Geoscience Frontiers 6 (2015) 557e570

HOSTED BY Contents lists available at ScienceDirect

China University of Geosciences (Beijing) Geoscience Frontiers

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

Research paper A comparative review of petrogenetic processes beneath the Volcanic Line: Geochemical constraints

Asobo N.E. Asaah a,b,*, Tetsuya Yokoyama a, Festus T. Aka c, Tomohiro Usui a, Mengnjo J. Wirmvem d, Boris Chako Tchamabe d, Takeshi Ohba d, Gregory Tanyileke c, J.V. Hell c a Department of Earth & Planetary Sciences, Tokyo Institute of Technology, 2-12-1, Ookayama, Meguro-ku, Tokyo 152-8551, Japan b Department of Mines, Ministry of Mines, Industry and Technological Development, Yaoundé, Cameroon c Institute of Mining & Geological Research, P.O. Box 4110, Yaoundé, Cameroon d Department of Chemistry, School of Science, Tokai University, Hiratsuka 259-1211, Japan article info abstract

Article history: The origin and petrogenesis of the Cameroon Volcanic Line (CVL), composed of volcanoes that form on Received 24 September 2013 both the ocean floor and the continental crust, are difficult to understand because of the diversity, het- Received in revised form erogeneity, and nature of available data. Major and trace elements, and Sr-Nd-Pb isotope data of volcanic 8 April 2014 rocks of the CVL spanning four decades have been compiled to reinterpret their origin and petrogenesis. Accepted 17 April 2014 Volcanic rocks range from nephelinite, basanite and alkali to , and with Available online 2 June 2014 the presence of a compositional gap between SiO2 58e64 wt.%. Similarities in geochemical characteristics, modeled results for two component mixing, and the existence of mantle xenoliths in most mafic rocks Keywords: fi Petrogenesis argue against signi cant crustal contamination. Major and trace element evidences indicate that the Cameroon volcanic line melting of mantle rocks to generate the CVL magma occurred dominantly in the garnet lherzolite stability Depleted MORB mantle field. Melting models suggest small degree (<3%) partial melting of mantle bearing (6e10%) garnet for Mt. Enriched mantle Etinde, the Ngaoundere Plateau and the Biu Plateau, and <5% of garnet for the oceanic sector of the CVL, Mantle heterogeneity Mt. Cameroon, Mt. Bambouto, Mt. Manengouba and the Oku Volcanic Group. The Sr-Nd-Pb isotope sys- tematics suggest that mixing in various proportions of Depleted MORB Mantle (DMM) with enriched mantle 1 and 2 (EM1 and EM2) could account for the complex isotopic characteristics of the CVL lavas. Low Mg number (Mg# ¼ 100 MgO/(MgO þ FeO)) and Ni, Cr and Co contents of the CVL maficlavas reveal their crystallization from fractionated melts. The absence of systematic variation in Nb/Ta and Zr/Hf ratios, and Sr-Nd isotope compositions between the mafic and felsic lavas indicates progressive evolution of magmas by fractional crystallization. Trace element ratios and their plots corroborate mantle het- erogeneity and reveal distinct geochemical signatures for individual the CVL volcanoes. Ó 2015, China University of Geosciences (Beijing) and Peking University. Production and hosting by Elsevier B.V. All rights reserved.

1. Introduction Masberg, 1998). Although continental and oceanic intraplate magmas show considerable variation in trace element abundances Volcanic rocks of continental rifts have received considerable and ratios, they are typically more enriched in incompatible trace attention in recent years. However, the diversity in their compo- elements than basalts erupted at subduction zones and mid-ocean sition means they are not explainable by a single process (Jung and ridges (MORs). As a result, trace element modeling for petrogenesis of intraplate magmas invariably requires their derivation from a mantle source that is chemically distinct from the upper-mantle magma source that produces the MORBs (Weaver, 1991). Conti- * Corresponding author. Department of Earth & Planetary Sciences, Tokyo nental intraplate rift magmatism is geochemically similar to ocean Institute of Technology, 2-12-1, Ookayama, Meguro-ku, Tokyo 152-8551, Japan. island basalts (OIB) (Weaver, 1991) and is thought to be generated Tel.: þ81 3 5734 3539; fax: þ81 3 5734 3538. E-mail address: [email protected] (A.N.E. Asaah). from a convectively upwelling asthenosphere during continental Peer-review under responsibility of China University of Geosciences (Beijing) extension associated with the formation of rift systems (Zou et al., http://dx.doi.org/10.1016/j.gsf.2014.04.012 1674-9871/Ó 2015, China University of Geosciences (Beijing) and Peking University. Production and hosting by Elsevier B.V. All rights reserved. 558 A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570

2000). The upper mantle beneath intraplate rift-related Cenozoic Meyers et al., 1998; King and Ritsema, 2000; Reusch et al., 2010; volcanoes is thought to have experienced metasomatism in the Milelli et al., 2012) and geochronological (Dunlop and Fitton, past (Wilson and Downes, 1991). These processes result in the 1979; Aka et al., 2004) data have been used to develop comple- production of hydrous minerals (amphibole and phlogopite), a mentary models to explain the origin of the CVL. Overall, mafic range of accessory mineral phases, and enrichment in incompatible rocks of the CVL show chemical features consistent with plume elements (Wilson and Downes, 1991; Wedepohl et al., 1994). activity outlined by their OIB major and trace elements and Unique to other intraplate rift related volcanoes, the Cameroon radiogenic isotope characteristics (Mbassa et al., 2012). Volcanic Line (CVL) in West Africa, consists of a chain of Cenozoic Geochemical data and its interpretation are challenging for the volcanoes developed on both the Atlantic oceanic floor and the CVL system, because of large variation. Lavas from the CVL range continental crust of the African plate (Fig. 1). There is no evidence of from mafic end-members e.g., nephelinite, basanite and alkali systematic age migration from one volcano to another (Lee et al., to felsic end-members e.g., , and rhyo- 1994; Marzoli et al., 2000; Aka et al., 2004) as observed in a lites. Available data show that some volcanoes (Mt. Etindé, Mt. typical hot spot settings such as the Hawaii-emperor chain. These Cameroon and Bui Plateau, Fig. 1) are composed only of maficlavas complex characteristics (clear volcano alignment but lack of a while others (Mounts Manengouba, Bambouto, the Oku Volcanic consistent time-space migration) of the CVL have generated some Group or OVG and the Ngaounderé Plateau) are bimodal, with the debate concerning its origin. Geological (Grunau et al., 1975; Gorini presence of a Daly gap. The degree of differentiation of lavas varies and Bryan, 1976; Freeth, 1978; Sibuet and Mascle, 1978; Fitton, from one volcano to another with no well-defined trend. There is 1980; Morgan, 1983; Burke, 2001), geochemical (Fitton and heterogeneity even at a local scale within each central volcano Dunlop, 1985; Deruelle et al., 1991; Lee et al., 1994; Njonfang (Nkouandou and Temdjim, 2011) which is thought to be due to a et al., 2011), structural (Moreau et al., 1987; King and Anderson, different degree and depth of partial melting (Marzoli et al., 2000; 1995), geophysical (Fairhead, 1988; Fairhead and Binks, 1991; Kamgang et al., 2013). The CVL magmatism is characterized by

Figure 1. Location of the Cameroon Volcanic Line. COB ¼ Continent Ocean Boundary, CAR ¼ Central African Republic, CASZ ¼ Central African Shear Zone, EARS ¼ East African Rift System (Modified from Marzoli et al., 2000; Rankenburg et al., 2005; Ngako et al., 2006; Nkouathio et al., 2008). A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 559 melting in the garnet lherzolite stability fields (Marzoli et al., 2000; granitic rocks. These basement rocks are either of Precambrian age Yokoyama et al., 2007; Kamgang et al., 2013), although melting in (Geze, 1943; Gouhier et al., 1974; Le Marechal, 1976; Lasserre, 1978; the spinel lherzolite stability fields have been reported in the Dunlop, 1983) or bear imprints of a penetrative Neo-Proterozoic Ngaoundere Plateau region (e.g., Lee et al., 1996; Nkouandou and Pan African deformation (Annor and Freeth, 1985; Toteu et al., Temdjim, 2011) where the lithosphere is significantly thin 2004). (Poudjom Djomani et al., 1992, 1997). In addition, a mixing of both The central volcanoes in the continental sector of the CVL garnet and spinel melting fields has been reported in Mt. Cameroon include Mounts Etindé, Cameroon, Manengouba, Bambouto, and (Tsafack et al., 2009). Both mafic and felsic rocks show chemical Oku, as well as two main plateaus; the Ngaounderé and Biu (Fitton, features consistent with plume activity outlined by their OIB 1987; Halliday et al., 1988; Marzoli et al., 1999; Marzoli et al., 2000; characters, and their isotopic ratios (Mbassa et al., 2012). Ngounouno et al., 2000; Nono et al., 2004; Rankenburg et al., 2005). Information based on trace element abundances and isotope Mount Oku forms part of the Oku range referred to in this study systematics of the CVL lavas shows that far from being a simple as the Oku Volcanic Group (OVG). Some central volcanoes alignment of individual volcanoes, the CVL is divided into three (Mt. Bambouto, the OVG, and the Ngaounderé Plateau) produce distinct sectors e the oceanic, ocean-continent boundary and the more chemically evolved lavas than others. They are dominantly continental sectors (Halliday et al., 1990; Lee et al., 1994, 1996). polygenetic, and alternate with less evolved monogenetic grabens These data can be complemented by in compatible trace elements (Sato et al., 1990; Nkouathio et al., 2008). North of the Oku massif with similar mineral/melt partition coefficients, as these are less (Fig. 1), the volcanic line bifurcates into a Y-shape; Mt. Mandara-Biu sensitive to variations in partial melting of the source mantle and Plateau branch in the north-west and Ngaounderé Plateau branch mineral crystallization in the magma plumbing system. The in the north-east (Fitton et al., 1983). Of these volcanoes, Bioko, Mt. geochemical complexity and tectonic setting of the CVL make it an Etindé, and Mt. Cameroon volcanic centers are located close to the ideal environment to investigate characteristics of source magmas seismically-defined “ocean-continent boundary” (OCB) zone in an intraplate setting, and their petrogenetic evolution. In addi- (Emery and Uchupi, 1984). tion, the parallelism (Fig. 1) between the CVL alignments and the There are four volcanoes in the oceanic sector of the CVL: central African shear zone strike (Ngako et al., 2006) could provide Annobon (813 m a.s.l.), São Tomé (2024 m a.s.l.), Principé (984 m insights into the potentials of mineral deposits associated with a.s.l.) and Bioko (2872 m a.s.l.). They are strato volcanoes (Annobon these structures (Njonfang et al., 2011). and São Tomé) and lava flows (Principé and Bioko). These volcanic More recent data from 2007e2013 (e.g., Njilah et al., 2007; islands are made up of volcanic rocks ranging from nephelinite, Yokoyama et al., 2007; Aka et al., 2008; Nkouathio et al., 2008; basanite and basalt to trachyte and phonolite (Piper and Suh et al., 2008; Tsafack et al., 2009; Kamgang et al., 2010; Richardson, 1972; Dunlop and Fitton, 1979; Cornen and Maury, Nkouandou and Temdjim, 2011; Fagny et al., 2012; Mbassa et al., 1980; Fitton and Dunlop, 1985; Fitton, 1987; Halliday et al., 1988, 2012; Tchouankoue et al., 2012; Kamgang et al., 2013) have been 1990; Deruelle et al., 1991; Lee et al., 1994). added to data sets for the CVL used in a review by Deruelle et al. Mt. Etindé is a small densely forested and highly dissected (2007). With the increase in published geochemical data available volcano located on the SW flank of Mt. Cameroon. It rises to a after this review from 2007, other models and interpretations for height of 1713 m a.s.l and is made up almost entirely of nephelinitic the CVL magmatism have been proposed. As a result, the debates on lavas rich in euhedral (3e7 mm) highly zoned clinopyroxene phe- its origin and petrogenesis are attracting more scientific interest. To nocrysts (Nkoumbou et al., 1995). enhance the debate, a framework based on updated data is needed. Mt. Cameroon is the highest (4095 m a.s.l) volcano in West Africa Therefore, the aim of this study is to (1) summarise existing models occupying a surface area of about 1300 km2. This volcano is the only for the origin of the CVL, and (2) investigate petrogenetic processes presently active member of the CVL with seven eruptions recorded of the CVL lavas including crustal contamination, nature of source in the last 100 years, i.e., 1909, 1922, 1954, 1959, 1982, 1999, and and degree of melting, mantle reservoirs and their interactions, as 2000 (Geze, 1943; Fitton et al., 1983; Suh et al., 2003). It is a com- well as the evolution from mafic to felsic lavas. Simultaneous posite volcano made of alkaline basanitic and basaltic flows inter- comparison of lavas from all the CVL central volcanoes can give bedded with small amounts of pyroclastic materials and numerous valuable information on the general nature and evolution of indi- cinder cones (Suh et al., 2003, 2008; Yokoyama et al., 2007). vidual volcanoes with respect to others along the chain. Mounts Bambouto (2700 m a.s.l.) and Oku (3011 m a.s.l) are Oligocene to Quaternary strato volcanoes with lava successions 2. Geologic and tectonic setting comprising a strongly bimodal basalt-trachyte-rhyolite suite (Marzoli et al., 2000; Njilah et al., 2004, 2007; Kamgang et al., 2010, The Cameroon Volcanic Line (CVL) is a 1600 km Y-shape chain of 2013; Tchouankoue et al., 2012). Mt. Manengouba (2411 m a.s.l) is a Tertiary to recent volcanoes (Fitton and Dunlop, 1985) that trends well preserved central volcano whose summit hosts two concentric from the Annobon Island in the Gulf of Guinea into the African plate (Elengoum and Eboga). Lavas range from basalts to tra- (Fig. 1). The tectonic setting of the CVL is described in detail in chytes, quartz trachytes, and rare (Fitton, 1987; Kagou previous studies (Grunau et al., 1975; Gorini and Bryan, 1976; Fitton et al., 2010). et al., 1983; Moreau et al., 1987; Ngako et al., 2006). The intraplate The Ngaounderé Plateau, in the northeastern part of the CVL, tectono-magmatic alignment is remarkable in that it developed consists of alkaline basalts and basanites capped by trachytes and simultaneously on the oceanic and continental sectors (Fitton, phonolitic flows. The most recent volcanism in this area consists of 1987). The CVL is oriented N30E and segmented by several cinder cones aligned in a WNWeESE direction, sometimes pro- N70E shear zones of Pan-African age (Deruelle et al., 2007), ducing small lava flows (Fitton, 1987; Nono et al., 2004). The Biu particularly evident on the continental sector. The oceanic sector Plateau, which is located in the northern part of the Ngaounderé consists of two seamounts in the Gulf of Guinea and four islands Plateau, consists of basaltic flows with a maximum thickness of (Annobon, São Tomé, Principé and Bioko) (Lee et al., 1994). On the 250 m. This plateau is composed of basanite to transitional basalts other hand, the continental sector consists of anorogenic plutonic (Rankenburg et al., 2005). complexes and volcanic constructs of varying sizes and shapes The history of tectono-magmatic activity in the CVL and adja- (Deruelle et al., 2007). Magmas of the continental CVL intruded into cent Jos Plateau and the Benoue Trough reveals four key points with the overlying Proterozoic meta-sedimentary, meta-granitic, and a general age progression from the northwest to southeast. These 560 A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 include: (1) The initial phase of magmatic activity in the Benue Trough (147 Ma; Coulon et al., 1996) occurred at about the same time as the emplacement of anorogenic granite ring complexes when the formation of the Jos plateau ended; (2) The transition from the first to second stages of magmatic activity in the Benue Trough (106e95 Ma) overlaps with the period (100 Ma) when the Equatorial Atlantic opened, leading to the separation between the South American and African plates (Wilson, 1992; Wilson and Guiraud, 1992); (3) The last stage of magmatic activity in the Benue Trough (68e49 Ma) was contemporaneous with the initial phase of intrusive activity of the continental sector of the CVL (66e30 Ma) (Lasserre, 1966); (4) Volcanic activity in the CVL (45 Maepresent, Fitton and Dunlop, 1985; Marzoli et al., 2000; Aka et al., 2004) (Fig. 1).

3. Data selection and treatment

The geochemical data used in this study consist of major and trace elements, and radiogenic isotopes (Sr-Nd-Pb) from pub- lished literature on the CVL spanning four decades. More recent data from 2007e2013 have been added to data sets for the CVL used in a previous review by Deruelle et al. (2007). A large data set of 580 samples covering the entire CVL is used here to address petrogenetic processes and source characteristics of the CVL magmas. Major elements were recalculated on a 100% volatile- free basis. Only representative internally consistent data sets were selected, with priority given to samples in which both major and trace elements and isotopes were analysed. Data for Annobon, São Tomé, Principé and Bioko that constitute the oceanic sector of the CVL are grouped and referred to as the oceanic CVL in the figures. On the other hand, data for each volcanic centre on the continental sector of the CVL (Mounts Cameroon, Etindé, Man- engouba, Bambouto, the OVG, and the Ngaounderé & the Biu Plateaux), are presented separately. PETROGRAPH (Petrelli et al., 2005), PETROMODELER (Ersoy, 2013) and AFC-Modeler (Keskin, 2013) software were used to generate diagrams and perform model calculations.

4. Geochemistry

4.1. Major elements

Representative samples (400) selected for major element sys- tematics have SiO2 ranging from 38 wt.% in a sample from the oceanic CVL to 79 wt.% in a sample from the OVG (Oku Volcanic Group). Based on the total alkali versus silica (TAS) classification (Fig. 2a; Le Bas et al., 1986), the samples representative of both the oceanic and continental sectors of the CVL are dominantly alkali basalts and basanites which are parental magmas of two major differentiation series: (1) alkali basalt e trachyandesite e trachyte erhyolite, and (2) basanite e tephrite - phonolite (Wilson, and Figure 2. (a) Total alkali vs. silica classification diagram (after Le Bas et al., 1986), (b) Downes, 1991). Nephelinite occurs in Mt. Etindé but has markedly # K2O vs. SiO2 (Le Maitre, 2002), (c) Mg vs. SiO2 for the CVL lavas. Hypothetical frac- different geochemical characteristics from the basanites and alkali tionation trends are also illustrated in Fig. 2c (subscripts 1, 2, 3 and 4 are labels to basalts. The nephelinites constitutes a separate group, not associ- differentiate dominant trends or directions identified on the diagram). Volcanic suite ated with differentiated lavas. Using the classification criteria pro- ranging from primitive basanite and alkali basalts to more evolved trachyte, phonolite and rhyolite with paucity in the trachy-andesite field. Nephelinite of Mt. Etinde con- posed by Le Maitre (2002), the lavas dominantly range from highly stitutes a separate group. (F ¼ foidite, BS ¼ basanite, PB ¼ pico-basalt, B ¼ basalt, # ¼ potassic to ultrapotassic (Fig. 2b). The Mg ( 100 MgO/ TB ¼ trachy-basalt, BA ¼ basaltic andesite, BTA ¼ basaltic trachy-andesite, (MgO þ FeO)) of mafic samples ranges from 35e69. This indicates PhT ¼ phono-tephrite, A ¼ andesite, TA ¼ trachy-andesite, TPh ¼ tephri-phonolite, that the lavas underwent some fractionation. The under-saturated Ph ¼ phonolite, T ¼ trachyte, D ¼ dacite, R ¼ rhyolite; Data source: The CVL literature nephelinite from Mt. Etindé has a lower Mg# (37e53) when data up to 2007 and recent data including: Njilah et al., 2007; Yokoyama et al., 2007; Aka et al., 2008; Nkouathio et al., 2008; Suh et al., 2008; Tsafack et al., 2009; Kamgang compared to those of basanites and alkali basalts (Fig. 2c). This et al., 2010; Nkouandou and Temdjim, 2011; Fagny et al., 2012; Mbassa et al., 2012; suggests an earlier fractionation of magnesian olivine and pyrox- Tchouankoue et al., 2012; Kamgang et al., 2013). enes (possibly augite). Hypothetical fractionation trends for the CVL lavas are illustrated in Fig. 2c (subscripts 1, 2, 3 and 4 are labels to A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 561

Table 1 Major element (wt.%) compositions (ranges) of representative CVL volcanic rocks.

t # Volcanic centres SiO2 Al2O3 Fe2O3 MgO CaO Na2OK2O TiO2 P2O5 MnO Mg Oceanic CVL 40e67 10e22 2e14 0e13 0e13 22e90e60e40e10e0.02 0e66 Mt. Etinde 39e43 13e17 11e15 3e912e16 2e60e44e51e20e0.4 37e53 Mt. Cameroon 43e50 10e17 11e14 5e12 9e14 2e51e23e40e10e0.3 46e65 Mt. Manengouba 44e54 12e18 2e14 2e12 5e12 3e61e32e30e10e0.2 35e65 Mt. Bambouto 41e70 12e20 4e14 0e11 0e11 3e81e60e40e10e0.4 1e63 OVG 42e79 10e17 1e15 0e13 0e11 2e70e60e50e20e0.3 1e65 Ngaoundere 41e68 12e22 2e15 0e11 0e11 2e11 0e60e40e10e0.4 7e68 Plateau Biu Plateau 44e53 13e16 8e13 6e13 7e11 2e411e32e30e10e0.2 58e69

differentiate supposed trends or directions identified on the Primitive-mantle normalized trace element patterns (Palme and diagram). O’Neill, 2003) for mafic rocks from the oceanic and continental The ranges for major element oxides are presented in Table 1. sectors of the CVL are similar to each other (Fig. 5). The patterns are Lavas of the oceanic CVL, Mounts Bambouto, the OVG and akin to OIB, suggesting the origin from a similar source. All the Ngaounderé Plateau have a wider range of major element oxides samples show a marked (LREEs) enrichment and a strong frac- especially for SiO2 when compared to those of Mt. Etindé, Mt. tionation of Heavy Rare Earth Elements (REEs) relative to LREEs. Cameroon, Mt. Manengouba and Biu Plateau. This suggests that the The absence of negative Eu anomaly defined by Eu* ¼ Eu(CI)/ former have experienced a higher degree of differentiation by [Sm(CI) þ Gd(CI)] (CI refers to chondrite normalized concentration) fractional crystallization than the latter. in the mafic rocks suggests an insignificant plagioclase fraction- The well-defined trends of compositional similarity among ation. These samples also show elevation of Ta and Nb compared to lavas of the oceanic and continental sectors of the CVL indicates the neighboring elements (K, La) in this diagram (Fitton and rapid ascent of magma in the continental CVL, avoiding signifi- Dunlop, 1985; Lee et al., 1994), and is similar to other ocean cant contamination from the continental crust. This is supported islands and mid ocean ridge basalts (Hofmann, 1988). Conversely, by the presence of mantle xenoliths in most basalts and basanite. island arcs and continental crustal rocks show negative Ta and Nb The observed insignificant crustal contamination of maficrocks patterns. The most striking features of Fig. 5 are: (1) higher trace of the continental sector of the CVL is reported in previous elemental abundances in nephelinite of Mt. Etindé, but similar studies (Fitton and Dunlop, 1985; Marzoli et al., 2000; abundances for highly incompatible elements (from Cs to K) Rankenburg et al., 2005; Deruelle et al., 2007), though there compared to the other CVL alkaline basalts; (2) relatively high are some who have reported crustal contamination in some lavas positive anomalies of Nb and Ta, high U and Th, alongside Ta, La, Ce, (see Section 5.2). Pb, and Eu for Mt. Cameroon and Biu Plateau which are not t Calcium oxide, MgO, Fe2O3 ,P2O5, and TiO2 decrease while Al2O3, observed for the OVG, the oceanic sector, and Mt. Manengouba Na2O, and K2O increase with increasing SiO2 content (Fig. 3aeh). lavas; and (3) the occurrence of a K-trough. Negative correlation of Na2O, K2O, and Al2O3 with SiO2 is observed The Ba/Nb ratio varies from 4.65 in a sample from Mt. for felsic samples (SiO2 >60 wt.%) (Fig. 3a, b, and f). The observed Cameroon to 15.91 in a sample from the OVG. The average Ba/Nb variation in major element oxides is consistent with the fraction- value is 10.03 for primitive mantle (McDonough and Sun, 1995) ation of (1) olivine, clinopyroxenes, titanomagnetite and apatite and 57.00 for continental crust (Rudnick and Gao, 2003). The CVL t (decreasing CaO, MgO, Fe2O3 , and TiO2) at an early crystallization samples with Ba/Nb ratios that fall out the range of the primitive stage, and (2) dominantly feldspar in later stages. The K2O content mantle (10 3) would have been affected by contamination and/or correlates positively with SiO2. This trend could indicate varying post magmatic alteration processes. This may be element enrich- degrees of melting of a common source for all the CVL lavas or ment or depletion at shallower depths within the crust. Some crustal contamination. The trend for each major oxide also varies samples from Mt. Cameroon and the Ngaounderé Plateau (and also from one volcano to the other. Biu) have Ba/Nb ratios less than 7.00 while those from the OVG, Mt. Bambouto and the oceanic CVL occur within or slightly above this 4.2. Trace elements range. In addition, Ba/Nb and La/Nb ratios show no clear rela- tionship suggesting that there is selective enrichment or removal The ranges for Ni, Cr, and Co (Ni: 90e400 ppm, Cr: 96e860 ppm, of Ba or La. Co: 36e70 ppm) in the mafic rocks overlap with those assumed for Barium, which is also a fluid mobile element like K, does not primary magmas (Ni: 300e400 ppm, Cr: 300e500 ppm, Co: show depletion. As a result, K-depletion cannot be attributed to 50e70 ppm) (Frey et al., 1978). Nickel and Cr contents decrease surface processes like weathering. However, the relatively con- with an increase in SiO2 content (Fig. 4a and b), suggesting frac- stant but low concentrations of K, coupled with high concentra- tionation of these elements along with crystallization of olivine and tions of Nb and La in alkali basalts and nephelinite suggests that K clinopyroxene. Lavas from the oceanic sector of the CVL, Mt. in the melts could have been buffered by residual amphibole, Cameroon, Biu Plateau, and Ngaounderé Plateau are relatively phlogopite and/or clinopyroxene during source enrichment pro- richer in these elements than Mt. Bambouto and the OVG. This may cesses and/or magma generation (e.g. Sun and Hanson, 1975; indicate that magmas from Mt. Bambouto and the OVG underwent Clague and Frey, 1982). Such residual K-bearing minerals are significant differentiation during their ascent, as well as storage in a accompanied by depletions of Rb and Cs in the CVL samples. In shallow magma chamber (Suh et al., 2008). Hypothetical fraction- addition, nearly constant elemental ratios of incompatible ele- ation trends for the CVL magmas are illustrated in Fig. 4b and are ments suggest that magma processes such as zone refining similar to the trends in Fig. 2c. On the other hand, the compositions melting, magma mixing, and extensive fractionation and replen- fi of Nd, La, and Rb show scattered trends with increasing SiO2 con- ishment, all of which can ef ciently fractionate incompatible ele- tent (Fig. 4c, e and f). The curved negative correlation of Sr with SiO2 ments, were not dominant processes during the generation of the would suggest fractionation of feldspars. CVL mafic lavas. 562 A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570

Figure 3. Major element (wt.% oxide) variation plots against SiO2 (wt.%) for the CVL lavas. Symbol, color and data source are same as in Fig. 2. A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 563

Figure 4. Trace element (ppm) variation plots against SiO2 (wt.%) for the CVL lavas. Symbols, color and data source are same as in Fig. 2. Hypothetical fractionation trends are also illustrated in Fig. 4b (subscripts 1, 2, 3 and 4 are labels to differentiate dominant trends or directions identified on the diagram).

4.3. Sr, Nd and Pb isotopes ratios, implying that they are more primitive than other conti- nental volcanic rocks. (Mt. Bambouto, Mt. Manengouba, and the Sr-Nd-Pb isotopic compositions of the CVL basalts overlap OVG). Isotope data for the OVG show a wider spread than those 20 those of OIB. In the 207Pb/204Pb vs. 6Pb/204Pb diagram (Fig. 6), of other CVL volcanoes. This difference is conspicuous in Pb the CVL data plot along the Northern Hemisphere Reference Line isotopes. A negative correlation is displayed between the (NHRL) of Hart (1984), and to the right of the 4.53 Ga geochron. 143Nd/144Nd and 87Sr/86Sr (Fig. 7) in which the correlation slope However, some data from the OVG with low 206Pb/204Pb and matches the mantle array of MORB-OIB samples. 207Pb/204Pb ratios overlap with MORB end members (Atlantic, On the contrary, plots of 87Sr/86Sr vs. 206Pb/204Pb (Fig. 8a and b) Indian, and Pacific MORBs). The mafic rocks show a limited range and 143Nd/144Nd vs. 206Pb/204Pb (Fig. 9a and b) show double cor- in their 87Sr/86Sr and 143Nd/144Nd ratios. The 87Sr/86Sr ratios relation (positive and negative) with different slopes, suggesting range from 0.70515 in a sample from Mt. Bambouto to 0.70286 in the involvement of enriched mantle components. Therefore, in a sample from the Biu Plateau and 143Nd/144Nd ratios vary from terms of identifying enriched mantle sources (EM1 and EM2), the 0.52771 in a sample from Mt. Cameroon to 0.51302 in a sample 87Sr/86Sr vs. 206Pb/204Pb and 143Nd/144Nd vs. 206Pb/204Pb diagrams from the Biu Plateau. Some lavas from the Biu Plateau and the are more powerful and easier than the 143Nd/144Nd vs. 87Sr/86Sr oceanic CVL show relatively low 87Sr/86Sr and high 143Nd/144Nd diagram. 564 A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570

Figure 5. Trace element patterns normalized to primitive mantle (Palme and O’Neill, 2003) for representative mafic samples. All samples show positive Ta and Nb anomalies. Data source are same as in Fig. 2.

5. Discussion 5.1.1. Structural interpretations The magmatic evolution of the CVL has been interpreted to be 5.1. Origin of the CVL related to the development of tectonic structures, including the reactivation of structures in the Cenozoic (in this case, reactivation The origin of the CVL is still controversial. Numerous models refers to the displacement of pre-existing tectonic structures and have been proposed, but there is no unifying model that can associated crustal melting) (Gorini and Bryan, 1976; Moreau et al., collectively explain the complex traits of the CVL. In the following, 1987; Fairhead, 1988; Deruelle et al., 1991). However, seismic pro- the various models that have been presented in the literature are file transects close to Annobon and São Tomé, the southwestern discussed; they are based on the geology, and geochemical, struc- most part of the CVL alignment of the oceanic sector, do not display tural, geophysical, and/or geochronological data. faults associated with the northeast-trending volcano alignment that would otherwise indicate magmatic-tectonic activity (Grunau et al., 1975). In addition, a reconstruction of the south American and African continents in the Santonian (87e83 Ma) by Sibuet and Mascle (1978) shows that the CVL is oblique to the ascension fracture system, which is located further to the southwest of the

Figure 6. 207Pb/204Pb vs. 206Pb/204Pb diagram for the CVL lavas studied. NHRL ¼ Northern Hemisphere Reference Line (Hart, 1984). Data plot on the right of the 4.53 Ga geochron. Samples are characterized by high 207Pb/204Pb and 206Pb/204Pb isotopic ratios. The OVG is distinct with a wide range in Pb isotopic ratios. Data source Figure 7. 143Nd/144Nd vs. 87Sr/86Sr diagram for the CVL lavas studied. Positive corre- is same as in Fig. 2. lation with different slopes for each volcanic centre. Data source is same as in Fig. 2. A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 565

Figure 8. 87Sr/86Sr vs. 206Pb/204Pb diagram for the CVL lavas studied. (a) The CVL lavas 143 144 206 204 plot dominantly in the FOZO field. High 87Sr/86Sr ratio for the OVG lavas. (b) Double Figure 9. (a) Nd/ Nd vs. Pb/ Pb diagram for the CVL lavas studied. (b) Double correlation (negative and positive) trends for the CVL lavas. Data source is same as in correlation (positive and negative) trends for the CVL lavas with a point of intersection. Fig. 2. Data source is same as in Fig. 2.

5.1.2. Displacement of the African plate alignment (Moreau et al., 1987; Fairhead, 1988; Fairhead and Binks, This model is based on the size and shape similarities between 1991). Thus, the fault-reactivation hypothesis to explain the gen- the CVL and the adjacent Benoue Trough (Fitton, 1980). Both eration of the CVL magmas in the southwestern-most part of the geologic features are Y-shaped, and can be superimposed by oceanic sector is difficult to confirm. In addition, if the reactivation rotating one with respect to the other. According to this model, model was possible, then crustal contamination of maficlavas there was a motion of the African Plate with respect to the e would be more evident. Also, the Central African Shear Zone (CASZ) asthenosphere (80 65 Ma) such that the hot asthenospheric with extension to the Central African Republic should have a wedge, which should have underlain the Benue Trough, was ori- complete alignment of volcanoes (i.e., there is a gap between Mt. ented beneath Cameroon and the Gulf of Guinea. Therefore, it Bambouto and the Ngaounderé Plateau (Fig. 1). Therefore, it seems suggests that there was migration of magmatism from the Jos fi that basement tectonic structures rather played a role in chan- plateau to the Benue Trough and nally to the currently active CVL. neling magmas to the surface than in the generation of the This is in agreement with the decrease in age of magmatic activity magmas. The alternating horst (islands and continental volcanoes) from Jos Plateau to the CVL (Fig. 1). and graben structure of the CVL (Deruelle et al., 1987, 1991)is related to membrane stresses developed as the African plate moved 5.1.3. Hotspots and hotlines away from the Equator (Freeth, 1978). However, according to this The fact that the CVL volcanoes are neatly aligned led to the “ ’’ model, the intrusive complexes were emplaced when the area hypothesis. According to this hypothesis, the CVL origi- should have been under compression. nated from the movement of the African plate over a stationary Based on the similarity in the structure of the Benoue Trough, sub-lithospheric hotspot that was periodically fed and melted by a the CVL and the East African Rift System with its extension in South deep mantle plume head (Morgan, 1983; Lee et al., 1994). The Africa, Njonfang et al. (2011) suggested a mechanism of episodic hotspot model is not applicable to the CVL because: (1) Geochro- emplacement of alkaline magmas through the entire African nologic data do not show any systematic age migration of volca- continent. The alignment of magmatic complexes resulted from a nism (Fitton and Dunlop, 1985; Lee et al., 1994; Aka et al., 2004), complex interaction between hot spots (mantle plumes acting in and (2) the only presently active volcano (Mt. Cameroon) is located succession) and lithospheric fractures (including Precambrian at the center of the line (Fig.1). Because of this, an alternative model faults) during African plate motion. to explain volcanic alignment along the CVL, referred to as the 566 A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570

‘hotline’ has been proposed (Meyers et al., 1998). According to this model, supported by seismic and gravimetric data in the oceanic sector, the mantle plume independently supplies each volcano of the CVL (Deruelle et al., 1991; Meyers et al., 1998).

5.1.4. Plate-wide swells Burke (2001) suggested a plate-wide shallow-mantle convec- tion system localized under the zone of extension. According to this model, the CVL originates from a series of three distinct phenom- ena (Burke, 2001). First, its position over the so-called 711 plume (‘711’ indicates latitude 70N and longitude 11500E) for the past 140 million years. Subsequently, the location shifted at a right- angle bend with respect to the continental margin, remaining at this position for nearly 125 million years. Finally, a new plate-wide pattern of shallow-mantle convection was established at 30 Ma when the African plate came to rest.

5.1.5. Edge convection and lithospheric instability Figure 10. Nb/Ta vs. SiO2 (wt.%) diagram for the CVL lavas. Limited variation in Nb/Ta Recently, a new geophysical model has been developed to ratios for most samples. Symbols, color and data source are same as in Fig. 2. explain the origin of the CVL. Based on evidence from seismic wave behavior in the upper mantle, Reusch et al. (2010) proposed an ‘edge convection’ model that is driven by lateral variations of lith- ospheric thickness. According to this model, lower temperatures in high (22e97) e.g., Mt. Bambouto, the oceanic CVL, the Ngaoundere thick continental roots compared to those of the adjacent thinner Plateau, and the OVG. Fig. 2c shows hypothetical multiple fractional lithosphere generate an edge ‘downwelling’ which feeds an ‘up- crystallization trends (FC1e4) for all the CVL lavas. FC1 and FC2 are welling’ beneath the thinner region (King and Anderson,1995; King the most common fractionation trends for the CVL lavas. It is and Ritsema, 2000). The structure of the CVL, which includes the exhibited by the oceanic CVL, Mt Cameroon, Mt. Manengouba, Mt. alternating horsts (main volcanoes) and grabens (plains), may Bambouto, the OVG and the Ngaoundere Plateau. FC3 and FC4 support this model. In an attempt to improve upon the work of trends are typical of Mt. Etinde and the Biu Plateau respectively. Reusch et al. (2010), Milelli et al. (2012) designed a new laboratory- Nevertheless, trends of FC1e3 show similar source characteristics based model, known as ‘lithospheric instability’. Their work in- and indicates that they might have resulted from a similar up- dicates that there may be instability within the sub-continental welling mantle plume, with subsequent divergence in evolution lithospheric mantle at the edge of a continent. due to different conditions. Magmas traveling to the surface from the mantle may become 5.1.6. Summary of current CVL models contaminated by assimilation of crustal material. However, the The models presented above are complimentary, each covering importance of crustal contamination and processes such as FC and the various aspects of the CVL origin and its evolution, including AFC in producing the variety of magma composition remains tectono-magmatic activity through time, and the manifestation of controversial (e.g McBirney and Noyes, 1979). The striking simi- plumes including convection. The ‘hotline’ model collectively ad- larity in Sr-Nd isotopic compositions between highly mafic rocks of dresses the various aspects of the CVL. This model invokes multiple the continental and oceanic sectors of the CVL strongly argues plumes originating from the same source (the upper mantle), each against significant crustal contamination. In addition, the existence of which manifests independently to form volcanoes at the surface of mantle xenoliths in most of the CVL mafic lavas (Aka et al., 2004, of the Earth. It should be noted that this model is not only consis- 2008; Temdjim et al., 2004) suggests that magmas for these rocks tent with the geophysical, structural and geochemical information, ascended rapidly, thus avoiding significant contamination and it also may explain the observed absence in the migration of vol- extensive fractional crystallization in crustal magma chambers. canic activity through time from one volcano to another (Fitton, However, in some samples (more evolved lavas) the possibility of 1987). This model involves the interaction of deep mantle plumes crustal contamination is likely, especially in samples from Mt. with the subcontinental lithospheric mantle. Bambouto and the OVG (Marzoli et al., 2000; Deruelle et al., 2007; Kamgang et al., 2008). This corroborates with the relatively high Sr 5.2. Assimilation and fractionation crystallization (AFC) isotopic compositions in these lavas (e.g. 0.70472 for the OVG and 0.70512 for Mt. Bambouto). This range is higher than the Sr isotopic Apart from using isotopes to determine genetic relationships in ratios for the oceanic CVL, and some of the continental CVL lavas volcanic suites, the systematic behavior of some highly incompat- (Lee et al., 1994; Rankenburg et al., 2005; Yokoyama et al., 2007). ible element ratios including Nb/Ta, Sr/Pr, Zr/Hf, Ce/Pb, Nb/U are The assimilation fractional crystallization (AFC) model using the powerful indicators. These ratios do not change significantly in equations of DePaolo (1981) was applied to CVL mafic and felsic lavas from the same source. Nb/Ta ratio for more than 90% of both lavas. The least fractionated sample P18 (MgO ¼ 12%, SiO2 ¼ 38%) mafic and felsic CVL lavas range from 13e18 with a mean of 16. was selected to represent the initial magma. In the model, trace fi AFC There is no systematic variation in this ratio from ma c to felsic element composition of a magma affected by AFC process (Clc )is lavas thus the lavas plot roughly horizontally on the Nb/Ta vs. SiO2 expressed by: diagram (Fig. 10). This indicates a genetic relationship amongst the " # fi ma c and felsic lavas of the CVL. AFC ¼ f z þ r Ca z Clc C0 F 1 F (1) The degree of differentiation, expressed by the differentiation r 1 zCf index (DI) for the CVL lavas varies from one volcano to another. 0 ¼ e f However, we group them into: low (DI 19 63) e.g., Mt. where C0 is the initial trace element composition of an element in Cameroon, Mt. Etinde, the Biu Plateau, and Mt. Manengouba; and magma, F is the mass fraction of the residual magma, and Ca is the A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 567 concentration of the element in the assimilated material (wall isotope systematic (Rankenburg et al., 2005; Yokoyama et al., 2007). rock). The r value describes the relative ratio of assimilated material Major and trace elements in most of the CVL mafic lavas are char- to crystallized material and is expressed by: acterized by elevated CaO/Al2O3, (La/Yb)N, Zr/Y ratios with nearly constant Y and Yb concentrations suggesting that the parental m r ¼ a (2) magmas originated from a dominantly garnet-bearing mantle mc source. However, some exceptional cases of melting in the spinel where ma is the amount of assimilated material and mc is the lherzolite have been reported in the continental sector of the CVL amount of crystallized material. The z value in the AFC equation is (e.g Lee et al., 1996; Nkouandou and Temdjim, 2011), suggesting expressed as: some degree of heterogeneity in the upper mantle of the conti- nental sector of the CVL which is controlled by the depth of melting. r þ D 1 During partial melting of mantle peridotite, incompatible ele- z ¼ 0 (3) r 1 ments are concentrated in the liquid phase. When two incompat- ible elements have identical mineral/melt partition coefficients (Kd) where, D0 is the bulk partition coefficient for the mineral during partial melting, their abundance ratio in the magma reflects assemblage. that of their source, regardless of the degree of partial melting; i.e., Modeled results using Rb and Zr (highly incompatible elements their abundance ratios do not vary with increasing concentrations with large contrast between initial magma and crustal material) are in the magma (Sun and McDonough, 1989). In contrast, the ratios of shown in Fig. 11. Most mafic lavas plot around the initial magma incompatible elements having different K values would vary along composition indicating low degree fractional crystallization and d with the degree of partial melting. The more incompatible an insignificant assimilation. However, evolved samples (increasing Zr element is, the more it will be preferentially enriched in the melt. and Rb) indicate crustal assimilation and plot within the AFC curves Based on this assumption, the mantle melting process was (r ¼ 0tor ¼ 0.85). Majority of samples in this range plot between modeled for the CVL lavas using Zr/Nb (high field strength elements r ¼ 0 (no contamination) and r ¼ 0.25 (less than 25% contamina- that are highly incompatible and are not fractionated during tion). Some samples indicate contamination of up to 70% and above. magmatic processes) and La/Yb (La is incompatible while Yb is Lavas most affected by AFC include the oceanic CVL, the Ngaoun- compatible and is fractionated by garnet). These parameters are dere Plateau, Mt. Etinde, the OVG and Mt. Bambouto. The source of good in distinguishing melting of spinel and garnet lherzolite crustal input to the oceanic CVL lavas is not clear at moment. mantle. The model calculations indicate that melting occurs in the However, different hypothesis may be included: (1) convection in garnet stability field (Fig. 12). However, the percentage of garnet in the upper mantle, (2) infiltrating fluids and (3) contamination from the mantle varies alongside the degree of melting. With the sediments or during sampling. On the other hand, contamination in exception of Mt. Etinde, the Ngaoundere Plateau and the Biu lavas from continental volcanoes arise from assimilation of wall Plateau, where the melts were generated in the presence of 6e10% rock during magma ascend through the continental crust or resi- garnet, melts in the other CVL volcanoes were generated below 5% dence at shallower magma chambers. garnet by a small (less than 3%) degree partial melting of the mantle source (Fig. 12). 5.3. Nature of the source and melting The high concentration of REEs in the nephelinites of Mt. Etindé is consistent with a low degree of partial melting in the presence of Previous studies have concluded that the CVL basic lavas were derived from a similar metasomatized asthenospheric mantle. These findings were based on Sr-Nd-Pb isotopes (Halliday et al., 1990; Lee et al., 1994), Hf isotopes (Ballentine et al., 1997), trace element variations (Marzoli et al., 2000) and U-Th-Pb and Os

Figure 12. Modeled melting (Zr/Nb vs. La/Yb) results for the CVL mafic rocks with MgO>4. Melts that produced most basanites and alkali basalts were generated by <3% Figure 11. Modeled AFC (Rb/Zr vs. Zr) for the CVL mafic and felsic lavas. Symbols, color partial melting of a dominantly garnet (<6%) bearing mantle lherzolite. Symbols, and data source are same as in Fig. 2. colors and data source are same as in Fig. 2. 568 A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 more garnet at relatively deeper levels than the basanites and alkali Contributions of EM-type end members to DMM accounts for basalts. the complex Sr-Nd-Pb characteristics of the CVL lavas. Mt. Cameroon and Mt. Etinde lavas show greater contribution from EM2. An intriguing aspect is the simultaneous involvement in both 5.4. Mantle source characteristics cases described above in individual volcanic centers of the CVL, especially in the Biu Plateau and the oceanic CVL. This, coupled with Discrete mantle components have been identified from isotopic trace element variations, suggests mantle heterogeneity at a local studies of MORB and OIB (Zindler and Hart, 1986). The following (single volcano) scale. The intersection between the two slopes in end members as defined in Stracke et al. (2005) will be used in this both Figs. 8 and 9 are characterized by a zone with the lowest paper; these are depleted MORB mantle (DMM), high-m (HIMU), 87Sr/86Sr, highest 143Nd/144Nd, and intermediate 206Pb/204Pb. This Enriched mantle 1 and 2 (EM1 and EM2), and Focal Zone (FOZO). zone might represent the isotopic composition for the common Combined Sr, Nd and Pb isotope studies of the CVL basalts mantle end-member for the CVL basalts (parent magma). It is thus demonstrate that the observed isotopic variations in these rocks reasonable to infer that this common mantle component is the cannot be generated by a single mantle process or mixing of two asthenospheric mantle source. This zone constitutes lavas domi- kinds of mantle components with well defined isotopic character- nantly from the Biu Plateau, the oceanic CVL and Mt. Manengouba istics. Based on the enriched mantle (EM1 and EM2) characteristics, which are considered to be less evolved. The OVG is very peculiar Zou et al. (2000) suggested that mixing between DMM and EM2 with its wide range in isotopic data. It presents an array which produces a positive correlation in the 87Sr/86Sr vs. 206Pb/204Pb di- extends towards EM1 and thus indicates the involvement of agram, and a negative correlation in the 143Nd/144Nd vs. 206Pb/204Pb dominantly EM1-like material during recycling. diagram and the reverse is true for the mixing between DMM and Unlike MORBs with a homogeneous chemical composition of the EM1. Unlike basalts from SE China where this assumption is upper mantle, the CVL magmas are plumes derived from deeper simplistic (two-component mixing) (Zou et al., 2000), basalts of the levels. The presence of mantle xenoliths in most alkaline basalts of CVL show a more complex pattern with the contribution of more the CVL suggests that the plumes from deeper levels, entrained than two components. Mixing between two end members (e.g., peridotite xenoliths at shallower levels (upper mantle). The inter- DMM with EM1 or EM2) will produce a linear array. However, the action of the plumes with the upper mantle at varying degrees CVL lavas do not plot on any of these arrays and cannot be imparts additional signatures to the magma and thus accounts for explained by only two component mixing. Nevertheless, in a three heterogeneity within the volcanoes. This, therefore, implies that component mixing, the proportion of each end member (EM1 and multiple plume strands with a homogenous composition interacted EM2) may significantly change the mixing slopes. with a heterogeneous subcontinental lithospheric (SCLM) mantle at In the 143Nd/144Nd vs. 87Sr/86Sr diagram (Fig. 7) all plots show varying degrees. The isotopic composition of the SCLM is locally a negative slope. Conversely, in the 87Sr/86Sr vs. 206Pb/204Pb different because of the complex history of mantle metasomatism. (Fig. 8)and143Nd/144Nd vs. 206Pb/204Pb (Fig. 9) diagrams, double The insignificant crustal contamination in mafic rocks suggests that contrasting slopes; negative followed by a positive slope in Fig. 8a both EM1 and EM2 components owe their origin to metasomatism and b, and positive followed by a negative slope in Fig. 9a and b and of the SCLM. However, the existence of multiple chemical res- are observed. These contrasting slopes suggest different degrees ervoirs is indicative that the mantle is not mixed up as would be the of contributions from EM1 and EM2 to DMM in the generation of case in mid-ocean ridges, where the whole mantle is convecting. the CVL magmas. In the first case (negative slope in 87Sr/86Sr vs. 206Pb/204Pb and positive slope in 143Nd/144Nd vs. 206Pb/204Pb), there is a greater contribution of EM1 to DMM when compared to 6. Conclusions the contribution from EM2. This leads to progressive depletion in 87Sr/86Sr, but increasing 143Nd/144Nd with an increase in This study includes the compilation, assessment, and interpre- 206Pb/204Pb. At the intersection point, equilibrium in contribu- tation of existing geochemical data of the ocean and continental tions from both EM1 and EM2 to DMM is attained (Figs. 8b and sectors of the CVL to investigate petrogenetic processes and mantle 9b). In the second case (positive slope in the 87Sr/86Sr vs. sources in the volcanic system. Main results are: 206Pb/204Pb and negative slope in 143Nd/144Nd vs. 206Pb/204Pb), the role of EM2 becomes dominant over EM1 and thus 87Sr/86Sr (1) The origin of the CVL stems from a series of related events, starts increasing while 143Nd/144Nd decreases with an increase in including mantle plumes and lithospheric structures. 206Pb/204Pb. Geochemical perspective is the most optimal vantage on the Even though basalts that plot on the recently redefined FOZO origin and evolution of the CVL. zone (Stracke et al., 2005) were initially interpreted as originating (2) The CVL lavas were not significantly contaminated by the from a HIMU source probably due to metasomatism from St. Helena continental crust. The CVL magmas were generated by small (Halliday et al., 1990; Lee et al., 1994), the initial concept lacks (<3%), but different degrees of partial melting of a dominantly sufficient evidence to support their high Pb isotope signature. garnet lherzolite (<10%) mantle. This accounts for the minor Although FOZO and HIMU may have similar sources and close variations in geochemical characteristics at both local (scale of characteristics, FOZO differs from HIMU in having lower individual volcano) and regional (entire CVL) scales 206,207,208Pb/204Pb, but higher 207Pb/206Pb and 208Pb/206Pb ratios. (3) Contribution in various proportions of EM1 and EM2 compo- 87Sr/86Sr ratios in FOZO basalts are higher than in HIMU basalts and nents to DMM accounts for the complex isotopic characteristics positively correlate with 206,207,208Pb/204Pb ratios. These charac- of the CVL lavas. The FOZO characteristics of the CVL can be teristics are observed in the CVL lavas formed by dominant explained by mixing between DMM and EM2. The isotopic contribution of EM2 to DMM (dominantly Mt. Cameroon, Mt. composition for the common end-member for the CVL basalts Etinde, the Oceanic CVL and Biu plateau). In contrast, HIMU-type is defined by the zone of intersection between the mixing OIBs seems to be a distinct group of OIBs, which only occurs at trends of DMM with EM1 and DMM with EM2. This zone two intra-oceanic localities, St. Helena in the Atlantic Ocean and the consisting of less evolved magmas might represent the isotopic Cook-Austral island chain in the South Pacific Ocean (Stracke et al., composition for the common mantle end-member for the CVL 2005). basalts. A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570 569

(4) The evolution of the CVL magmas was controlled by a small Lubala, R.T. (Eds.), Magmatism in Extensional Structural Settings: The Phaner- e degree fractional crystallization (for mafic lavas) and AFC from ozoic African Plate. Springer, Berlin, pp. 275 327. fi Deruelle, B., Ngounouno, I., Demaiffe, D., 2007. The Cameroon Hot Line (CHL): a ma c to felsic lavas. Four major fractional crystallization trends unique example of active alkaline intraplate structure in both oceanic and e FC1e4 characterize the differentiation pattern for the CVL lavas. continental lithospheres. Comptes Rendus Geoscience 339, 589 600. Dunlop, H.M., 1983. Strontium Isotope Geochemistry and Potassium e Argon Studies on Volcanic Rocks from the Cameroon-Line, West Africa. Ph.D Thesis. Although the entire CVL has been fairly studied, intriguing Univ. Edinburgh, 347 pp. questions still remain unanswered. The OVG which is located close Dunlop, H.M., Fitton, J.D., 1979. A K-Ar and Sr-isotope study of the volcanic rocks of to the junction where the CVL bifurcates and also hosts the infa- the island of principle, West Africa e evidence for mantle heterogeneity beneath the Gulf of Guinea. Contributions to Mineralogy and Petrology 71, mous , presents the most variable geochemical charac- 125e131. teristics. Existing Sr-Nd-Pb data for the OVG show wide variation Emery, K.O., Uchupi, E., 1984. The Geology of the Atlantic Ocean. Springer, New York. with arrays that are uncommon in other CVL volcanoes. However, it Ersoy, E.Y., 2013. PETROMODELER (Petrological Modeler): a Microsoft Excel is one of the least studied along the CVL in terms of isotope sys- spreadsheet program for modelling melting, mixing, crystallization and assimilation processes in magmatic systems. Turkish Journal of Earth Sciences tematics. Future updates of isotopic data, especially on the entirely 22, 115e125. continental volcanoes, would significantly complement existing Fagny, A.M., Nkouandou, O.F., Déruelle, B., Ngounouno, I., 2012. Revised petrology studies to better constrain the origin of the CVL and its petrogen- and new chronological data on the peralkaline felsic lavas of Ngaoundéré volcanism (Adamawa plateau, Cameroon, Central Africa): evidence of open- esis. Application of statistical methods like independent compo- system magmatic processes. Analele Stiintifice ale Universitatii “Al. I. Cuza” nent analysis and principal component analysis might be a din Iasi, Seria Geologie 58 (2), 5e22. worthwhile additional analytical tool in unraveling valuable infor- Fairhead, J.D., 1988. Mesozoic plate tectonic reconstructions of central South Atlantic Ocean: the role of West and Central Rift system. Tectonophysics 155, mation. This work is on-going. 181e191. Fairhead, J.D., Binks, R.M., 1991. Differential opening of the Central and South Atlantic Oceans and the opening of the West African system. Tectonophysics Acknowledgments 187, 191e203. Fitton, J.D., 1980. The Benue Trough and the e a migrating rift system e The numerous sources of published data on the Cameroon Vol- in West Africa. Earth and Planetary Science Letters 51, 132 138. Fitton, J.G., 1987. The Cameroon Line, West Africa: a comparison between oceanic canic Line used in this paper are heartily acknowledged. Nick M.W. and continental alkaline volcanism. In: Fitton, J.G., Upton, B.G.J. (Eds.), Roberts and two anonymous reviewers are thanked for handling and Alkaline Igneous Rocks, 30. Geological Society London Special Publications, e reviewing the manuscript. We are grateful to James DOHM for pp. 273 291. Fitton, J.D., Dunlop, H.M., 1985. The Cameroon line, West Africa and its bearing on helpful discussions and corrections made on the earlier version of the origin of oceanic and continental alkali basalts. Earth and Planetary Science the manuscript. This work is part of a running Ph.D by ANEA sup- Letters 72, 23e38. ported by Science and Technology Research Partnership for Sus- Fitton, J.D., Kilburn, C.R.J., Thirwall, M.F., Hughes, D.J., 1983. 1982 eruption of Mt. Cameroon, West Africa. Nature 306, 327e332. tainable Development (SATREPS) project titled: Magmatic Fluid Freeth, S.J., 1978. A model for tectonic activity in West Africa and the Gulf of Guinea Supply into Lakes Nyos and Monoun, and Mitigation of Natural Di- during the last 90 m.y. based on membrane tectonics. Geology Rundsch 67, sasters through capacity building in Cameroon. Material and 675e688. fi Frey, F.A., Green, D.H., Roy, S.D., 1978. Integrated models of basalt petrogenesis: a nancial support is being provided by the Japan Science and Tech- study of quartz tholeiites to olivine melilitites from South Eastern Australia uti- nology Agency (JST), Japan International Cooperation Agency (JICA) lizing geochemical and experimental petrological data. Journal of Petrology 19, and the Institute of Geological and Mining Research (IRGM) of the 463e513. fi Geze, B., 1943. Geographie physique et geologie du Cameroun occidental, 17. Mem. Cameroon Ministry of Scienti c Research and Innovation (MINRESI). Mus. natl Hist. nat. Paris, Nouv. Ser, pp. 1e272. Gorini, M.A., Bryan, G.M., 1976. The tectonic fabric of Equatorial Atlantic and adjoining continental margins: Gulf of Guinea to northeastern Brazil. Anais da Appendix A. Supplementary data Academia Brasileira de Ciências 48, 10e119. Gouhier, J., Nougier, J., Nougier, R., 1974. Contribution a l’etude volcanologique du ” e Supplementary data related to this article can be found at http:// Cameroun (Ligne du Cameroun Adamaoua). Annales de la Faculté des Sci- ences, 3e48. dx.doi.org/10.1016/j.gsf.2014.04.012. Grunau, H.R., Lehner, P., Cleintuar, M.R., Allenbach, A., Bakker, G., 1975. New radiometric ages and seismic data from Fuerteventura (Canary Islands), Maio (Cape Verde Island) and Sao Tome (Gulf of Guinea). In: Borradaile, G.J., References Ritsema, A.R., Rondeel, H.E., Simon, O.J. (Eds.), Progress in Geodynamics. North- Holland, Amsterdam, pp. 90e118. Aka, F.T., Nagao, K., Kusakabe, M., Sumino, H., Tanyileke, G., Ateba, B., Hell, J.V., 2004. Halliday, A.N., Dicken, A.P., Fallick, A.E., Fitton, J.D., 1988. Mantle dynamics: a Nd, Sr, Symmetrical helium isotope distribution on the Cameroon Volcanic Line. West Pb and Os isotope study of the Cameroon line volcanic chain. Journal of Africa Chemical Geology 203, 205e223. Petrology 29, 181e211. Aka, F.T., Yokoyama, T., Kusakabe, M., Nakamura, E., Tanyileke, G., Ateba, B., Halliday, A.N., Davidson, J.P., Holden, P., Dewolf, C.P., Lee, D.-C., Fitton, D.G., 1990. Ngako, V., Nnange, J.M., Hell, J.V., 2008. U-series dating of Lake Nyos maar Trace element fractionation in plume and the origin of HIMU mantle beneath basalts, Cameroon (West Africa); implications for potential hazards on the Lake the Cameroon Line. Nature 347, 523e528. Nyos dam. Journal of Volcanology and Geothermal Research 176, 212e224. Hart, S.R., 1984. A large-scale isotope anomaly in the Southern Hemisphere mantle. Annor, A.E., Freeth, S.J., 1985. Thermo-tectonic evolution of the basement complex Nature 309, 753e757. around Okene, Nigeria, with special reference to deformation mechanism. Hofmann, A.W., 1988. Chemical differentiation of the Earth: the relationship be- Precambrian Research 28, 269e281. tween mantle, continental crust, and the oceanic crust. Earth and Planetary Ballentine, C.J., Lee, D.C., Halliday, A.N., 1997. Hafnium isotopic studies of the Science Letters 90, 297e314. Cameroon Line and new HIMU paradoxes. Chemical Geology 139, 111e124. Jung, S., Masberg, P., 1998. Major-and-trace element systematics and isotope Burke, K., 2001. Origin of the Cameroon Line of volcano-capped swells. The Journal geochemistry of Cenozoic mafic volcanic rocks from the Vogelsberg (Central of Geology 109, 349e362. Germany) e constraints on the origin of continental alkaline and tholeiitic Clague, D.A., Frey, F.A., 1982. Petrology and trace element geochemistry of the basalts and their mantle sources. Journal of Volcanology and Geothermal Honolulu Volcanics, Oahu: implications for the oceanic mantle below Hawaii. Research 86, 151e177. Journal of Petrology 23, 447e504. Kagou, D.A., Nkouathio, D., Pouclet, A., Bardintzeff, J.M., Wandji, P., Nono, A., Cornen, G., Maury, R.C., 1980. Petrology of the volcanic island of Annobon, Gulf of Guillou, H., 2010. The discovery of late Quaternary basalt on Mount Bambouto: Guinea. Marine Geology 36, 253e267. implications for recent widespread volcanic activity in the southern Cameroon DePaolo, D.J., 1981. Trace elements and isotopic effects of combined wallrock Line. Journal of African Earth Sciences 57, 96e108. assimilation and fractional crystallization. Earth and Planetary Science Letters Kamgang, P., Chazot, G., Njonfang, E., Tchoua, F., 2008. Geochemistry and 53, 189e202. geochronology of mafic rocks from Bamenda Mountains (Cameroon): source Deruelle, B., N’ni, J., Kambou, R., 1987. Mount Cameroon: an active volcano of the composition and crustal contamination along the Cameroon Volcanic Line. Cameroon Line. Journal of African Earth Sciences 6, 197e214. Comptes Rendus Geoscience 340, 850e857. Deruelle, B., Moreau, C., Nkoubou, C., Kambou, R., Lissom, J., Njonfang, E., Kamgang, P., Njonfang, E., Nono, A., Gountie, D.M., Tchoua, F., 2010. Petrogenesis of a Ghogomu, R.T., Nono, A., 1991. The Cameroon line: a review. In: Kampunzu, A.B., silicic magma system: geochemical evidence from Bamenda Mountains, NW 570 A.N.E. Asaah et al. / Geoscience Frontiers 6 (2015) 557e570

Cameroon, Cameroon Volcanic Line. Journal of African Earth Sciences 58, the Cenozoic Cameroon Line (Africa): implications for tectonic evolution and 285e304. mantle source composition. Mineralogy and Petrology 94, 287e303. Kamgang, P., Chazot, G., Njonfang, E., Ngongang, N.B.T., Tchoua, F.M., 2013. Mantle Nkoumbou, C., Deruelle, B., Velde, D., 1995. Petrology of Mt. Etinde Nephelinite sources and magma evolution beneath the Cameroon Volcanic Line: series. Journal of Petrology 36, 373e395. geochemistry of mafic rocks from the Bamenda Mountains (NW Cameroon). Nono, A., Njonfang, E., Kagou Dongmo, A., Nkouathio, D.G., Tchoua, F.M., 2004. Gondwana Research 24, 727e741. Pyroclastic deposits of the Bambouto volcano (Cameroon Line, Central Africa): Keskin, M., 2013. AFC-Modeler: a Microsoft Excel workbook program for modeling evidence of an initial strombolian phase. Journal of African Earth Sciences 39, assimilation combined with fractional crystallization (AFC) process in magmatic 409e414. systems by using equations of DePaolo (1981). Turkish Journal of Earth Sciences Palme, H., O’Neill, H.StC., 2003. Cosmochemical Estimates of mantle composition. 22, 304e319. In: Carlson, R.W. (Eds.). The Mantle and Core. In: Holland, H.D., Turekian, K.K. King, S.D., Anderson, D.L., 1995. An alternative mechanism of flood basalt formation. (Eds.), Treatise on Geochemistry, vol. 2, Elsevier-Pergamon, Oxford, pp. 1e38. Earth and Planetary Science Letters 136, 269e279. Petrelli, M., Poli, G., Perugini, D., Peccerillo, A., 2005. Petrograph: a new software to King, S.D., Ritsema, J., 2000. African hotspot volcanism: small scale convection in visualize. model, and present geochemical data in igneous petrology. the uppermantlebeneathcratons. Science 290, 1137e1140. Geochemistry, Geophysics, Geosystems 6, Q07011. Lasserre, M., 1966. Confirmation de l’existence d’une séries de granites tertiaires au Piper, J.D.A., Richardson, A., 1972. The paleomagnetism of the Gulf of Guinea vol- Cameroun. Bull BRGM 3, 141e148. canic province, West Africa. Geophysical Journal of the Royal Astronomical Lasserre, M., 1978. Mise au point sue les granitoïdes dits “ultime” du Cameroun: Society 29, 147e171. gisement, petrographie et geochronologie. Bull BRGM (2) IV, 143e159. Poudjom Djomani, Y.H., Diament, M., Albouy, Y., 1992. Mechanical behavior of the Le Bas, M.J., LeMaitre, R.W., Streckeisen, A., Zanettin, B., 1986. A chemical classifi- lithosphere beneath the Adamawa uplift (Cameroon, West Africa) based on cation of volcanic rocks based on the total alkali silica diagram. Journal of gravity data. Journal of African Earth Sciences 15, 81e90. Petrology 27, 745e750. Poudjom Djomani, Y.H., Diament, M., Wilson, M., 1997. Lithospheric structure across Le Maitre, R.W. (Ed.), 2002. Igneous Rocks. A Classification and Glossary of Terms. the Adamawa Plateau (Cameroon) from gravity studies. Tectonophysics 273, Recommendations of the International Union of Geological Sciences Subcom- 317e327. mission on the Systematics of Igneous Rocks. Rankenburg, K., Lassiter, J.C., Brey, G., 2005. The role of continental crust and lith- Le Marechal, A., 1976. Geologie et Géochimie des sources thermo minérales du ospheric mantle in the genesis of Cameroon Volcanic Line lavas: constraints Cameroun. No. 59. Collection des Memoires de l’ORSTOM, Paris. from isotopic variations in lavas and megacrysts from Biu and Jos Plateaux. Lee, D.C., Halliday, A.N., Fitton, J.F., Poli, G., 1994. Isotopic variation with distance and Journal of Petrology 46 (1), 169e190. time in the volcanic Islands of the Cameroon line: evidence for a mantle plume Reusch, A.M., Nyblade, A.A., Wiens, D.A., Shore, P.J., Ateba, B., Tabod, C.T., origin. Earth and Planetary Science Letters 123, 119e138. Nnange, J.M., 2010. Upper mantle structure beneath Cameroon from body wave Lee, D.C., Halliday, A.N., Davies, G.R., Essene, E.J., Fitton, J.G., Temdjim, R., 1996. Melt tomography and the origin of the Cameroon Volcanic Line. Geochemistry, enrichment of shallow depleted mantle: a detailed petrological, trace element Geophysics, Geosystems 11, Q10W07. http://dx.doi.org/10.1029/2010GC003200. and isotopic study of mantle-derived xenoliths and megacrysts from the Rudnick, R.L., and Gao, S., 2003. Composition of the continental crust, pp. 1e64. In Cameroon Line. Journal of Petrology 37, 415e441. the crust (ed. R.L.Rudnick) vol. 3. Treatise on Geochemistry (eds. H.D. Holland Marzoli, A., Renne, P.R., Peccirillo, E.M., Castorina, F., Bellieni, G., Melfi, A.G., and K.K. Turekian), Elsevier-Pergamon, Oxford. Nyobe, J.B., N’ni, J., 1999. Silicic magmas from the continental Cameroon Vol- Sato, H., Aramaki, S., Kusakabe, M., Hirabayashi, J., Sano, Y., Nojiri, V., Tchoua, F.M., canic Line (Oku, Bambouto and Ngaoundere): 40Ar-39Ar dates, petrology, Sr- 1990. Geochemical difference of basalts between polygenetic and monogenetic Nd-O isotopes and their petrogenetic significance. Contributions to Miner- volcanoes in the central part of the Cameroon volcanic line. Geochemical alogy and Petrology 135, 133e150. Journal 24, 357e370. Marzoli, A., Peccirillo, E.M., Renne, P.R., Bellieni, G., Iacumin, M., Nyobe, J.B., Sibuet, J.C., Mascle, J., 1978. Plate kinematic implications Atlantic Equatorial fracture Tongwa, A.T., 2000. The Cameroon Volcanic Line revisited: petrogenesis of zone trends. Journal of Geophysical Research 83, 3401e3421. continental basaltic magmas from lithospheric and asthenospheric mantle Stracke, A., Hofmann, A.W., Hart, S.R., 2005. FOZO, HIMU, and the rest of the mantle sources. Journal of Petrology 41, 87e109. zoo. Geochemistry, Geophysics, Geosystems 6, Q05007. http://dx.doi.org/ Mbassa, B.J., Njonfang, E., Benoit, M., Kamgang, P., Grégoire, M., Duchene, S., 10.1029/2004GC000824. Brunet, P., Ateba, B., Tchoua, F.M., 2012. Mineralogy, geochemistry and petro- Suh, C.E., Sparks, R.S.J., Fitton, J.G., Ayonghe, S.N., 2003. The 1999 and 2000 erup- genesis of the recent magmatic formations from Mbengwi, a continental sector tions of Mount Cameroon: eruption behavior and petrochemistry of lava. of the Cameroon Volcanic Line (CVL), Central Africa. Mineralogy and Petrology Bulletin of Volcanology 65, 267e281. 106, 217e242. Suh, C.E., Luhr, J.F., Njome, M.S., 2008. Olivine-hosted glass inclusions from Scoriae McBirney, A.R., Noyes, R.M., 1979. Crystallisation and layering of the Skaergard erupted in 1954-2000 at Mount Cameroon volcano, West Africa. Journal of intrusion. Journal of Petrology 20, 487e554. Volcanology and Geothermal Research 169, 1e33. McDonough, W.F., Sun, S., 1995. The composition of the Earth. Chemical Geology Sun, S., Hanson, G.N., 1975. Origin of Ross Island basanitoids and limitations upon 120, 223e253. the heterogeneity of mantle sources for alkali basalts and nephelinites. Con- Meyers, J., Rosendalh, B.R., Harrison, C.G., Ding, Z.D., 1998. Deep-imaging seismic tributions to Mineralogy and Petrology 52, 77e106. and gravity results from the offshore Cameroon Volcanic Line, and speculation Sun, S., McDonough, W., 1989. Chemical and isotopic systematics of oceanic basalts: of African hotlines. Tectonophysics 284, 31e63. implications for mantle composition and processes. Geological Society of Lon- Milelli, L., Fourel, L., Jaupart, C., 2012. A lithospheric instability origin for the don Special Publications 42, 313e345. Cameroon Volcanic Line. Earth and Planetary Science Letters 335e336, 80e87. Tchouankoue, J.P., Wambo, N.A.S., Kagou, D.A., Wörner, G., 2012. Petrology, Moreau, C., Regnoult, J.M., Deruelle, B., Robineau, B., 1987. A new tectonic model for geochemistry, and Geodynamic implications of basaltic Dyke Swarms from the the Cameroon Line, Central Africa. Tectonophysics 139, 317e334. southern Continental part of the Cameroon volcanic line, Central Africa. The Morgan, W.J., 1983. Hotspot tracks and the early rifting of the Atlantic. Tectono- Open Geology Journal 6, 72e84. physics 94, 123e139. Tsafack, T.J.P., Wandji, P., Bardintzeff, J.M., Bellon, H., Guillou, H., 2009. The Mount Ngako, V., Njonfang, E., Aka, F.T., Affaton, P., Nnange, J.M., 2006. The North-South Cameroon (Cameroon volcanic line, Central Africa): petrology, Paleozoic to Quaternary trend of alkaline magmatism from Niger-Nigeria to geochemistry, isotope and age data. Geochemistry, Mineralogy and Petrology. Cameroon: complex interaction between hotspots and Precambrian faults. SOFIA 47, 65e78. Journal of African Earth Sciences 45 (3), 241e256. Weaver, B.L., 1991. The origin of ocean island basalt end-member compositions: Ngounouno, I., Déruelle, D., Demaiffe, D., 2000. Petrology of the bimodal Cenozoïc trace element and isotopic constraints. Earth and Planetary Science Letters 104, volcanism of the Kapsiki Plateau (northernmost Cameroon, Central Africa). 381e397. Journal of Volcanology and Geothermal Research 102, 21e44. Wedepohl, K.H., Gohn, E., Hartmann, G., 1994. Cenozoic alkali basaltic magmas of Njilah, I.K., Ajonina, H.N., Kamgang, K.V., Tchinjang, M., 2004. K-Ar ages, mineralogy, western Germany and their products of differentiation. Contributions to major and trace element geochemistry of the Tertiary-Quatenary lavas from the Mineralogy and Petrology 115, 253e278. Ndu Volcanic ridge NW Cameroon. African Journal of Science and Technology 5, Wilson, M., 1992. Magmatism and continental rifting during the opening of the 47e56. South Atlantic Ocean: a consequence of Lower Cretaceous super-plume activ- Njilah, K., Temdjim, R., Nzolang, C., Tchuitchou, R., Ajonina, H., 2007. Geochemistry ity? in: magmatism and the causes of continental Break-up. Geological Society of Tertiary-Quartenary lavas of Mt. Oku, Northwest Cameroon, 040. Revisita Special Publications 68, 241e255. facultad de ingeenieria Univeridad de Antioquia, Junio, pp. 59e75. Wilson, M., Downes, H., 1991. Tertiary-Quaternary extension-related alkaline Njonfang, E., Nono, A., Kamgang, P., Ngako, V., Tchoua, F.M., 2011. Cameroon line magmatism in Western and Central Europe. Journal of Petrology 32, 811e849. magmatism (Central Africa): a reappraisal. In: Beccaluva, L., Bianchini, G., Yokoyama, T., Aka, F.T., Kusakabe, M., Nakamura, E., 2007. Plume-lithosphere Wilson, M. (Eds.), Volcanism and Evolution of the African Lithosphere. GSA, interaction beneath Mt. Cameroon volcano, West Africa: constraints from Special Paper, 478, pp. 173e191. 238U- 230Th-226Ra and Sr Nd-Pb isotopic systematics. Geochimica et Cos- Nkouandou, O.F., Temdjim, R., 2011. Petrology of spinel lherzolite xenoliths and host mochimica Acta 71, 1835e1854. basaltic lava from Ngao Voglar volcano, Adamawa Massif (Cameroon Volcanic Zindler, A., Hart, S.R., 1986. Chemical geodynamics. Annual Review of Earth and Line, West Africa): equilibrium conditions and mantle characteristics. Journal of Planetary Sciences 14, 493e571. Geoscience 56, 375e387. Zou, H., Zindler, A., Xu, X., Qi, Q., 2000. Major, trace element, and Nd, Sr and Pb Nkouathio, D.G., Kagou Dongmo, A., Bardintzeff, J.M., Wandji, P., Bellon, H., isotope studies of Cenozoic basalts in SE China: mantle sources, regional vari- Pouclet, A., 2008. Evolution of volcanism in graben and horst structures along ations, and tectonic significance. Chemical Geology 171, 33e47.