1 Introduction
The West-African craton (Fig. 1) was stabilized at about 2000 ± 200 Ma, at the end of the Eburnean orogeny (Black and Fabre, 1983). It underlies a vast Neoproterozoic to Quaternary cover represented by the Taoudeni Basin. This cover is relatively homogeneous, with a lateral continuity of facies (Deynoux et al., 1978; Trompette, 1973). The Taoudeni basin consists of two circular sub-basins and four troughs (Villeneuve, 2005). The so-called Gourma basin in the South-East is one of the four troughs. The classic lithostratigraphy of the Taoudeni basin is described in the Adrar of Mauritania (Trompette, 1973). It includes a Neoproterozoic glaciogenic deposit crowned by a “cap carbonate” that is virtually continuous from the central part to the south-eastern one. The cap carbonate does not exist in the south-western part of the Taoudeni basin where there is a high relief (Villeneuve, 2006). The glaciogenic deposit is dated at about 635 Ma (Shields et al., 2007) and testifies to the Marinoan glaciation, one of the two major worldwide glaciations (the Sturtian putting into place between 750–700 Ma and Marinoan glaciations) which have been the basis for the “Snowball Earth” hypothesis (Hoffman and Schrag, 2002; Hoffman et al., 1998; Kirschvink, 1992). These worldwide glaciations may have taken place while all the continental blocks, resulting from the break-up of the Rodinia Supercontinent, were still at low latitudes (Hoffman and Schrag, 2002; Hoffman et al., 1998; Kirschvink, 1992).
Schematic geological map of the West-African craton, after Deynoux et al. (2006) simplified. 1- Archean and Paleoproterozoic basement; 2- Pan-African and Caledono-Hercynian fold belts; 3- Neoproterozoic and Paleozoic cover; 4- Post-Paleozoic cover. Sites: 1- Toussiana, 2- Guinguette, 3- Banarodougou, 4-Tiara, 5- Digouera, 6- Tin Dioulaf, 7- Tin Akof.
Carte géologique schématique du craton Ouest-Africain, d’après Deynoux et al. (2006) simplifiée. 1- Socle archéen et paléoprotérozoïque ; 2- Chaînes plissées panafricaine et calédono-hercynienne ; 3- Couverture néoprotérozoïque et paléozoïque. Sites : 1- Toussiana, 2-Guinguette, 3- Banarodougou, 4- Tiara, 5- Digouéra, 6- Tin Dioulaf, 7- Tin Akof.
In the research for the palaeolatitudes of deposition of glaciogenic rocks, paleomagnetic methods are the most reliable. They allow the geographic positions of continental blocks during their evolution in the history of the earth to be constrained. The West-African craton is characterized by a meagre palaeomagnetic database, which cannot allow the determination of its Neoproterozoic palaeolatitude with certainty. And some available Neoproterozoic poles were perhaps mostly remagnetized in the Permian (Evans, 2000; Perrin and Prévot, 1988).
The data presented in this paleomagnetic study come from samples of Neoproterozoic formations of the Gourma sub-basin and its Bobo Dioulasso prolongation and are reported in lithostratigraphic section (Fig. 2 b,c) (Miningou, 2006; Ouédraogo, 1983).
Lithostratigraphy of the Gourma and Bobo Dioulasso sub-basins: a) from Ouédraogo, 1983; b) from Miningou (2006). Sites: T- Toussiana, G- Guinguette, B- Banarodougou, Tr- Tiara, D- Digouera, TD- Tin Dioulaf, TA- Tin Akof.
Lithostratigraphie des sous-bassins du Gourma et de Bobo Dioulasso : a) d’après Ouédraogo (1983) ; b) d’après Miningou (2006). Sites : T- Toussiana, G- Guinguette, B- Banarodougou, Tr- Tiara, D- Digouéra, TD- Tin Dioulaf, TA- Tin Akof.
Two primary paleomagnetic poles, constrained by folding and reversal tests, are proposed and lead to determine the paleolatitude of the West-African craton during the Neoproterozoic.
2 Geology
The Gourma sub-basin forms the southeastern portion of the large Taoudeni basin and is bounded to the east by the nappes of the Gourma belt. In the adjacent Bobo-Dioulasso region, the lithostratigraphic sequence is composed of three unconformable groups (Fig. 2a; Ouédraogo, 1983). The lower group is about 1200 m thick and consists of fine to medium-grained sandstones or quartzitic sandstones, siltstones and conglomeratic horizons. The middle group is about 1400 m thick and contains fine to coarse-grained quartzitic sandstones, conglomerates with glaciogenic characteristics, shales, glauconitic siltstones, dolomites, stromatolite-bearing dolomitic limestones and silexites. The upper group is thinner, about 50m, and consists of coarse-grained reddish sandstones, conglomerates and fine to medium-grained quartzitic sandstones. In the Beli area, situated in the North-East of Burkina-Faso, the lithostratigraphy consists of eight formations (Fig. 2b; Miningou, 2006). The lower part is made of quartzitic sandstones with conglomeratic lenses. The middle part contains a triad, consisting of glaciogenic deposit-carbonate-silexitic complex, overlain by a sequence of shales and siltstones, including silexite lenses and limestone and sandstone intercalations. Above is a dolomitic sequence. The upper part contains folded calcschists and quartzo-schists which are perhaps thrust over the underlying formation. The whole is capped by molassic sediment.
3 Palaeomagnetism
Seven sites were sampled (Fig. 1). The 118 core samples collected were prepared in the Geophysics and Planetology laboratory at the CEREGE.
3.1 Method
The collected samples are represented by silexites, siltstones, fine-grained sandstones and carbonates. They were drilled and oriented in situ, then cut off in cylindrical specimens and stored in a non-magnetic chamber. They were subjected to thermal demagnetization to isolate their characteristic remanent magnetization (ChRM). Demagnetization in alternating fields was tried on four samples from In Tangoun and gave unsatisfactory results.
Thermal demagnetization was done using a laboratory furnace. The remanent magnetization was measured using a JR5 spinner magnetometer (Agico) and a SQUID (2G enterprises) cryogenic magnetometer. The magnetic susceptibility of each sample at ambient temperature was measured after each heating stage using a Kappabridge KLY5 (Agico) susceptibilimeter to detect possible mineralogical changes.
Two or three samples from each site have been subjected to rock magnetic studies in order to identify the main ferromagnetic carriers. Three methods were used: (1) Thermomagnetic (susceptibility/temperature) curves were done on powders using an MFK1-MFA susceptibilimeter equipped with a CS3 furnace. Samples were heated to 675 °C in air or argon (with similar results), and then cooled; (2) The thermal demagnetisations of three isothermal remanent magnetizations (IRM) (Lowrie, 1990) were measured using JR5 magnetometer. The IRM was acquired using an MMPM9 field pulse magnetizer; (3) The hysteresis loops were generated using a micromag vibrating sample magnetometer capable of reaching 1 T.
For palaeomagnetic studies, the mean direction of the different components of natural remanent magnetization (NRM) in each specimen has been computed using principal component analysis (Kirschvink, 1980). The mean direction within each site was calculated on statistics method (Fisher, 1953). The data were processed using Paleomac 6.1 software (Cogné, 2003). Statistical folding and reversal tests were made and used to characterize the relative age of magnetization.
3.2 Results
The samples possess a natural remanent magnetization (NRM) showing variable intensity (Table 1). The magnetic susceptibility, measured at ambient temperature after each heating stage, is quite stable and does not show any significant mineralogical transformation.
Directions moyennes et PGVs des sous-bassins du Gourma et Bobo-Dioulasso.
| South Virtual Geomagnetic Pole | ||||||||||||||||||||
| N/n | Jr | Mx | In situ | After dip correction | In situ | After dip correction | F.T. | R.T. | ||||||||||||
| Sites | (10-3A/m) | D̊ | I̊ | k1 | α95̊ | D̊ | I̊ | k2 | α95̊ | Lon̊ | Lat̊ | Lon̊ | Lat̊ | A95̊ | Plat̊ | γc̊ | γo̊ | |||
| Gourma sub-basin and Bobo Dioulasso | ||||||||||||||||||||
| Toussiana | 9/14 | 4.9 | h | 299.6 | –23.0 | 28.2 | 9.9 | 300.2 | –26.1 | 27.0 | 10.1 | 75.4 | –25.9 | 73.4 | –25.9 | 10.9 | –13.8 | 18.9 | 18.4 | |
| Guinguette | 8/11 | 100 | h | 323.4 | –10.5 | 10.5 | 17.9 | 322.6 | –11.5 | 10.0 | 18.5 | 71.9 | –50.1 | 71.7 | –49.1 | 18.8 | –5.8 | |||
| Banarodougou | 15/21 | 20 | m, h | 345.2 | –7.5 | 57.8 | 5.13 | 345.0 | –10.8 | 75.5 | 4.4 | 49.6 | –69.0 | 46.9 | –67.6 | 4.5 | –5.4 | |||
| Tiara | 18/20 | 5.6 | m, h | 342.8 | –6.9 | 115.3 | 3.2 | 343.1 | –2.2 | 107 | 3.2 | 54.6 | –67.5 | 59.1 | –69.2 | 3.2 | –1.1 | |||
| Digouéra | 13/20 | 4.2 | h | 345.6 | –4.1 | 12.6 | 12.1 | 346.0 | –8.6 | 13.0 | 11.9 | 52.8 | –70.5 | 47.3 | –69.2 | 12.0 | –4.3 | |||
| Tin Dioulaf | 10/14 | 1 | m, h | 129.3 | 24.8 | 12.9 | 14.0 | 130.5 | 16.2 | 48.5 | 7.0 | 63.9 | –32.7 | 68.5 | –35.8 | 7.2 | –8.3 | + | ||
| Tin Akof | 7/18 | 0.8 | m | 328.9 | 9.3 | 13.2 | 17.3 | 332.1 | –0.1 | 14.2 | 16.6 | 75.3 | –57.8 | 64.6 | –58.8 | 16.6 | –0.1 |
3.2.1 Magnetic mineralogy
Thermomagnetic curves are reversible or irreversible with a dominant effect by magnetite (Fig. 3), and with the presence of hematite. The irreversible curves show an increase in susceptibility between 400 and 500 °C during heating, and about 580 °C during cooling (Fig. 3b, c) due to mineralogical alteration. The thermal demagnetization of IRM shows the unblocking temperatures situated either at about 575 °C (Figs. 4b–d) or above 650 °C (Fig. 4a), corresponding to magnetite and hematite respectively. Goethite is sometimes identified by a drop at 120 °C in the hard fraction (Fig. 4b). On hysteresis loops (Fig. 5), minerals are still not saturated after application of 1000 mT field. There is slightly wasp-waisted form (Fig. 5a) or a larger form (Fig. 5b) due to coercivity of hematite.
Susceptibility/temperature curves of the specimens. a) Tiara 7677; b) Tiara 7685; c) Tin Dioulaf 31. Temperature is in degrees Celsius. Susceptibility is in 10−6 S.I.
Courbes de susceptibilité/température des échantillons. a) Tiara 7677 ; b) Tiara 7685 ; c) Tin Dioulaf 31. La température est en degrés Celsius. La susceptibilité est en 10−6 S.I.
Example of thermal demagnetization of three axes of isothermal remanent magnetization based on Lowrie (1990) method applied to some samples. : Hard fraction, Δ: Medium fraction; O: Soft fraction. (a) TD: Tin Dioulaf; (b) BAA: Banarodougou A; (c) BAB: Banarodougou B; (d): TR: Tiara.
Exemple de désaimantation thermique des trois axes de l’aimantation rémanente isotherme, basée sur la méthode de Lowrie (1990) appliquée à quelques échantillons. : fraction dure ; Δ : fraction moyenne ; O : fraction molle. (a) TD : Tin Dioulaf ; (b) BAA : Banarodougou A ; (c) BAB : Banarodougou B ; (d) : TR : Tiara.
Hysteresis cycles with slope correction. a: Banarodougou 19; b: Tin Dioulaf 53.
Cycles d’hystérésis avec correction de pente. a : Banarodougou 19 ; b : Tin Dioulaf 53.
3.2.2 Paleomagnetic analysis
The analysis of principal components reveals one or three components in the rocks (Fig. 6). (1) One component is identified (Fig. 6a1). (2) Two components are identified (Figs. 6a2, 6a3). (3) Three components are identified (Fig. 6a4). In the sample with two or three components, the first or the two first components are destroyed at a low to medium temperature (250–470 °C) when carried by magnetite, and at high temperature (575 °C) when carried by hematite.
Thermal demagnetization of characteristic samples: (1) Tiara, (2) Toussiana, (3) Tin Dioulaf, (4) Digouera. a) Intensity curves as a function of temperature ; b) Zijderveld's diagrams (The open/full circles correspond to projections on the vertical/horizontal plane).
Désaimantation thermique des échantillons caractéristiques : (1) Tiara, (2) Toussiana, (3) Tin Dioulaf, (4) Digouera. a) Courbes d’intensité en fonction de la température ; b) diagrammes de Zijdervelds (les cercles vides/pleins correspondent aux projections sur les plans verticaux/horizontaux).
4 Interpretation and discussion
4.1 Interpretation
The mean paleomagnetic directions are measured on the high temperature component and reported in situ and after dip correction (Table 1 and Fig. 7).
Stereographic projections of individual directions and the mean of site in the Gourma and Bobo-Dioulasso sub-basins in increasing stratigraphic order: (Fisher, 1953) calculated in restoring the data to the same polarity: Individual samples; a- Toussiana; b- Guinguette; c- Banarodougou; d- Tiara; e- Digouera; f- Tin Dioulaf; g- Tin Akof.: Current geomagnetic field.
Projections stéréographiques des directions individuelles et moyennes des sites dans les sous-bassins du Gourma et de Bobo-Dioulasso dans l’ordre stratgraphique croissant. Fisher (1953) calculé en ramenant les données à la même polarité : Échantillons individuels; a- Toussiana ; b- Guinguette ; c -Banarodougou ; d- Tiara ; e- Digouéra ; f- Tin Dioulaf ; g- Tin Akof. : Champ géomagnétique actuel.
The Toussiana site displays two groups of components, one group with normal polarity and another group with reverse polarity (Fig. 7a). A reversal test (McFadden and McElhinny, 1990) to check the lack of superimposition of several components (Henry et al., 2004) is applied at this site. The angle separating the mean direction of these two components γo = 18.4°, is lower than the critical angle γc = 18.9°. The reversal test is positive of class C (10° <γc ≤ 20°).
The Guinguette site gives a mean direction of normal polarity (Fig. 7b) with precision parameter very similar in situ and after dip correction. No statistic test is usable.
The Banarodougou, Tiara and Digouera sites gives very close mean directions (Figs. 7c–e) which is consistent with a similar stratigraphic position. These mean directions diverge from those at Toussiana and Guinguette and are between the Tin Akof direction and the present field. The Banarodougou site is located in a folded area. The mean direction shows k2 = 75.5 (after dip correction) higher than k1 = 57.8 (in situ) but the ration k2/k1 gives inconclusive tests (McElhinny, 1964; McFadden, 1990) and non-synfolding magnetization. The Banarodougou and Tiara sites display only a component of normal polarity while the two polarities are recorded in Digouera but the scattered directions of reversed polarity do not permit a conclusive reversal test.
The Tin Dioulaf site (Fig. 7f) is also located in a folded area. It represents a Marinoan cap carbonate. All directions present a reversed polarity and the mean gives precision parameters, k1 = 12.9° in situ and k2 = 48.5° after dip correction. The ratio k2/k1 = 3.75 gives conclusive tests at 95% to 99% (McElhinny, 1964) and 95% (McFadden, 1990). The synfolding test is negative. These positive fold tests shows that the magnetization at the Tin Dioulaf site of the Marinoan cap carbonate considered putting into place at about 635 Ma (Shields et al., 2007) was acquired before the Pan-African folding dated in this area between 616 ± 7 Ma and 585 ± 14 Ma (Affaton et al., 2000; Bertrand et al., 1978). This magnetization should therefore be of Neoproterozoic age.
The Tin Akof site (Fig. 7 g) shows two groups of components, one group with normal polarity and another group with reverse polarity but a reversal test (McFadden and McElhinny, 1990) is indeterminate.
The South Virtual Geomagnetic Poles (VGPs) of our sites are compared with those of other sites from the literature in West Africa and Amazonia (Fig. 8). The paleolatitudes have low values situated between 13.4°S and 0.1°S, which place the craton in subequatorial position.
Paleomagnetic Poles: Comparison of Gourma and Bobo-Dioulasso sub-basins (this study) with some African and Amazonian paleopoles presented in southern hemisphere with African APWP (Cogné, 2003).
Pôles paléomagnétiques : Comparaison des sous-bassins du Gourma et de Bobo-Dioulasso (cette étude) avec quelques paléopôles africains et amazoniens présentés dans l’hémisphère sud avec l’APWP de l’Afrique (Cogné, 2003).
4.2 Discussion
Data processing has shown that the samples carry, often, several magnetization components. The mean minerals, carriers of magnetization, are magnetite and hematite. The mean directions and VGPs are compared with those known in the West-African and Amazonian cratons (Table 2) (Perrin and Prevot, 1988; Perrin et al., 1988; Tohver et al., 2006; Trindade et al., 2003). Some poles in the West-African craton are considered to be Permian remagnetizations (Evans, 2000; Perrin and Prévot, 1988).
Pôles néoprotérozoïques de l’Afrique de l’Ouest et de l’Amazonie.
| Formation | Âge (Ma) | South Virtual Geomagnetic pole | Precision | Reference | |||
| Minimum | Maximum | Long (̊) | Lat (̊) | dp (̊) | dm (̊) | ||
| West Africa | |||||||
| Atar Group I9 | 775 | 866 | 66.9 | –34 | 4.6 | 8.6 | a, b |
| Amouslek Tuffs | 563 | 593 | 70 | –41 | – | – | c |
| Amazonia | |||||||
| Puga carbonate cap A | 580 | 593 | 69 | –51 | 6 | 9 | d |
In the investigated area, the oldest formations are those at Toussiana and Guinguette sites. These formations are considered as the lateral equivalents of the Assabet el Hassiane Group of the Adrar of Mauritania, whose age is probably posterior to 775 ± 52 Ma (K-Ar age obtained in the underlying Atar Group (Clauer et al., 1982)). The Toussiana results (Table 1) are constrained by the type C reversal positive test. Its VGP coincides with the pole of the Neoproterozoic unit I9 of the Atar Group (Table 2) considered to have been remagnetized in the Permian (Perrin and Prevot, 1988; Tohver et al., 2006). The Guinguette site gives VGP (Table 1) not constrained by statistical test.
Near the Toussiana and Atar unit I9 VGPs, is the Tin Dioulaf VGP whose direction was constrained by positive fold test (McElhinny, 1964; McFadden, 1990). The Tin Dioulaf VGP predates the folding of at least one of the Panafrican deformational phases of the Gourma fold belt that affects the “cap carbonate” and the underlying formations. It is a primary pole that postdates 635 Ma, the age of the underlying Marinoan diamictite (Deynoux et al., 2006; Shields et al., 2007). This conclusion is consistent with the position of the Tin Dioulaf VGP close to the VGP of the Puga “cap carbonate” Ad (Trindade et al., 2003) rotated to Africa coordinates using Euler pole rotation (Lawver and Scotese, 1987) see also (Tohver et al., 2006). The Puga cap carbonate Ad dated at 630-580 Ma (Trindade et al., 2003), belongs to the Amazonian craton which coincide to the West-African craton during the Neoproterozoic (Meert and Torsvik, 2003; Rino et al., 2008), after the closure of the Neoproterozoic ocean suggested between those two cratons (Klein et al., 2005; Nomade et al., 2002). The Amouslek tuff VGP, considered as Permian remagnetization, is closed to the Tin Dioulaf. This cluster allows suggesting that Amouslek tuff VGP could be also Neoproterozoic.
Banarodougou, Tiara and Digouera sites show closed VGPs (Table 1) which are situated between Tin Akof VGP and the current pole. Without paleomagnetic test these three VGPs could be related to a Late Panafrican deformation remagnetization or possibly to opening Atlantic Ocean but their proximity to the other Neoproterozoic poles permits also to consider a possible primary origin.
The Tin Akof site is stratigraphically above the “Triad”. Its VGP (Table 1) diverges from those at Toussiana and Tin Dioulaf and remains near to Puga cap carbonate Ad in Amazonia.
All the sites give relatively low palaeolatitudes (Table 1) situated between 13.4°S and 8.3°N. Such low paleolatitudes were proposed for West-African and Amazonian Neoproterozoic formations (Lefort et al., 2004; Perrin et al., 1988; Trindade et al., 2003). They determine the subequatorial position of the West-African craton during the period preceding and following the Marinoan glaciation. The Amazonian craton, which was intimately related to the one in West Africa during its history, in most reconstruction models of the Rodinia and Gondwana supercontinents (Meert and Torsvik, 2003; Rino et al., 2008), is considered to have been at low latitude during the Puga glaciation (Trindade et al., 2003). The low paleolatitudes obtained in the Gourma enhance the “Snowball Earth” hypothesis. Actually, the “Snowball Earth” hypothesis places the continents at low latitudes around the two major (Sturtian and Marinoan) Neoproterozoic glaciations that took place during the intervals 750–730 Ma and 650–635 Ma (Hoffman and Schrag, 2002; Hoffman et al., 1998; Shields et al., 2007).
5 Conclusion
This palaeomagnetic study, carried out in the South-East of Gourma and Bobo Dioulasso (South-East of the Taoudeni basin), contributes significantly to the knowledge on Neoproterozoic paleomagnetic poles in the West-African craton.
The Toussiana sandstones give a VGP dated at 775 ± 52 Ma. This Toussiana PGV in antipode coincides with the unit I9 VGP of Atar Group which was considered as remagnetized in the Permian (Perrin and Prévot, 1988; Tohver et al., 2006). The Tin Dioulaf cap carbonate defines a pole with an age close to 635 Ma, which corresponds to the age of the Marinoan tillite on which it reposes (Deynoux et al., 2006; Shields et al., 2007). The Tin Akof carbonates give a younger VGP belonging to the 635–600 Ma interval (i.e. between the age of the underlying formations and the one of the Panafrican orogeny (Affaton et al., 2000; Bertrand et al., 1978; Black et al., 1979; Caby et al., 1982). Banarodougou, Tiara and Digouera VGPs are much more doubtful; they can correspond to a period close to those of Tin Akof but are probably related to later Panafrican deformation remagnetization or a Meso-Cenozoic remagnetization.
Paleolatitudes data place the West-African craton at low latitudes during the Neoproterozoic time.
Acknowledgements
The authors would like to acknowledge the “Bureau des Mines et de la Géologie du Burkina Faso” (BUMIGEB) and the “Institut de Recherche pour le Développement” (IRD), especially A. Blot, for their important help for field work. B. Henry and M. Villeneuve are also acknowledged for their reviews of the first drafts of the manuscript.