Ga to 1.6 Ga and 1.1 Ga that correspond to the amalgamation of Sclavia and Superia (Kenorland), Columbia (or Nuna), Rodinia and the later fifth, Pangaea (~600 Ma) which broke up into Laurasia to become North America and Eurasia and Gondwana (Suess, 1885). The geological evidence is more compelling for the younger supercontinents (Pangaea and Gondwana) as the rock record is better preserved compared to older lithologies. Consequently, older supercontinents are less constrained than the younger supercontinents.

5 1.3. Relationship between Rodinia and Gondwana Rodinia supercontinent

The name “Rodinia” (Fig. 1.1) was first used in McMenamin and McMenamin (1990).

Rodinia is a Mesoproterozoic supercontinent formed after the breakup of an older supercontinent, i.e. Columbia (Zhao et al., 2002) and was surrounded by the superocean called Mirovic Ocean. The evidence for the existence of Rodinia was found during the 1970s by the discovery of orogenic events from 1.23 Ga to 1.0 Ga that exist on virtually all cratons (i.e. the Grenville orogeny in America, the Dalslandian orogeny in Europe; Dewey and Burke, 1973).

Since then, based on the orogens correlation on different cratons, a number of configurations have been proposed.

Figure 1.1. Reconstruction of Rodinia supercontinent by ca. 1000 Ma [modified after Meert and Torsvik, 2003)]

In 2009, International Geological Correlation Project 440 named “Rodinia Assembly and Breakup” recognized that from 825 to 550 Ma, Rodinia broke up in four different stages to form Gondwana. They are:

6

 825 to 800 Ma: correlative evidence (crustal arching, coeval magmatism, and accumulation of thick rift-type sedimentary) of a super plume event recorded in the South Australia, South China, Tarim, Kalahari, India, and the Arabian - Nubian cratons.

 800 to 750 Ma: rifting proceeded within the same cratons and spread to Laurentia and Siberia;

 750 to 700 Ma: a new pulse of magmatism and rifting continued along most margins of Laurentia;

 650 to 550 Ma: the Pan-African orogenies, opening of the Iapetus Ocean, and closure of the Mozambique ocean all resulted in the formation of Gondwana.

The break-up of Rodinia led to the formation of smaller continental blocks that eventually amalgamated into Gondwana (Meredith et al., 2016). Most of evidence for this plate reconfiguration is found within the African and neighbouring (South America, Antarctica, India and Australia) cratonic domains as a number of orogenic belts developed from ~750 Ma until ~500 Ma. Of particular note is the north-south trending East African Orogen (EAO) that formed as a consequence of closure of the Mozambique Ocean collision with India (Collins and Pisarevsky, 2005; Fritz et al., 2013).

Gondwana supercontinent

Gondwana (Fig. 1.2), named by Suess (1885), was a supercontinent that fully assembled during the Neoproterozoic (~550 Ma) and the first stage of its breakup began during Carboniferous (~320 Ma). The remnants of Gondwana make up about two-thirds of today's continental area, including Africa, Antarctica, South America, Australia, and India. The formation of Gondwana was initiated from 800 to 650 Ma and completed by 600 to 530 Ma with the collision of India with eastern Africa and South America with western Africa (Meert and Van Der Voo, 1997). The break-up of Gondwana occurred in a number of stages. During

7

Mesozoic, supercontinent Pangaea started to break-up and separation of Gondwana into different individual continents (Australia, Antarctica, Africa, Madagascar, South America and India). Beginning in the middle Jurassic, around 167 Ma, recorded the separation of Gondwana into East Gondwana (comprising Antarctica, Madagascar, India and Australia) and West Gondwana (comprising Africa and South America). The next stage, Africa separated from South America to form the South Atlantic and India move northward from Antarctica. Between 95 and 84 Ma, rifting separated India from Madagascar. Subsequently, oceanic crust began to form within the Mascarene basin causing a rotation of the Seychelles/India land mass. This continued until ca. 66 Ma when formation of the currently active Carlsberg Ridge severed the Seychelles from India (Plummer and Belle, 1995; Torsvik et al., 1998). At that time, India collided with Asia, resulting in the formation of the Himalayas and Antarctica separated from Australia.

Figure 1.2. Peri-Gondwanan terranes at ∼570 Ma [modified from Linnemann et al. (2007)]

(Mad: Madagascar, Sey: Seychelles)

8 1.4. East African Orogen

The Pan-African Orogeny (sensu lato) formed by collisional events resulting from the convergence of lithospheric microplates created by the break-up of Rodinia supercontinent.

The term ‘Pan-African’ is widely used to denote to the principal part of the global orogeny that led to the formation of the supercontinent Gondwana (Kennedy, 1964). There are two types of orogens within the associated with the Pan-African belts. The first type consists of Neoproterozoic supracrustal and magmatic assemblages, which have mantle-derived origin, and contain ophiolites as a diagnostic feature. Such belts include the Arabian-Nubian Shield (ANS) of Arabia and NE Africa. This type exposes rocks from the upper to middle crust. The second type of mobile belt generally contains polydeformed high-grade metamorphic assemblages, which expose rocks from the middle to lower crust. The assemblages consist of Mesoproterozoic to Archaean continental crust that was strongly reworked during Neoproterozoic. One of the best examples of the high-grade belts is the Mozambique Belt of East Africa. Together, the thin-skinned ANS and the thick-skinned high-grade Mozambique Belt comprise the EAO (Stern, 1994; Stern, 2002). This orogen is about 6,000 km long, stretches from ANS through Tanzania and northern-central Madagascar into the southern India (Dharwar craton), Sri Lanka and eastern Antarctica (Fig. 1.3). The EAO is made up of deformed and metamorphosed crust from Neoproterozoic to Cambrian. The tectonic evolution of the EAO started with rifting and break-up at ~900-850 Ma. Rifting is followed by seafloor spreading (870 to 690 Ma) that led to arc and back-arc basin formation. It is followed by continent-continent collision from 630 to 600 Ma and ended by further crustal shortening (orogenic collapse, extension) that will eventually be exploited during the Mesozoic break-up of Gondwana (Stern, 1994).

9 Arabian - Nubian Shield

The ANS is an exposure of Precambrian rocks and the largest section of predominantly juvenile Neoproterozoic crust in Africa (Kröner and Stern, 2004). The evolution history of the shield is intimately linked with two supercontinent cycles, the amalgamation of Gondwana, results from the fragmentation of Rodinia and the closure of the Mozambique Ocean. The ANS was fully assembled by 550 Ma after undergoing ~300 Ma growth that included terrane suturing and island arc convergence starting at 780 Ma.

Figure 1.3. The reconstructed Gondwana supercontinent and East African Orogen [modified from Zhou et al. (2018)]

Madagascar

Madagascar (Fig. 1.4) is an ideal area to study EAO, especially the central and northern parts. There are five principal tectonic units identified in that area including the Archean Antongil-Masora block, the Antananarivo block, Tsaratanana complex, the Neoproterozoic Bemarivo Belt and the Proterozoic metasedimentary rock of Itremo group (Collins et al., 2001).

The Antongil-Masora block of eastern Madagascar is comprised of 3.2 Ga gneiss intruded by

10

~2.5 Ga granites that have only experienced low-grade metamorphism since ~2.5 Ga. The Antananarivo block, which underlies the central highland area, is comprised of 2.5 Ga gneiss interlayered with 820 to 740 Ma granitoids and gabbros. It is overlain by Itremo group on the central-west that contains metasedimentary rocks of the Mesoproterozoic to early Neoproterozoic Itremo Group thrust over, and imbricated with, rocks of the Antananarivo block (Cox et al., 1998). The Tsaratanana complex overlies the northern Antananarivo, dominated by Neoarchean orthogneiss (2.7 to 2.5 Ga) and cut by 800 to 760 Ma gabbro intrusions. The northern Madagascar, Bemarivo Belt, juxtaposed the Antongil and Antananarivo, consists of amphibolite-granulite gneiss on the southern and granitic massif on the northern (Collins, 2000; Kröner, 2000; Collins and Windley, 2002).

Figure 1.4. Simplified geological map of Madagascar [modified from Collins and Windley (2002)]

11 India - Dharwar craton

The Indian Shield as shown in Fig. 1.5., is comprised of four main cratons that formed and stabilized during Archean to Paleoproterozoic and Meso- to Neoproterozoic mobile belts (Meert and Pandit, 2015; Saha and Mazumder, 2012). The four cratons of the Indian Shield include: the Dhawar craton, the Bastar craton, the Singhbhum craton, and the Aravalli-Bundelkhand craton (Srivastava, 2008).

Figure 1.5. Simplified geological map of India [modified from Srivastava (2008)]

(MIS: Malani Igneous Suite)

The Dharwar craton of southern India (Fig. 1.5) is comprised of four major units: the Archean tonalite-trondhjemite-granodiorite gneiss, the supracrustal belts, the Proterozoic metasedimentary basin and the late potassic granite, such as Closepet granite. The Chitradurga schist belts divided the craton into two cratonic domains named the Western Dharwar (WDC, ca. 3.3 to 2.7 Ga) and the Eastern Dharwar (EDC, ca. 3.0 to 2.5 Ga) (Meert and Pandit, 2015).

The WDC contains most of tonalite-trondhjemite-granodiorite gneiss while the EDC is a suite of granitic rocks, intrusive volcanics, and middle Proterozoic to recently sedimentary basins.

12

The emplacement of Closepet granite during ~2.5 Ga indicated the amalgamation of WDC and EDC as well (Saha and Mazumder, 2012; Chadwick et al., 2000).

India - MIS

The term MIS was first used by Murthy at al. (1961). MIS is located in northwestern Indian Shield and covers an area of ~50,000 km². The magmatic activities in MIS began with mafic and felsic flows and the accumulation of conglomerate. Next followed stage is dominant by the felsic flows. At this stages, granite plutonism occurred mostly at central MIS, especially Siwana and Jalore. The end of Malani magmatism was marked by felsic and basic dyke intrusions (Sharma, 2005). The Siwana is a large ring structure, around 30 km diameter, and consists of mostly intruded peralkaline granite. The Jalore shows a variety range from quartz syenite to granite (Eby and Kochhar, 1990).

1.5.The Seychelles microcontinent

The elliptical-shaped block located at the northern end of the Mascarene plateau in the Indian Ocean is known as the Seychelles microcontinent. The surface exposure is a series of granitic islands that are ~1,000 km from Madagascar and East Africa (Fig. 1.6). The Mascarene plateau is the second largest submarine plateau in the Indian Ocean. It is a shallow (8 m to 150 m deep), extensive rise that forms a crescent shape from Mauritius in the south to the Seychelles in the north. The Mascarene plateau extends nearly 2,000 km and covers an area of ~115,000 km².

13

Figure 1.6. Regional topography of Mascarene plateau and the western Indian Ocean, eastern Africa, Madagascar

Figure 1.7. Simplified map of the Seychelles microcontinent [modified after Baker (1963) and Suwa et al. (1983)]

There are two distinctive regions in the Seychelles, the granitic islands and the coralline outer islands. The granitic group is concentrated within a 56 km radius and consists of 45 covering a total area of 247.2 km². The main islands of this group are Mahé, Praslin, Silhouette,

14

La Digue, and North Island. The outer islands contain five groups of coralline islands, comprise 211.3 km² (46% of the Seychelles). The coral islands show elevated coral reefs at different stages of formation in flat terranes. The Main Islands of the Seychelles are only islands in the world that are entirely composed of granite and they are also the oldest islands in the world as they have a cluster of zircon concordant ages in the range of 750-760 Ma although Tucker et al. (2001) reported one (Île aux Récifs) high-precision U-Pb zircon age of 808.8±1.9 Ma, interpreted as magmatic crystallization ages.

The Seychelles is a key region as it was located at the centre of the EAO and preserved rocks that recorded the Rodinia and Gondwana supercontinent cycle. There are three main types of granite that comprise the Main Islands of the Seychelles and include: the dominant grey granite on Mahé, the dominant pink granite on Praslin-La Digue and the younger syenite on Silhouette and North Island (Fig. 1.7). Both Wegener (1924) and Du Toit (1937) interpreted the Seychelles microcontinent as a continental fragment left behind during Gondwana break-up. Once sandwiched between Africa and India as part of Gondwanaland, the Seychelles split away from India during the Early Paleogene (Plummer, 1995; Shellnutt et al., 2015).

1.6. Purpose of the study

The granitic islands of the Seychelles have been considered as a geological curiosity for a long time. The interest in the Seychelles is inspired by their unique geologic history with a great deal of research on its geology, especially the debate regarding their tectonomagmatic evolution as well as their cratonic affinity. Currently, there are a number of first order issues regarding the Seychelles that have yet to be resolved. Specifically, whether the Seychelles represents a fragment of ancient crust derived from India or Africa. Moreover, there is debate concerning the tectonomagmatic evolution of the Neoproterozoic granites vis-à-vis subduction-related vs. extension-related.

15

Weis and Deutsch (1984) measured the Nd and Pb isotopes and suggested Precambrian Seychelles rocks are alkaline anorogenic and mantle-derived with slight upper-crust interaction.

Taylor (1974) published the oxygen isotopes and suggested an extension, hotspot- or rift-generated magmatism. Plummer (1995) illustrated the rocks are characteristics of mildly alkaline and concluded three emplacement phases, i.e. hotspot- and rift-related, within-plate magmatism. Alternatively, Tucker et al. (2001) suggested the Precambrian Seychelles magmatism is consistent with a continental arc setting than an extensional environment. He also suggested 100 Myr of magmatic activity in Seychelles based on U-Pb zircon ages of 703 Ma and his U-Pb age of 808 Ma. Based on chemical and isotopic data, Ashwal et al. (2002) supported the continental arc setting for the Precambrian Seychelles magmatism and emphasized that they were likely formed at an Andean-type arc.

Figure 1.8. Palaeogeographic reconstruction of Greater India, the Seychelles microcontinent and Madagascar [modified from Torsvik et al. (2001b)]

16

Figure 1.9. Palaeomagnetic reconstruction of the Seychelles [modified after Suwa et al. (1994)]

Palinspastic reconstructions by Torsvik et al. (2001a, b) show the spatial contiguity at

~750 Ma of the Seychelles and northwestern India. The new Seychelles - India fitting proposes that the Seychelles and the MIS, northwestern India were separated by ~600 km. The age and geochemistry of Seychelles igneous rocks by Tucker et al. (2001) are also correlative with the volcanic and plutonic rocks in the MIS. In contrast, others consider the Seychelles to be a fragment of Africa (Fig. 1.9). Rabinowitz et al. (1982) showed that Madagascar was located off the east coast of Africa bordering southern Somalia, Kenya and Tanzania before parting Gondwana. Additionally, paleomagnetic reconstruction of Seychelles suggests that it rotated by 10^o clockwise relative to Madagascar and ~17^o clockwise to Africa (Suwa et al., 1994). The Seychelles pre-drift position consequently is thought to be near the eastern end of the Horn of Africa (Hargraves and Duncan, 1990). Reconstruction of the Seychelles, therefore, first occupied the eastern of the Horn of Africa and then drifted to the currently position. Suwa et

17

al. (1983, 1994) showed petrological, geological, geochronological and geochemical data to conclude that Neporoterozoic granitic rock of the Seychelles are well correlated with and similar to those of ANS.

In this report, we used in situ zircon U-Pb geochronology and whole rock geochemistry in order to: 1) determine age of emplacement of Mahé and Praslin granites, 2) identify recycled components (inherited zircon clusters) within the granites, 3) used whole rock geochemical data and Sr-Nd isotopic data to address the origin of the Seychelles microcontinent regarding the India-Africa dichotomy and 4) address the tectonomagmatic evolution of the Mahé and Praslin granites.

18

CHAPTER 2. GEOLOGICAL BACKGROUND

The seminal work on the geology of the Seychelles was written by Baker (1963).

Baker’s field observations and maps provided a reference frame for all subsequent studies. The bedrock geology of the Seychelles is primarily comprised of late Precambrian granitic rocks (Mahé and Praslin groups) that are cross-cut by doleritic dykes. There are 42 granitic islands within the the Seychelles, in descendant order of size: Mahé, Praslin, Silhouette, La Digue, Curieuse, Félicité, Frégate, Ste. Anne, North, Cerf, Marianne, Grand Sœur, Thérèse, Aride, Conception, Petite Sœur, Cousin, Cousine, Long, Récif, Round (Praslin), Anonyme, Mamelles, Moyenne, Ile aux Vaches Marines, L'Islette, Beacon (Ile Sèche), Cachée, Cocos, Round (Mahé), L'Ilot Frégate, Booby, Chauve Souris (Mahé), Chauve Souris (Praslin), Ile La Fouche, Hodoul, L'Ilot, Rat, Souris, St. Pierre (Praslin), Zavé, Harrison Rocks (Grand Rocher). Lithified sedimentary and metamorphic rocks are absent but there is a minor volume of volcanic and volcaniclastic rocks. The youngest igneous rocks are the syenites, diorites, and rhyolites from North Island and Silhouette that yielded Early Palæogene ages (Shellnutt et al., 2017). The oldest rocks of the Seychelles are the Precambrian granites that form the principal island groups – the Mahé and Praslin groups. The granites of Mahé and Praslin groups were first suggested to be Neoproterozoic (663±17 Ma) by Miller and Mudie (1961).

19 2.1. Mahé group island

Figure 2.1. Simplified geological map of Mahé islands [modified from Suwa et al. (1983)]

Mahé is the largest and tallest island in the Seychelles at 145 km² (~27 km long, maximum width of ~11 km). The island is mountainous with a central line of peak up to 910 m elevation. The central mountain rises steeply from the sea and from the localized coastal plateau. Mahé is almost entirely composed of coarse grained granular hornblende granite with minor porphyries. Xenoliths, enclaves, and associated hybrid rocks of tonalitic, dioritic and gabbroic composition are found in localized zones on the coast. The granitic massif is cross-cut by dolerite and basalt dykes. It is necessary to mention of the variability of exposure on Mahé. The coastal exposures are excellent and tend to form low cliff and have wave-beaten surfaces. However, away from the coastline there is a slope of the red-brown soils with granite

20

boulders. The boulders are likely derived from the hinterland and were transported by landslips and soil liquefaction, or by the progressive lowering of the surface by erosion. Deeply weathered granite areas and deep-seated surface blocks conceal much of the solid geology (Baker, 1963).

The Mahé granite

The Mahé granite is the basis of the greater part of the island and is considered as a

The Mahé granite is the basis of the greater part of the island and is considered as a