Magnetic Fabrics of the Miocene Ignimbrites from West-Cameroon: Implications for Pyroclastic ﬂow Source and Sedimentation

Journal of Volcanology and Geothermal Research 203 (2011) 113–132

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

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

Magnetic fabrics of the Miocene ignimbrites from West-Cameroon: Implications for pyroclastic ﬂow source and sedimentation

M. Gountié Dedzo a,⁎, A. Nédélec b, A. Nono a, T. Njanko a, E. Font c, P. Kamgang d, E. Njonfang e, P. Launeau f a LGE, Department of Earth Sciences, University of Dschang, B.P. 67 Dschang, Cameroon b UMR5563-LMTG-OMP, University of Toulouse-CNRS-IRD, 14 Av. Edouard-Belin, 31400 Toulouse, France c IDL-FUL, Instituto Dom Luis. Universidade de Lisboa, 1749-016 Lisboa, Portugal d Department of Earth Sciences, University of Yaoundé I, B.P. 812 Yaoundé, Cameroon e Laboratoire de Géologie, Ecole Normale Supérieure, Université de Yaoundé I, B.P. 47, Yaoundé, Cameroon f UMR-CNRS 6112, Laboratoire de Planétologie et Géodynamique, Université de Nantes, rue de la Houssinière, 44322 Nantes, France article info abstract

Article history: The Miocene ignimbrites of Mounts Bambouto and Bamenda located in the central part of Cameroon Volcanic Received 24 September 2010 Line are generally made of welded and non-welded massive lapilli tuff and lithic breccias. These discontinuous Accepted 18 April 2011 deposits cover a total area of 180 km2 with thickness ranging from 25 to 200 m. The different facies contain Available online 5 May 2011 several lithic fragments of mainly trachytic nature. The devitrified matrix of the welded ignimbrites is constituted by sanidine, anorthoclase, quartz, plagioclase, clinopyroxene, biotite, Fe–Ti oxides and devitrified Keywords: fi AMS ammes. Anisotropy of magnetic susceptibility (AMS) is used to characterize magnetic fabrics and to provide fl fl magnetic mineralogy an estimate of ow direction of each ignimbrite sheet. Magnetic mineralogy results from different ow units Miocene ignimbrite show that titanomagnetite, titanohematite, maghemite and goethite with grain size ranging from coarse MD caldera to very fine SP are the main magnetic carriers of these ignimbrites. Inferred transport directions based on the Western-Cameroon AMS data and field indicators show that Bambouto caldera is the source of main pyroclastic deposits of Mount Bambouto. In southwestern Mount Bamenda, Santa-Mbu caldera or Bambouto caldera constituted the probable emission center of Mbengwi, Bamenda and Mbu ignimbrite sheets, whereas magnetic fabrics of Bambili, Sabga and Big Babanki ignimbrites demonstrate that these deposits were emitted from a northeastern source, namely Oku vent in Mount Oku. A small number of subvertical AMS fabrics correspond to rocks possibly modified by an elutriation process. © 2011 Elsevier B.V. All rights reserved.

1. Introduction (Moundi et al., 2007) and Mount Bangou between 44.7 and 43.1 Ma (Fosso et al., 2005), and is still active at Mount Cameroon (1999 and The Cameroon Volcanic Line (CVL), a 1600 km long mega-shear 2000 eruption). zone in central Africa, shows a characteristic alignment of volcanoes, Most of the CVL volcanics are made up of mafic and felsic lavas. anorogenic complexes and grabens. This chain of Tertiary to recent Ignimbritic flow deposits are scarce and mainly observed in the generally alkaline volcanoes extends over more than 900 km across central part of the CVL, especially in Mounts Bambouto and Bamenda Cameroon from the Bui and Adamawa plateaux in the north to Mt (Fig. 1c). Other small flow deposits are present in Nkogam massif and Cameroon and Equatorial Guinea southwestwards (in the gulf of in Mount Oku. Pyroclastic flow deposits, ranging from non-welded to Guinea, Fig. 1a, b). These volcanoes are separated by plains, low areas welded ignimbrites, form an important component of rock formations corresponding to collapsed grabens, namely: Tombel, Mbo and Noun in the studied massifs of Mounts Bambouto and Bamenda. (Fitton, 1987; Deruelle et al., 1991, 2007; Nkouathio et al., 2002, 2008; Some of our understanding of pyroclastic flow behavior has been Itiga et al., 2004; Fosso et al., 2005). It continues seawards for a further gained by observations of recent pyroclastic eruptions; e.g. (i) 1998 and 700 km through the Atlantic Islands of Bioko, Principe, São Tomé and 2003 Soufriere Hills volcano eruptions in the Carrribean (Hart et al., Pagalu. 2004; Edmonds et al., 2006), (ii) 2006 and 2007 eruptions of Mt Etna in The volcanism along the CVL begun during the Eocene with the Italy (Behncke et al., 2008; Behncke, 2009) (iii) 2006 pyroclastic flows emplacement of the Bamoun plateau between 51.8 and 46.7 Ma on Mt Merapi in Java, Indonesia (Charbonnier and Gertisser, 2008). On the other hand, most of our understanding of the mechanisms and flow ⁎ Corresponding author. Tel.: +237 75 08 85 86. dynamics of pyroclastic density currents comes from examination of the E-mail address: [email protected] (M. Gountié Dedzo). deposits they left behind. Detailed examination of the facies architecture

0377-0273/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jvolgeores.2011.04.012 114 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132

Fig. 1. a) Location map of the Cameroon Volcanic Line (CVL); the main geologic features of Africa are indicated. b) The main volcanic centers of the Cameroon Volcanic Line. Central Cameroon Shear Zone according to Ngako et al. (2006); fracture zones following Lee et al. (1994) and Ballentine et al. (1997). c) Digital elevation model (DEM) of Mounts Bambouto and Bamenda and the close surroundings; the different calderas are also indicated.

and particle fabric of pyroclastic flow deposits gives insight into their Anisotropy of magnetic susceptibility (AMS) measurements can be transport and deposition. Determination of flow directions in ignim- used for determining lineations and foliations in ignimbrites. This brites is often uneasy. In some cases, flow directions have been method detects the alignment of magnetic minerals in the ignimbrite, successfully determined from textural indicators. These include the and has been used in many studies to evaluate flow directions in use of imbricated logs (Froggatt et al., 1981), as well as orientation of pyroclastic deposits in order to determine vent location or, more glass shards, crystals, pumice, and lithic fragments (Elston and Smith, recently, to provide information about transport and depositional 1970; Suzuki and Ui, 1982, 1983, 1988; Potter and Oberthal, 1987; Ui et process involved at different distances from the vent (Ellwood, 1982; al., 1989; Buesch, 1992; Seaman and Williams, 1992). Imbrication of Incoronato et al., 1983; Knight et al., 1986; Wolff et al., 1989; pumice and other features indicate not only the flow direction but also McDonald and Palmer, 1990; Hillhouse and Wells, 1991; Palmer, et al., the sense of motion (Kamata and Mimura, 1983; Suzuki and Ui, 1988). 1991; Fisher et al., 1993; Ort, 1993; Bear et al., 1997; Cagnoli and These methods, however, are time consuming and the features are Tarling, 1997; Le Pennec et al., 1998; MacDonald et al., 1998; Ort et al., commonly absent in specific outcrops. 1999, 2003; Palmer and MacDonald, 1999; Le Pennec, 2000; Wang M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 115 et al., 2001; Porreca et al., 2003; Alva-Valdivia et al., 2005; Giordano of outcrops due to their general re-covering by more recent basaltic et al., 2008; Petronis and Geissman, 2008). flows or due to their rapid erosion, makes difficult the determination of the corresponding emission centers. Therefore, systematic AMS deter- 1.1. Aim of study mination was the selected method, here used for the first time in the CVL. AMS results, combined to magnetic mineralogy and image analysis Determination of flow directions in ignimbrites of Bambouto and studies, help to connect discontinuous Bambouto and Bamenda flow Bamenda massifs is difficult, because the fabric in these deposits is deposits to their sources and to discuss magnetic fabric acquisition and visually nearly isotropic in most outcrops. In addition, the discontinuity emplacement mode of these ignimbrites.

Fig. 2. Geologic sketch maps of a) Mounts Bambouto and b) Bamenda showing the sampled stations and locations of dated samples. 116 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132

2. Geological setting et al., 2001; Nzolang et al., 2003). Mt Bamenda culminates at 2621 m (Bambili Lake borders) and is characterized by two elliptic calderas After Mounts Cameroon and Manengouba, Mount Bambouto (Fig. 1c): Santa-Mbu caldera (6km×4km) and Lefo caldera constitutes the third most important volcano (in volume) of the CVL (4 km×3 km) (Gountié Dedzo et al., 2009), whose basements are with Mount Mélétan (2740 m) as its highest point (Figs. 1c, 2a). It is essentially made of trachytic domes that are also abundant on the made up of volcanic products dated from 21.12 Ma to 0.5 Ma, and external versant of the massif. According to Kamgang et al. (2007, 2008), comprising basalts, trachytes, phonolites, rhyolite and ignimbritic felsic and intermediate lavas (27.40–18.98 Ma) are made of mugearites, deposits constituted by various facies (Nono et al., 2003, 2004). benmoreites, trachytes, rhyolites. Mafic lavas (basanites, basalts, and Geochemical, mineralogical and crystallographic data (Marzoli et al., hawaiites) are dated from 17.4 Ma to the present; rhyolitic ignimbrite 2000; Saviulo et al., 2000; Njonfang and Nono, 2003)showgenetic flow deposits are inserted between the granito-gneissic basement at relationships between different petrographic types. The Mount Bam- their floor and the lateritized old basalts on top. bouto caldera (Fig. 1c) is a dissymmetric roughly elliptic depression (13×8 km), which opens to the west of the volcano. On the southeastern side, it shows subvertical walls that rise 1300 m above 3. Field observations and petrography of ignimbrites its floor made of trachytic and phonolitic domes and dome-flows. A synthetic revision of the volcanic story of Mount Bambouto is 3.1. Ignimbrites of Mount Bambouto proposed as follows by Kagou Dongmo et al. (2010). The first stage, ca. 21 Ma, corresponds to the building of the initial basaltic shield Ignimbrites outcrops in Mount Bambouto are discontinuous and volcano. The second stage, from 18.5 to 15.3 Ma, is marked by the cover approximately 17% (≈135 km2) of the massif (Fig. 2a, 3) with 30 collapse of the caldera linked to the pouring out of ignimbritic to 120 m in thickness. The volume of these pyroclastic deposits rhyolites and trachytes. The third stage, from 15 to 4.5 Ma, renews estimated at 13.5 km3, is actually much larger because these formations with basaltic effusive activity, together with post-caldera extrusions are covered by generally lateritized basalts in the southern part of the of trachytes and phonolites. The 0.5 Ma Totap basaltic effusive activity massif. In the lower zone, they lay on a metamorphic basement, while in could indicate the beginning of a fourth phase. the upper zone, they cover trachytic lavas. Welded and non welded Mount Bamenda, the fourth more important volcano of the CVL massif lapilli tuff (Tlm) and massif lithic breccias (Brlm) (following the appears as the NE prolongation of Mounts Bambouto with which there is classification of Branney and Kokelaar, 2002) are the two ignimbritic no clear limit. The massifs lay on a granitic Pan-African basement (Toteu facies cropping out in this massif.

Fig. 3. Stratigraphic sections of the Mount Bambouto ignimbrites in some localities. The ages are from Gouhier et al. (1974), Tchoua (1973), Youmen et al. (2005), Nkouathio et al., 2008. M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 117

The southernmost ignimbrites of Dschang (9 km2;18.1Ma,Youmen flow was moved roughly toward azimuths 200–215° at Dschang and et al., 2005) outcrop in sheets in the Menoua valley and are overlain in 349–356° at Lepo. In the whole ignimbritic sheets, the matrix is places by aphyric basaltic flow (Fig. 3a; 12.52 Ma; Nkouathio et al., devitrified and bears numerous lithic fragments (0.3×0.4 cm to 2008) intersected to the west by a dyke of phonolite (6.61 Ma; 1.2×1.6 cm, 10–35%), most of them displaying a trachytic texture. Nkouathio et al., 2008). The ignimbrites of Mbeng, Mbou, Lepo, Nzemla Other components include: rhyolites, granitic basement, obsidian, I and Nzemla II are the main outcrops (82 km2) of ignimbritic flow scoria (essentially at Mbou) and fragments of carbonized wood. The deposits east of Fongo-Tongo (Figs. 2a, 3b–d) and are covered in places devitrified matrix (with non devitrified glass shards at Mbou; Fig. 5a) with basalts (14.08 Ma; Nkouathio et al., 2008). The ignimbrites of is made of alkali feldspar (sanidine and anorthoclase; 10–35%), quartz Baranka (44 km2) outcrop at the summit area of the volcano and cover (3–5%), plagioclase (1–3%), oxides (1–3%), biotite (1–3%) and discontinuously the south-western rim and the bottom of the Bambouto clinopyroxene (1%) (Fig. 5b). Subvertical elutriation pipes (up to caldera (Fig. 2a). Their thickness is difficult to estimate because of steep 1.2 cm in diameter) are common in the middle part and at the top of relief and dense vegetation. Several phases of ignimbrites deposit are deposits of the WU (Nzemla II; Fig. 5c) and of the DGU (Mbeng) present with intercalations of trachytic and basaltic flows (Fig. 3e). indicating the release of fluid phase after deposition. In the different stratigraphic sections (Fig. 3), the welded parts of these deposits consists of one simple cooling unit made of one or two 3.2. Ignimbrites of Mount Bamenda flow units represented by dark gray unit (DGU), light gray unit (LGU) and whitish unit (WU). The lower parts of welded ignimbrites has a The ignimbrites of Mount Bamenda (Figs. 2b, 4) cover about 10% eutaxitic texture characterized by deformed and devitrified fiammes (60 km2) of the massif with thickness ranging between 50 and 200 m. (5–20%) with preferentially oriented direction which suggest that Their volume which is really most important is estimated to be

Fig. 4. Stratigraphic sections of the Mount Bamenda ignimbrites in some localities. 118 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 119

Fig. 6. Representative thermomagnetic curves (susceptibility versus low and high temperature) of Mounts Bambouto and Bamenda ignimbrites. The arrows indicate the heating and cooling curves.

6.3 km3. As in the case of Mount Bambouto, the deposits are (3 km2) Sabga (6 km2) and Big Babanki (2 km2) lie on the basement discontinuous and lie on the basement rocks made of granite, syenite, rocks made of granite, micaschist and migmatite (Fig. 4d, e, f). The WU micaschist and migmatite. of welded Tlm are cover with a Brlm facies (not sampled) in the two The ignimbrite sheets of Bamenda (6 km2), Mbu (3 km2) and first localities. Mbengwi (13 km2) lie on granitic (Fig. 4a, b) and syenitic (Fig. 4c) The basal and middle parts of Bambili DGU are much consolidated basements in the SW of Mt Bamenda. The upper parts of the deposits and display eutaxitic texture (Fig. 4d); they are characterized by black in the two first localities and the entire unit of Mbengwi are made of and/or white preferentially oriented devitrified fiammes (10–30%) Brlm facies (only sampled in Mbengwi). A non-welded Tlm was also showing that transport direction of pyroclastic current was roughly recognized at Mbu, but only the lowermost welded DGU (Tlm) was directed toward azimuth 235 (site GM24: Fig. 5d). At Bamenda (site GM sampled in this sheet. In the NE of massif, the ignimbrites of Bambili 28), field indicators represented by elongate imbricated pumice and

Fig. 5. Photomicrographs and photographs of Mounts Bambouto and Bamenda ignimbrites: a) Ignimbrites (WU) of Mbou showing ashy matrix and multiple glass shards. b) Enclave of trachyte and granite in the DGU of Nzemla I ignimbrites. c) Presence of multiple elutriation pipes at the top of WU of Nzemla II ignimbrite; the reddish color is due to rubefaction. d) Oriented thin section of Bambili ignimbrite (DGU) showing aligned fiammes which defined flow direction along azimuth 235°. e) Devitrified matrix of Nzemla II (WU) ignimbrites with several Fe–Ti oxides. f) Crystals of titanohematite with wide lens-shape exsolution (DGU, GM7) and thin exsolution lamellae of ilmenite (DGU, GM31). g) Association of goethite and titanohematite in the DGU of Nzemla I ignimbrite. h) Crystals of pyrite in Baranka ignimbrites. San: sanidine; Anor: anorthose; Pl: plagioclase; Qtz: quartz; Pyr: pyrite. 120 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 lithic fragments suggest that flow was directed toward azimuth 226°. identifying the mineral phases that contribute to the magnetic The incipient zones of the welded WU of Bambili is characterized by properties of the ignimbrites. vertical elutriation pipes (few millimeters to 1.5 cm in diameter) and Low field versus high temperature susceptibility measurements multiple pumice (40%). Rock enclaves (≤1.9×2.2 cm, 10–25%) are (K-T curves) in air were carried out using a CS-3 apparatus coupled to mainly trachytic with minor proportion of rhyolite, granite, ignimbrite the KLY-2 bridge susceptometer. Selected samples were progressively and obsidian. In the breccificated zones, proportions of lithics (up to heated, generally up to 700 °C at heating rate 18 °C/min and then 10×15 cm at Bambili) range between 30 and 70%. The devitrified matrix subsequently cooled to room temperature at the same rate. In some of Tlm in different units is made of alkali feldspar (sanidine and cases, low temperature (from −200 °C to room temperature) anortoclase; 5–10%), quartz (2–3%), plagioclase (b3%), oxides (1–3%), susceptibility was also recorded. biotite (1–2%) and of clinopyroxene (1%). The characterization of ferromagnetic minerals and their respective concentration was obtained by acquisition of isothermal remanent 4. Analytical methods magnetization (IRM) and anhysteretic remanent magnetization (ARM) on 10 representative samples. IRM curves were acquired in progres- 4.1. Magnetic mineral characterization sively increasing magnetizing field using a pulse magnetometer MMPM10 at fields up to 2T. For ARM measurements, samples were To ascertain the interpretation of the AMS results, standard rock first demagnetized at 100 mT and then subjected to an alternating field magnetic experiments were conducted with the principal goal of (AF) peak of 100 mT plus a DC field of 0.1 mT in an LDA-3A apparatus. In

Fig. 7. a) IRM acquisition curves for Mounts Bambouto and Bamenda ignimbrites: (i) presence of one component; (ii) presence of two components. b) Examples of treatment of IRM curves by cumulative log-Gaussian (Robertson and France, 1994) functions using the software developed by Kruiver et al. (2001). M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 121 all cases, magnetic remanence was measured with a JR5-A the susceptibility ellipsoid is characterized by the T parameter with T= magnetometer. (2lnK2 −lnK1 −lnK3)/(lnK1 −lnK3)(Jelinek, 1981) ranging from +1 In order to characterize the mineral assemblages and textures of for oblate ellipsoid to −1 for prolate ellipsoid. Triaxial ellipsoids are ignimbrites, optical studies using transmitted and reflected-light characterized by values between −0.5 and +0.5. microscope were undertaken on polished thin sections from representative samples. Nature of ferromagnetic minerals was investigated 4.3. Shape fabric determination by image analysis by electron microprobe analyses on CAMECA SX50 apparatus operating at the usual conditions. The magnetic fabric is supposed to be coaxial with the mineral fabric of the magnetic carriers. This point is now well established in plutonic rocks (e.g. Grégoire et al., 1998) and even in basaltic flows (Canon-Tapia 4.2. Sampling methods and AMS measurements et al., 1995). However, the peculiar emplacement modes of ignimbrite and some unexpected AMS orientations required that magnetic fabric The magnetic studies were performed on different flow units was compared with the fabrics of other elements (feldspars, quartz and (Figs. 2–4). All measurements were made in LMTG (Laboratoire des fiammes). This determination was carried out on oriented thin sections Mécanismes et Transferts en Géologie) at Paul Sabatier University in which were cut from three samples (GM7, GM29 and GM40) in Toulouse, France. horizontal, NS and EW sections to build a shape preferred orientation Sampling of the Bambouto (244 core samples) and Bamenda (115 (SPO) ellipsoid. The grain size and shape fabric data were extracted out core samples) ignimbrites was performed on 41 stations (Fig. 2). of the digitized images by the SPO program of Launeau (2004) using the Stations were distributed somewhat homogeneously over each ignim- intercepts method by the Intercepts 2003 software of Launeau and brite sheet. At each site, a total of 6–10 oriented cores were collected Robin. All minerals and fiammes are digitized in gray levels. Intercepts using a portable, gasoline-powered drill-machine with a non-magnetic are extracted directly from the gray level image following the Launeau et diamond-tipped drill bit in an area covering approximately 5 to 10 m2. al. (2010) algorithm from each boundary between white, gray and black Samples were oriented using magnetic compass. In order to avoid pixels. SPO ellipsoids are given by a combination of 3 mutually breakage, the cores were collected in the welded part of the deposit. In orthogonal sections (one horizontal and two vertical, namely NS and laboratory, each core sample was cut into 2.2×2.5 cm cylinder EW) following the procedure of Launeau and Robin (2005) and Robin specimens, using a diamond tipped, non-magnetic saw blade. Up to (2002). These sections do not need to be chosen in relation with the four specimens per sample were obtained, hence a total of 957 AMS principal axes. specimens. AMS measurements were performed on a Kappabridge susceptometer (KLY-3S, Agico, Czech Republic) operating at a low alternating field (4×10−4 T at 920 Hz) with a sensitivity of about 5. Results 2×10−7 SI, allowing anisotropy discrimination below 0.2% over a wide range of susceptibility. 5.1. Magnetic mineralogy They measure the orientation of the magnetic carriers in a rock. AMS measurement of one rock specimen results in an ellipsoid of magnetic Petrographic determination and microprobe analysis were per- susceptibility (K) defined by the length and orientation of its three formed on opaque minerals with appropriate grain sizes. In order to identify the magnetic carriers bearing the AMS signal we performed orthogonal axes, K1 ≥K2 ≥K3, which are the three eigenvectors of the susceptibility tensor, representing the maximum, intermediate and thermomagnetic analyses in low and high temperatures and acquisi- minimum susceptibility directions respectively (Tarling and Hrouda, tion of Isothermal Remanent Magnetization (IRM) and ARM/IRM ratios. 1993). The long axis K1 defines the magnetic lineation, meanwhile the short axis K3, is the foliation pole, i.e. the normal to the plane of magnetic foliation. The mean magnetic susceptibility (Km) is the arithmetic mean 5.1.1. Opaque mineral textures and compositions of the lengths of the principal axes (Km=[K1 +K2 +K3]/3). The Petrographic analyses indicate that opaque minerals are in technique also quantifies the anisotropy percentage (P%=[(K1/K3)− proportions ranging from 1% to 3% of the bulk-rock composition. 1]×100). In addition, the linear (L%=[(K1/K2)−1]×100) and planar Transmitted and reflected-light observations enable the identification (F=[(K2/K3)−1]×100) anisotropies are also provided. The shape of of different types of opaque minerals: (i) fine- to medium-grained

Table 1

Results of the LAP–GAP–SAP treatment (Kruiver et al., 2001). SIRM (A/m) represents the value of IRM at saturation, B1/2 (mT) is the mean coercivity of the magnetic phase and DP is the dispersion parameter (DP). Values of the ARM/SIRM ratio at 100 mT are also indicated.

Samples LAP–GAP–SAP ARM/SIRM ratio

Component Contribution SIRM B1/2 DP ARM SIRM ARM/SIRM (%) (A/m) (mT) at 100 mT (A/m) at 100 mT (A/m) at 100 mT (%)

GM2 1 100 28.5 201.7 0.31 0.23 4.5 5.1 GM6 1 100 3.7 85.1 0.40 0.17 2.04 8.3 GM7 1 35 0.95 18.2 0.33 0.13 1.50 8.6 2 65 1.77 149.1 0.33 GM14 1 28 0.64 23 0.41 0.12 0.89 13.5 2 72 1.65 187.3 0.32 GM15 1 100 24.5 125.9 0.33 0.44 8.71 5.1 GM20 1 100 6.6 39.8 0.31 0.58 5.97 9.7 GM21 1 52 0.53 13.2 0.4 0.10 0.61 16.4 2 48 0.49 220.6 0.33 GM22 1 100 140 177.8 0.32 2.08 27.95 7.5 GM23 1 100 12.5 95.5 0.42 0.98 6.21 15.8 GM24 1 100 0.55 63.1 0.42 0.55 5.97 9.2 122 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132

Table 2

Magnetic data of Mounts Bambouto and Bamenda. K1,K2 and K3 are the maximum, intermediate and minimum susceptibility intensities respectively; Km is the mean magnetic susceptibility; L% is the linear anisotropy percentage; F% is the planar anisotropy percentage; P% is the total anisotropy percentage; T is the Jelinek's shape parameter (Jelinek, 1981);

D: declination in degrees; I: inclination in degrees; N is the number of samples for the site; E1–2,E1–2, and E1–2 are semi-angles of the 95% conﬁdence ellipses around the principal susceptibility axes (derived from Jelinek, 1978); samples with a star represent rejected samples due to high within site dispersion.

Locality Unit Site Lon Lat N AMS parameters Mean eigenvectors Semi-angles (°) (°) of conﬁdence K1 K2 K3

Km P% L% F% T D I D I D I E1–2 E2–3 E3–1 (μSI)

Mount Bambouto Dschang DGU GM02 10.03694 5.44306 24 1763.5 7.1 1 6 0.7 31 7 121 0 212 83 31.6 6.4 5.4 GM04 10.03694 5.44306 27 1712.9 8.2 0.8 7.3 0.8 224 1 134 6 325 84 24.5 3.1 2.8 GM05 10.03194 5.44889 18 2198.8 6.3 0.3 6 0.9 208 2 298 6 102 83 78.3 12.9 12.3 LGU GM01 10.02639 5.43778 23 196.6 2.5 0.1 2.3 0.9 70 10 161 2 264 80 64.5 7.1 6.7 GM03 10.04583 5.44806 23 195.7 0.3 0 0.3 0.8 18 10 287 1 190 80 79.7 34.3 31.3 Mbou WU GM11a* 10.14167 5.49417 30 208.2 0.2 0.1 0.1 0.1 206 30 302 11 49 57 73.7 68.3 55.4 GM11b 10.14167 5.49417 9 250.5 1.9 0.3 1.7 0.7 359 23 260 18 154 68 77.4 35.7 31.7 Mbeng DGU GM10 10.10778 5.48556 22 187.8 0.5 0.2 0.3 0.3 132 65 330 24 237 7 42.6 27.3 18.3 Lepo DGU GM13 10.04917 5.52333 22 571.7 3.7 0.4 3.2 0.8 71 2 341 12 168 78 49.6 9.5 8.4 GM14 10.06528 5.51667 23 512.9 0.2 0.1 0.1 -0.3 41 3 310 18 139 72 28.8 28.8 18.7 GM15 10.05389 5.51889 17 1303.4 0.3 0.1 0.2 0.6 26 4 295 8 143 81 65 25.6 21.4 GM17 10.07333 5.50417 23 419.1 2.1 0.3 1.8 0.7 43 2 313 3 163 87 36.9 8.3 7 GM38 10.10028 5.54083 32 801.5 1.9 0.1 1.7 0.8 60 5 330 8 182 80 73.5 15.8 14.7 LGU GM35 10.03167 5.51944 27 211.2 1.1 0.1 1 0.9 27 6 118 8 261 80 67.5 6.8 6.5 GM36 10.01083 5.53639 33 614.4 3.5 0.2 3.3 0.9 77 4 347 2 233 86 59.8 7.5 7 GM37 10.10028 5.54083 27 130.6 0.4 0.1 0.3 0.6 334 2 64 2 208 88 47.4 13.9 11.4 Nzemla I DGU GM06a 10.10500 5.53944 10 819.1 0.4 0 0.4 0.8 324 56 195 23 94 24 67.7 17.2 15.4 GM06b 10.10500 5.53944 10 856.9 0.4 0.1 0.3 0.5 285 51 25 8 121 38 24.1 7.9 6.1 GM06c 10.10500 5.53944 8 952.5 0.4 0.1 0.4 0.7 302 45 168 35 59 24 62.3 21.5 18.1 GM07 10.11056 5.53639 9 359.1 1 0.4 0.6 0.1 59 4 328 8 179 81 18.5 14.4 8.3 GM09 10.08306 5.52583 17 1610.5 0.2 0 0.1 0.6 35 75 265 10 174 11 73.1 39.6 33.5 GM18 10.07194 5.52667 21 3473.2 0.2 0 0.2 0.9 127 4 216 1 283 86 81.8 10 9.8 GM34 10.10028 5.54083 35 3374.7 0.2 0.1 0.1 0.1 19 10 286 17 139 70 42.1 36.4 22.1 GM41 10.11047 5.53689 16 227.3 0.2 0.1 0.1 0.0 107 81 304 8 214 3 38.8 37.6 21.5 Nzemla II WU GM08 10.11306 5.51778 25 153.7 0.3 0.2 0.1 −0.2 215 75 84 10 352 12 28.3 28.3 16.9 GM12 10.07000 5.50806 20 24.2 1.2 1 0.3 −1.0 309 75 54 4 145 15 50.2 50.2 43.6 GM16 10.07944 5.50944 24 211.1 0.7 0.5 0.2 −0.6 238 80 132 3 42 10 11.2 11.2 8.8 GM39 10.12778 5.52500 24 107.5 0.7 0.7 0.1 −0.7 230 86 4 3 94 3 9.6 9.6 8.9 GM40 10.13306 5.52389 20 99.8 0.5 0.3 0.2 −0.6 177 74 274 2 4 16 43.6 43.6 32.6 Baranka LGU GM19 10.08361 5.68528 17 354.7 2 0.4 1.6 0.7 335 26 72 15 188 60 61 21.7 18.1 GM20 10.03861 5.67889 25 703.7 1.1 0.7 0.4 −0.3 7 47 221 37 117 18 12 12 8 Mount Bamenda Bamenda LGU GM27 10.15417 5.96444 23 147.3 0.7 0.3 0.4 0.1 142 2 233 11 45 79 53.5 49.5 32.1 GM28 10.17389 5.95806 24 88.9 0.7 0 0.7 0.9 254 35 14 35 134 36 64 8.6 8 GM29 10.16500 5.94500 22 375.5 0.6 0.4 0.2 −0.3 278 81 43 6 134 8 25.3 25.3 17.6 Mbu DGU GM21 10.09472 5.85611 20 911.1 0.8 0.3 0.4 0.1 213 13 307 13 79 71 41.4 37 22.1 Mbengwi DGU GM31 10.06222 6.00056 33 2062.5 0.2 0 0.1 0.5 182 24 274 4 12 66 48 17.7 14 GM32 10.05528 5.97917 32 725.6 1.4 0.1 1.2 0.8 197 1 100 13 310 87 75 36.7 31.9 Bambili DGU GM24 10.27944 6.00417 24 3253.2 0.2 0 0.2 0.6 315 2 45 6 202 84 40 10.4 8.5 GM26 10.27139 6.01333 12 2579.5 0.2 0.1 0.1 0.0 50 26 144 8 249 62 32.9 23.3 14.5 WU GM25* 10.27278 6.00694 32 66.3 0.4 0.2 0.2 0.0 144 89 282 1 12 1 64.8 60.3 43.9 GM33 10.27139 6.00944 27 66.6 0.1 0.1 0.1 0.1 39 84 290 2 199 6 22 22 11.6 Sabga DGU GM22a 10.33944 6.01917 11 1220.6 1.7 0.7 1 0.2 315 32 73 37 197 37 17.1 11 6.8 GM22b 10.33944 6.01917 9 1143 1.6 0.5 1.1 0.4 128 20 241 47 23 36 23.3 11.6 7.9 LGU GM23a 10.33833 6.00778 7 391.9 0.5 0.3 0.1 −0.3 221 8 340 73 129 15 17.9 17.9 12.7 GM23b 10.33833 6.00778 9 379.8 0.5 0.2 0.3 0.0 225 6 116 73 317 16 24.3 23.7 12.5 Big-Babanki WU GM30 10.25944 6.12056 12 55.5 0.5 0.2 0.3 0.0 56 6 326 0 235 84 60.6 11.3 10.2

oxides included in feldspar and in pore boundaries; (ii) free coarse Finally, sulfide minerals, such as pyrite (Fig. 5h), were also identified opaque minerals in glassy matrix (Fig. 5c, e). Results of microprobe in a few facies of ignimbrites; they are generally non-magnetic. analysis show that opaque minerals are mainly Fe–Ti oxides. Among these oxides, hematite–ilmenite solid solution (titanohematite) is 5.1.2. Thermomagnetic curves dominant. Crystals show generally high temperature exsolution of Ti- Thermomagnetic curves were obtained from 16 samples in Mount rich titanohematite close to the ilmenite end-member. The grains are Bambouto and from 9 samples in Mount Bamenda for which typical either relatively homogeneous, with thin exsolution lamellae, or examples are illustrated in Fig. 6. Most of the thermomagnetic curves display wide lens-shape exsolution of ilmenite (Fig. 5f and g). Their exhibit irreversible thermomagnetic behavior with higher suscepti- width is in the average of 30–100 μm and their length is up to 200 μm. bility values for the cooling curves indicating thermally induced The ilmenite-rich areas are often intensely altered to pseudo-rutile or conversion of iron-bearing minerals to new (ferro)magnetic phases. other Ti-rich products. In some cases, large grains of (titano) Curie temperature (Tc) indicate that titanomagnetite with low magnetite, sometimes weathered to goethite, are also identified (Tc=100–300 °C: samples GM2 and GM13) to high (Tc=420– (Fig. 5g). Small grains of magnetite were not detected at this stage. 580 °C: samples GM8, 11, 14 and 20) Ti-content is the principal M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 123

5.2. AMS

The AMS results are presented in Table 2.

5.2.1. Scalar data In the whole set of AMS samples, Km values range from 24 μSI, i.e. 24×10− 6 SI to 3473 μSI in Mts Bambouto and Bamenda (Table 2). In first approximation, Km valuesb500 μSI are regarded as typical of paramagnetic rocks, i.e. rocks devoid of any magnetite, whereas higher Km values generally point to the existence of some magnetite, which has a much higher susceptibility than other minerals, i.e. ferromagnetic rocks sensu lato (Rochette, 1987). There is a rough correlation between Km values and rock color, hence rock mineralogy. Km values are lower than 500 μSI in all WU ignimbrite flows and most LGU (71% in Mt Bambouto and 100% in Mt Bamenda). The DGU generally display higher values, suggesting a dominant ferromagnetic behavior (76% of samples of Mt Bambouto, 100% in Mt Bamenda). The anisotropy percentage (P%) vary from 0.1 to 8.2% in the two massifs with all the stations displaying very weak anisotropy (P%b2%) in Mt Bamenda (Table 2). According to Giordano et al. (2008), these values of P% are typical values for pyroclastic flow deposits. It should be noted that there is no systematic relationship between the susceptibility magnitude and the anisotropy percentage (Fig. 8a). The samples that have the highest Km are characterized by very low values of P%. Nevertheless, the DGU of Dschang in Mt Bambouto display the highest anisotropies (P%N5%) as well as rather high susceptibility magnitudes; the LGU and WU always display low values (P%b3%) in both massifs. Fig. 8. Relationship between bulk susceptibility and (a) percentage of anisotropy (P%) The planar anisotropy percentages (F%) vary from 0.1 to 7.3% in Mts and (b) shape of the susceptibility ellipsoid (T).The plots are made with all core Bambouto and Bamenda (Table 1). The DGU of the two massifs show the samples. highest values. As expected, linear anisotropy percentages (L%) always present very weak values (b1%). The balance between linear and planar fabrics varies from one flow unit to another, as shown by the shape parameter (T) of Jelinek (1981). Although covering almost its whole magnetic carrier in most samples. In addition to titanomagnetite, range of variation, from −1to0.9(Table 1), this parameter points to titanohematite is also present in sample GM32 (Tc=~620 °C; Fig. 6k) dominant planar fabrics (positive values, mostly higher than 0.5). and is ubiquitous in sample GM31 (Tc=600–640 °C; Fig. 6l). Low Negative values, i.e. dominant linear fabrics, especially characterize the amounts of goethite (Tc=~100 °C) are also depicted in few samples Nzemla II WU of Mt Bambouto (Fig. 8b). (GM7, GM14, and GM32) corroborating with our petrographic analyses. A small decrease in the intensity of the magnetic 5.2.2. Directional data susceptibility is also observed between 300 and 360 °C in few samples Table 2 gives the orientations of K1,K2 and K3 with respect to the (GM20 and GM32) that may correspond to the destruction temper- north as declinations (azimuths) and inclinations (plunges). E1–2, ature of maghemite. E2–3,E3–1 are the semi-angles of the 95% confidence ellipses around the principal susceptibility axes; their values witness the fact that K3 5.1.3. IRM and ARM analyses (magnetic foliation pole) is generally more well-defined than K1 IRM analyses were performed in 10 samples from the Mont (magnetic lineation), as can be seen also in the lower hemisphere Bambouto and the Mont Bamenda and were treated by the cumulative projections for each analyzed sites presented in Fig. 9. Comparison of log-Gaussian method (Robertson and France, 1994) using the software K1 and K3 eigenvectors at specific sites and as an overall pattern of Kruiver et al. (2001) (Fig. 7; Table 1). Seven samples show that a low provides the basis for estimating overall transport directions for coercive phase is the principal magnetic carrier in these rocks while each ignimbrite deposits. Trends of magnetic fabrics vary between three samples exhibit a bimodal distribution of coercivity spectra each ignimbrite sheet, but are relatively well defined in some cases

(Fig. 7). Values of B1/2 (mT) are comprised between 13 and 220 mT and (Fig. 9). Based on the orientation of the principal axes K1 and K3 are in the typical range of titanomagnetite (e.g., Font et al., 2009). The (Wang et al., 2001), the AMS fabrics of the massifs can be classiﬁed heterogeneity observed in B1/2 (mT) values is interpreted here to result into: (i) normal fabrics in which K1 axis dips at angle of less than 30°, from different concentrations in Titanium as it was already noted whilst K3 axis is nearly vertical or steeply plunging (i.e. the magnetic in the unblocking temperatures obtained from thermomagnetic foliation is subhorizontal), and (ii) inverse fabric in which K1 dips at analyses. Bimodal distribution of coercivity spectra is thus inter- high angle (generally N50°), whilst K3 is subhorizontal (i.e. the preted to correspond to a mixture of low (Ti-rich) and high (Ti- magnetic foliation is subvertical). Normal fabrics may correspond to poor) titanomagnetites. two different situations after Giordano et al. (2008).Insomecases,

ARM/SIRM ratio is an indicator of the presence of ultrafine K1 is consistent with the flow direction inferred from the foliation superparamagnetic (SP) particles. In remagnetized carbonates, pres- imbrication, whereas in other cases it is normal to the flow direction ence of SP particles are depicted when ARM/SIRMN10% (Jackson et al., (Fig. 10). Such a transverse-to-flow position of K1 is rather common

1992, 1993). In our selected samples, ARM/IRM ratios vary from 5% to in ignimbrites. Therefore, K3 is preferred to K1 to derive ﬂow direction in 17% suggesting that SP particles contribute signiﬁcantly to the bulk ignimbrites, a situation that is different from the case of other magmatic properties (Table 1). rocks (Bouchez, 2000). More precisely, the sense of transport is based on 124 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 125

Fig. 9. Lower hemisphere projections of all AMS data encountered in Mounts Bambouto and Bamenda with inferred ﬂow direction using imbrication of magnetic foliation (normal fabrics); squares, triangles and circles represent K1,K2 and K3 respectively; ﬁlled symbols represent means of K1,K2 and K3; stars represent the pole of the paleotopography. Site locations are shown in Fig. 2.

the imbricated pattern of magnetic foliations, provided that the remaining two stations (GM4 and GM5), magnetic foliations dip very paleotopography is either recognized in the ﬁeld (sites GM1, GM31, gently (6°SE and 7°SSE) with K1 perpendicular to the imbrication of K3 GM35, and GM36) or assumed with respect to the source. In the case of axes. inverse fabrics, no ﬂow direction can be deduced from the magnetic AMS stations at Lepo show synthetic stereograms with a best pole data. Fig. 11 presents the AMS results at the formation level, with mean of magnetic foliation at 190/84 and a best line of magnetic lineation at site data plotted on the maps and all sample data plotted in the lineation 43/3 (Fig. 11a). All stations present normal fabrics and yield readily and foliation pole synthetic diagrams. interpretable AMS directional data. Three sites (GM36, GM37 and GM38) have magnetic foliation planes gently inclined (2–10°) to NE,

5.3. Mount Bambouto while site GM35 display K3 axe inclined (10°) approximately to E. The four remaining sites (GM13, GM14, GM15, and GM17) are imbricated In Mount Bambouto, 17 sites in the gray units and 1 site in whitish to the NW (3–28°), as can be seen in Fig. 11a. Magnetic line- unit yield interpretable AMS directional data used to infer transport ations generally present a NE–SW direction with low plunges (2 direction. The magnetic lineations have mostly very low plunges (2° to to 6°).

10°) and magnetic foliation planes dip very gently (2 to 12°). For the At Nzemla I, eight AMS sites display K1 and K3 axes somewhat majority of stations, K1 is approximately perpendicular to the flow dispersed with a general concentration respectively trending to the direction deduced from K3 axis. In the remaining sites (3 at Dschang and NW and SE (best pole of magnetic foliation at 154/77 and best line at 1 at Nzemla I), K1 is parallel to the transport directions. At Nzemla I, 23/9; Fig. 11a). In detail, 3 sites display normal fabrics with gently Mbeng and Baranka, for most sites, lineation tends to be subvertical (45 inclined magnetic foliation planes (best foliation poles at 283/86, to 81°) and K1–K2 planes have very high inclination (45 to 87°). 179/81 and 139/70 for GM18, GM7 and GM34 respectively) and In Dschang, all stations present normal fabrics. Magnetic lineations subhorizontal lineations. These 3 stations suggest a poorly defined (best line at 38/4; Fig. 11 a) have low plunges (1 to 10°) and generally transport direction toward a sector from the south-east to the west, define a NE–SW axis. The magnetic foliations have a best pole at 248/88 consistent with field data indicating transport toward the south-west (Fig. 11a). For three of these sites (GM1, GM2, and GM3), magnetic at site GM 7. The remaining 5 stations display inverse fabrics with foliation planes dip range from 7° to 10°NE with a magnetic lineation steeply dipping foliation planes (52–87°) and moderately to highly consistent with the flow direction inferred from K3 (Fig. 9). For the plunging lineations (45–81°). 126 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132

Fig. 10. Conceptual model for imbrication of magnetic foliation in two ideal cases. a) Magnetic foliation can be imbricated with Kmax axis oriented parallel and b) perpendicular to the ﬂow direction, and the related c), d) stereoplots of AMS axes. Note that the orientation of the magnetic foliation is parallel to the inferred ﬂow direction in both cases (redrawn from Giordano et al., 2008).

At Mbeng locality, site GM10 shows inverse fabric with a dipping 24° to the SSW; the magnetic lineation has approximately the subvertical foliation plane (83°) and a high plunge of lineation (65°). same direction, i.e. is parallel to flow. The magnetic foliation at the site In Mount Bambouto caldera where the ignimbrites of Baranka GM32 is also well defined and inclined slightly (3°) to the SE, whereas outcrop, station GM20 displays a subvertical foliation with moderate the magnetic lineation seems transverse to the flow. plunge of magnetic lineation of 47° to the north. Station GM19 shows At Mbu, the site GM21 is characterized by a magnetic foliation a moderately dipping magnetic foliation (pole at 188/60) and a plane imbricated (19°) to the WSW. transverse-to-flow magnetic lineation (trending at 335/26). At Bamenda, two sites (GM28 and GM29) display inverse fabrics. In the whitish units, only one AMS site at Mbou (GM11b with a Site GM27 exhibits normal fabric with magnetic foliation plane normal fabric) was used to infer the flow pattern in this locality: the slightly inclined (11°) to NW. The magnetic lineation for this location magnetic foliation pole points to the northwest, suggesting a NW to SE is subperpendicular to the direction of imbrication of magnetic flow. The remaining five sites at Nzemla II present nearly vertical fabrics foliation, i.e. appears transverse to flow. (foliation dips vary from 75° to 87° and plunges of lineation from 51° to Further north, at Sabga (Fig. 11b), most magnetic foliation planes 86°). Such fabrics where K3 and K1 appear to have been exchanged are are steeply dipping (53–75°) in all sites. called inverse fabrics (Rochette et al., 1999). However, it is worth At Bambili, the magnetic foliation in the dark gray units is well noticing that these sites also correspond to a prolate shape of their AMS defined and dips gently at 6° and 28° to the NE respectively for the sites ellipsoids (Tb0). These data question the significance of magnetic GM24 and GM26. Magnetic lineations have approximately the same NE directional data in ignimbrites. In our cases, the so-called inverse fabrics direction for the site GM26; on the other hand, at site GM24, the correspond to locations where abundant elutriation pipes were magnetic lineation is subperpendicular to the direction of imbrication of recognized in the field (Fig. 5c). Therefore, it is suggested that the magnetic foliation, i.e. transverse to flow. Station GM33 at Bambili vertical AMS fabrics may be a consequence of the elutriation process. displays an inverse fabric with subvertical foliation plane and lineation. This point will be examined later in the light of the SPO data. At site GM30, the magnetic foliation dips moderately (6°) to the NE and the magnetic lineation has the same NE trend. 5.4. Mount Bamenda 5.5. Shape preferred orientation fabrics In Mount Bamenda, 6 sites in the gray units yield interpretable AMS directional data that show a relatively consistent transport Shape preferred orientation (SPO) fabrics were studied in three pattern inferred by magnetic foliation or direction of imbrication dip. selected samples and the results are presented in Fig. 12. Only one of These sites situated at Mbengwi, Bambili, Bamenda and Mbu display these samples displays a normal AMS fabrics (GM 7 from Nzemla I dark well defined magnetic fabrics; the magnetic lineations have mostly gray ignimbrite). The two others display an inverse (or vertical) AMS low plunges (1 to 26°) and magnetic foliation planes are inclined fabric: GM 29 (Bamenda, light gray) displays a pronounced horizontal gently (3 to 24°). The remaining six sites at Sabga and Bamenda foliation, easily identified with the naked eye and GM 40 (Nzemla II; present mostly highly plunging (20 to 81°) of magnetic lineations and whitish) displays a steeply dipping foliation with field data pointing to a steeply dipping foliations (52 to 82°). transport towards SW. SPO fabrics are provided for all objects (crystals, AMS stations at Mbengwi, Mbu and Bamenda show synthetic fiammes, etc.). Fig. 12 presents the orientations of the principal axes stereograms with two concentrations of their magnetic foliation (ANBNC) of the SPO ellipsoids reconstructed from the 3 mutually poles, one corresponding to mildly dipping foliations and the other to orthogonal sections. All ellipsoids are oblate and well-defined (semi- steeply dipping foliations (Fig. 11b). angles of confidence ellipses b28° and incompatibility index √Fb10%). The two AMS sites of Mbengwi locality display normal fabrics. In the GM7 and GM29 provide especially good SPOs, despite the fact that site GM31, the result shows a well defined magnetic foliation plane foliation was not easily recognized with the naked eye in GM7. The flow

Fig. 11. Magnetic foliation and magnetic lineation map obtained for accepted AMS stations from (a) Mounts Bambouto and (b) Bamenda ignimbrites. Projection diagrams (lower hemisphere, 2% contour intervals) for foliation poles and lineation are also shown; the plots are made with all specimens. M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 127 128 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132

Fig. 12. SPO results for samples GM7, GM29 and GM40. Lower hemisphere projection diagrams showing all data and mean principal directions (A, B and C) of the SPO fabric ellipsoid in the center of their respective error ellipse (1σ); P’ (anisotropy degree) and T (shape parameter) are calculated using the formulae of Jelinek (1981); s1 and s2 are the semi-angles of the error ellipse; √F (%) is the incompatibility index (Launeau et al., 2010). direction is easily derived from the orientation of the long axis (A) of foliation and an oblate ellipsoid shape. In these cases, the source of their SPOs. In sample GM7 with normal AMS fabric, this direction fits AMS is likely due to a primary orientation and distribution of iron-rich with the flow direction derived from the magnetic foliation pole. The minerals, whatever their nature and size, and AMS can be used for the SPOfabricinGM29suggestsaflow direction towards the north, determination of ignimbrite fabrics, and thus for the recognition of whereas its inverse magnetic fabric could not be used. Examination of their flow direction in agreement with Knight et al. (1986), Wolff et al. the sample confirms the existence of vertical channels possibly related (1989), Buesch (1992, 1990) and Seaman et al. (1991). to fluid or gas escape traced by secondary magnetic phases. Therefore, its AMS fabric is a post-depositional fabric. Sample GM 40 contains many 6.2. Flow directions and identification of emission centers objects and its SPO anisotropy (P’) is lesser than in the other two samples. Position of the C axis is consistent with the flow direction For a normal magnetic fabric, the mean magnetic foliation plane, deduced in the field, hence corresponding to a high imbrication angle. i.e. the K1–K2 plane approximates the flow plane. Commonly, the The A axis is transverse to the flow, a situation that can happen in clast- magnetic foliation may be inclined with respect to the depositional rich currents (Archanjo et al., 2006). The SPO fabric can still be used for surface, or stratification, similar to the imbrications of sediments flow determination, whereas the vertical AMS fabric is likely post- (Laberge et al., 2008). Thus, the magnetic foliation plane differs in depositional. orientation (imbrication angle) relative to the flow plane (Ellwood, 1982; Knight et al., 1986) and dips in direction opposite to the flow 6. Discussion direction. The magnetic lineation can be either parallel or perpendicular to the flow direction (Fig. 10), due to the rolling or saltation of 6.1. Origin and use of AMS grains within the flow (Ort et al., 2003). The imbrication dip direction is inferred to point towards the source area. Comparing both the trend

As was just shown above, there is a very good consistency between of the K1 axis and dip direction of the K1–K2 plane to evaluate the ﬂow the magnetic fabric and the rock fabric (as can be recognized by pattern provides a way to quickly identify anomalous K1 lineation studying major minerals or ﬁammes), when the AMS fabric is direction (e.g., Knight et al., 1986; Hillhouse and Wells, 1991; Seaman “normal”, that is with a subhorizontal to mildly dipping (imbricate) et al., 1991; Ort, 1993; Le Pennec et al., 1998). M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 129

Fig. 13. Map showing mean ﬂow directions inferred from the orientation of imbrications of magnetic foliation plane and from ﬁeld indicators.

These considerations were already applied to our AMS directional At Dschang, the general flow direction inferred by magnetic data is data. Flow directions for each site characterized by a normal fabric are towards SW for sites GM1, GM2 and GM3. Flow directions inferred plotted in Fig. 13. The source of the ignimbrites is assumed to be the from the magnetic data are in close agreement with the direction area from which the flow lines radiate, that is the intersection of suggested by field indicators because imbricated fiammes (around extended flow directions inferred from the AMS (Palmer and site GM1) also suggest that flow was directed towards SW. The AMS MacDonald, 1999; Alva-valdivia et al., 2005; Laberge et al., 2008; data for the remaining two stations (GM4 and GM5) are nearly Petronis and Geissman, 2008). orthogonal to the general trend. We infer that change in the flow 130 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132

imbrications angle and K1 directions across the area reﬂect variations processes (Rochette et al., 1999), such as hydrothermalism or post in local slope which increases the overall dispersion of the transport emplacement modiﬁcation (tectonic effect), or to the presence of SD data. Indeed, the effect of subtle topography on AMS directional data grains in the rocks. In the samples consisting mainly of SD grains, has been recognized in other ignimbrites (e.g., Buesch, 1992; Ort et al., inverse fabric can be produced by exchange of K1 and K3 positions 2003; Petronis and Geissman, 2008). Ignimbrite of Dschang is situated compared with those in which MD grains are predominant (Rochette, in the NE–SW trending of Menoua valley, and a similar topographic 1988; Tarling and Hrouda, 1993; Dunlop and Ozdemir, 1997). effect may have channeled the pyroclastic density current from the Taira and Scholle (1979) and Mimura (1984) demonstrated that caldera of Mounts Bambouto to this locality. imbrications of paramagnetic minerals caused by grain collision can

At Lepo, excepted site GM35 with approximately E–W flow result in higher plunges of K1 up to 30°, and even to over 50°. This may direction, AMS and field transport directions clearly show that the have occurred because of high percentages of crystals (35–45%) in the emission center of this pyroclastic flow is situated north of this area, ignimbrites which show inverse fabrics (Mbeng and Nzemla II). which is the position of the Bambouto caldera. However, it may not be sufficient to explain actual nearly vertical At Nzemla I and Mbou, excepted site GM18 (with anomalous ESE– lineations and foliations. WNW transport direction, certainly also due to variation of local The SPO fabric analysis has revealed that the vertical magnetic topography), general flow direction inferred from AMS is to SE fabrics are different from the SPO fabrics of all objects (Fig. 12). It is (Fig. 13), consistent with flow direction infer from field indicators also worth to notice that these vertical fabrics are generally tenuous (imbricated lithic fragments) around station GM11. Thus, magnetic and prolate. Subvertical elutriation pipes have been recognized in the fabric and field indicators also evidently indicate Bambouto caldera as upper part of incipiently welded ignimbrites of Bambili, Nzemla II source area in the NE of these localities. (WU, GM12 and GM33) and Mbeng (DGU, GM10). Post-depositional At Nzemla II and Mbeng, because of the vertical magnetic fabrics, vertical structures were also observed in GM29. It is suggested that no flow direction can be recognized. In addition, all stations display a the elutriation process may have been responsible for the vertical relatively low magnetic anisotropy and susceptibility. Field indicators fabrics. Tiny secondary oxide minerals may have crystallized or be represented by elongate imbricated pumices and lithic fragments in deposited in the pore boundaries produced from separation and one site (GM40) yield a flow azimuth toward 245°, confirmed by the upwards migration of dust-loaded vapor phases produced during SPO analysis, that may suggest eruption from a source located to the post-emplacement of ignimbrites. NE of this locality, possibly covered with younger basaltic flows. Unfortunately, there is no other evidence supporting this hypothesis. 7. Conclusions At Mbengwi, Mbu and Bamenda the general flow direction varies from one locality to another but seems to radiate from the same AMS data reported here for the Bambouto and Bamenda massifs region, namely the Mbu caldera as probable emission center. This indicate distinct transport directions of pyroclastic currents respon- location is also revealed by field indicators around site GM28 which sible for the ignimbritic deposits. More precisely, AMS can be used indicated flow was towards azimuth 46° (NE). when the AMS fabric has a so-called normal orientation, i.e. is At Sabga, Bambili and Big Babanki, flow direction is from NE to SW. characterized by a subhorizontal to mildly inclined imbrication The field indicators around site GM22 and GM24 are imbricated foliation. The Bambouto caldera seems to be the source of the main fiammes indicating that flow pattern was approximately toward ignimbrite sheets of the Mount Bambouto. In Mounts Bamenda, flow azimuth 222° and 235° respectively. The suspected emission center is patterns indicate two different emission centers of ignimbrites. The then localized to the NE of these localities. In Mount Oku the important Oku crater, located at the NE of Mount Bamenda, is the probable depression (1.8×2.2 km) in which Oku Lake is localized, is the probable source of Bambili, Sabga and Big Babanki ignimbrites, whereas Santa- source of these ignimbrite sheets (or at least their welded Tlm parts). Mbu caldera is the suspected source of Mbengwi, Bamenda and Mbu This unexpected result definitely strengthens the interest of AMS ignimbrites. studies to infer flow direction and emission center in ignimbrites. Some sites display vertical AMS fabrics, different from the SPO fabrics of all objects (crystals, fiammes, etc.). These vertical and 6.3. Ignimbrite emplacement mode prolate fabrics are secondary fabrics attributed to a post-depositional elutriation process also recognized in the field. In these cases, the clast The AMS data from the Mounts Bambouto and Bamenda and inclusion shape fabrics obtained by image analysis is the more ignimbrites yield important information about the depositional reliable way to determine flow directions and sources. system of the pyroclastic flow. In most stations, magnetic foliation planes (and sometimes K axes) consistently parallel downhill 1 Acknowledgments directions, commonly with an upslope imbrication. Flow directions interpreted from these data indicate deposition from pyroclastic This work has been supported by the SCAC (Service de Coopération density currents that traveled downhill. et d'Action Culturelle de la France au Cameroun), EGIDE (Centre The current that transported the pyroclasts out from the vent must français pour l'accueil et les échanges internationaux) and the French have moved in a radial direction at Lepo (Mts Bambouto) and in Mts Government. Field work was partially supported by the IRD-CORUS 2 Bamenda commonly around the probable emission center (Fig. 13). project of M. Jessell and J.L. Bouchez from LMTG (Toulouse). Technical There were likely no major topographic obstacles to affect the current assistance by R. Siqueira, F. and de Parseval is warmly acknowledged. in these cases. At the other locations (Dschang), the depositional system was strongly affected by topography or channeled by valleys. This type of References ignimbrite is generated by a dense pyroclastic flow current which Alva-Valdivia, L.M., Rosas-Elguera, J., Bravo-Medina, T., Urrutia-Fucugauchi, J., Henry, B., remains generally confined in the valley (Branney and Kokelaar, 2002). Caballero, C., Rivas-Sanchez, M.L., Goguitchaichvili, A., López-Loera, H., 2005. Paleomagnetic and magnetic fabric studies of the San Gaspar ignimbrite, western fi Mexico: constraints on emplacement mode and source vents. J. Volcanol. 6.4. Signi cance of vertical fabrics Geotherm. Res. 147, 68–80. Archanjo, C.J., Silva, M.G., Castro, J.C., Launeau, P., TRindade, R.I.F., Macedo, J.W.P., 2006. The so-called inverse AMS fabrics, actually vertical (and generally AMS and grain shape fabric of the Late Palaeozoic diamictites of the Southeastern Paraná Basin, Brazil. J. Geol. Soc. London 163, 95–106. prolate) fabrics, are the most frequent fabrics in the WU of Nzemla II Ballentine, C.J., Lee, D.C., Halliday, A.N., 1997. Hafnium isotopic studies of the Cameroon ignimbrite. Inverse fabrics have been attributed to secondary line and new HIMU paradoxes. Chem. Geol. 139, 111–124. M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132 131

Bear, E.M., Fisher, R.V., Fuller, M., Valentine, G., 1997. Turbulent transport and Jackson, M., Sun, W.W., Craddock, J.P., 1992. The rock magnetic fingerprint of chemical deposition of Ito pyroclastic flow: determinations using anisotropy of magnetic remagnetization in midcontinental paleozoic carbonates. Geophys. Res. Lett. 19, susceptibility. J. Geophys. Res. 102, 22565–22586. 781–784. Behncke, B., 2009. Hazards from pyroclastic density currents at Mt. Etna (Italy). J. Jelinek, V., 1981. Characterization of the magnetic fabric of rocks. Tectonophysics 79, Volcanol. Geotherm. Res. 180, 148–160. 563–567. Behncke, B., Calvari, S., Giammanco, S., Neri, M., Pinkerton, H., 2008. Pyroclastic density Jelinek, V., 1978. Statistical processing of anisotropy of magnetic susceptibility currents resulting from the interaction of basaltic magma with hydrothermally measured on groups of specimens. Stud. Geophys. Geod. 22, 50–62. altered rock: an example from the 2006 summit eruptions of Mount Etna. Italy. Kagou Dongmo, A., Nkouathio, D.G., Pouclet, A., Bardintzeff, J.-M., Wandji, P., Nono, A., Bull. Volcanol. 70, 1249–1268. doi:10.1007/s00445-008-0200-7. Guillou, H., 2010. The discovery of Late Quaternary basalt on Mount Bambouto: Bouchez, J.L., 2000. Anisotropie de susceptibilité magnétique et fabrique des granites. implications for recent widespread volcanic activity in the southern Cameroon C.R.Acad.Sci.Paris330,1–14. Line. J. Afr. Earth. 57, 87–108. Branney, M.J., Kokelaar, P., 2002. Pyroclastic density currents and the sedimentation of Kamata, H., Mimura, K., 1983. Flow directions inferred from imbrication in the Handa ignimbrites. Geol. Soc. London Mem. 27, 152p. pyroclastic flow deposit in Japan. Bull. Volcanol. 46, 277–282. Buesch, D.C., 1992. Incorporation and redistribution of locally-derived lithic fragments Kamgang, P., Njonfang, E., Chazot, G., Tchoua, F.M., 2007. Géochimie et géochronologie within a pyroclastic flow. Geol. Soc. Am. Bull. 104, 1193–1207. des laves felsiques des Mounts Bamenda (ligne volcanique du Cameroun). C. R Cagnoli, B., Tarling, D.H., 1997. The reliability of anisotropy of magnetic susceptibility Géoscience 339, 659–666. (AMS) data as flow direction indicators in friable base surge and ignimbrite Kamgang, P., Chazot, G., Njonfang, E., Tchoua, F.M., 2008. Geochemistry and deposits: Italian examples. J. Volcanol. Geotherm. Res. 75, 309–320. geochronology of mafic rocks from Bamenda Mountains (Cameroon): source Canon-Tapia, E., Walker, G.P.L., Herrero-Bervera, E., 1995. Magnetic fabric and flow composition and crustal contamination along the Cameroon Volcanic Line. C. R direction in basaltic pahoehoe lava of Xitle Volcano, Mexico. J. Volcanol. Geotherm. Géoscience 340, 850–857. Res. 65, 249–263. Knight, M.D., Walker, G.P.L., Ellwood, B.B., Diehl, J.F., 1986. Stratigraphy, paleomagne- Charbonnier, S.J., Gertisser, R., 2008. Field observations and surface characteristics of tism and magnetic fabric of the Toba tuffs; constraints on the sources and eruptive pristine block-and-ash flow deposits from the 2006 eruption of Merapi Volcano, style. J. Geophys. Res. B 91, 10355–10382. Java, Indonesia. J. Volcanol. Geotherm. Res. 117, 971–982. Kruiver, P.P., Dekkers, M.J., Heslop, D., 2001. Quantification of magnetic coercivity Deruelle, B., Moreau, C., Nkoumbou, C., Kambou, R., Lissom, J., Njonfang, E., Ghogomu, components by the analysis of acquisition curves of isothermal remanent R.T.,Nono,A.,1991.TheCameroonline:areview.In:Kampuzu,A.B.,Lubala,R.T. magnetization. Earth planet. Sci. Lett. 189, 269–276. (Eds.), Magmatism in Extensional Structural Settings. The Phanerozoic African LaBerge, R.D., Porreca, M., Massimo Mattei, M., Giordano, G., Cas, R.A.F., 2008. Plate. Springer-Verlag, Heidelbelg, pp. 237–274. Meandering flow of a pyroclastic density current documented by the anisotropy Déruelle, B., Ngounouno, I., Demaiffe, D., 2007. The “Cameroon Hot Line” (CHL): a of magnetic susceptibility (AMS) in the quartz latite ignimbrite of the Pleistocene unique example of active alkaline intraplate structure in both oceanic and Mounte Cimino volcanic centre (central Italy). Tectonophysics 466, 64–78. continental lithospheres. C. R. Geosci. 339, 589–600. Launeau, P., 2004. Evidence of magnetic flow by 2-D image analysis of 3-D shape Dunlop, D.P., Ozdemir, O., 1997. Rock magnetism: fundamentals and frontiers. preferred orientation distributions. Bull. Soc. Géol. Fr. 175, 331–350. Cambridge Univ. Press, Cambridge, UK. 573p. Launeau, P., Robin, P.Y., 2005. Determination of fabric and strain ellipsoids from Edmonds, M., Herd, R.A., Strutt, M.H., 2006. Tephra deposits associated with a large lava measured sectional ellipses—implementation and applications. J. Structural Geolog. dome collapse, Soufriere Hills volcano, Montserrat, 12–15 July 2003. J. Volcanol. 27, 2223–2233. Geotherm Res. 153, 313–330. Launeau, P., Archanjo, C.J., Picard, D., Arbaret, L., Robin, P.Y., 2010. Two- and three- Ellwood, B.B., 1982. Estimate of flow direction for calc-alkaline welded tuffs and dimensional shape fabric analysis by the intercept method in grey levels. paleomagnetism data reliability from anisotropy of magnetic susceptibility Tectonophysics 492, 230–239. measurements: Central San Juan Mountains, southwest Colorado. Earth Planet. Le Pennec, J.L., Chen, Y., Diot, H., Froger, J.L., Gourgaud, A., 1998. Interpretation of Sci. Lett. 59, 303–314. anisotropy of magnetic susceptibility fabric of ignimbrites in terms of kinematic Elston, W.E., Smith, E.I., 1970. Determination of flow direction of rhyolitic ash-flow tuffs and sedimentological mechanisms: an Anatolian case-study. Earth Planet. Sci. Lett. from fluidal textures. Geol. Soc. Am. Bull. 81, 3393–3406. 157, 105–127. Fisher, R.V., Orsi, G., Ort, M.H., Heiken, G., 1993. Mobility of a large-volume pyroclastic Le Pennec, J.L., 2000. Identifying ash flow sources with directional data: an applica- flow-emplacement of the Campanian Ignimbrite. Italy J. Volcanol. Geotherm. Res. tion to the Kizilkaya ignimbrite, central Anatolia. J. Geophys. Res. B 105, 56, 205–220. 28427–28441. Fitton, J.G., 1987. The Cameroon Line, West Africa: a comparison between oceanic and Lee, D.C., Halliday, A.N., Fitton, J.G., Poli, G., 1994. Isotopic variations with distance and continental alkaline volcanism. In: Fitton, J.G., Upton, B.G.J. (Eds.), Alkaline Igneous time in the volcanic islands of the Cameroon Line: evidence for a mantle plume Rocks: Geol. Soc. Spec. Publ., 30, pp. 273–291. origin. Earth Planet Sci Lett. 123, 119–138. Font, E., Ernesto, M., Silva, P.F., Correia, P.B., Nascimento, M.A.L., 2009. Palaeomagnetism, MacDonald, W.D., Palmer, H.C., Hayatsu, A., 1998. Structural rotation and volcanic rock magnetism and AMS of the Cabo Magmatic Province, NE Brazil, and the opening source implications of magnetic data from Eocene volcanic rocks. SW Idaho. Earth of South Atlantic. Geophys. J. Int.. doi:10.1111/j.1365-246X.2009.04333.x Planet Sci Lett. 156, 225–237. Fosso, J., Menard, J.-J., Bardintzeff, J.-M., Wandji, P., Tchoua, F.M., Bellon, H., 2005. Les Marzoli, A., Piccirillo, E.M., Renne, P.R., Bellieni, G., Iacumin, M., Nyobe, J.B., Tongwa, laves du Mount Bangou: une première manifestation volcanique éocène, à affinité A.T., 2000. The Cameroon volcanic Line revisited: petrogenesis of continental transitionnelle, de la Ligne du Cameroun. C. R. Géosci. 337, 315–325. basaltic magma from lithospheric and asthenospheric mantle source. J. Petrol. Froggatt, P.C., Wilson, C.J.N., Walker, G.P.L., 1981. Orientation of logs in the Taupo 41, 87–109. ignimbrite as an indicator of flow direction and vent position. Geology 9, McDonald, W.D., Palmer, H.C., 1990. Flow directions in ash flow tuffs: a comparison of 109–111. geological and magnetic susceptibility measurements. Tshirege member (upper Giordano, G., Porreca, M., Musacchio, P., Mattei, M., 2008. The Holocene Secche di Bandelir Tuff) Valles caldera. New Mexico. USA. Bull. Volcanol. 53, 45–59. Lazzaro phreatomagmatic succession (Stromboli, Italy): evidence of pyroclastic Mimura, K., 1984. Imbrication, flow direction and possible source areas of the pumice- density current origin deduced by facies analysis and AMS flow directions. Bull. flow tuffs near Bend, Oregon. USA. J. Volcanol. Geotherm. Res. 21, 45–60. Volcanol. 70, 1221–1236. Moundi, A., Wandji, P., Bardintzeff, J.-M., Ménard, J.-J., Okomo Atouba, L.C., Farouk Gouhier, J., Nougier, J., Nougier, D., 1974. Contribution à l'étude volcanologique du Mouncherou, O., Reusser, E., Bellon, H., Tchoua, F.M., 2007. Les basaltes Eocène à Cameroun («Ligne du Cameroun»-Adamawa), 17. Ann. Fac. Sci. Univ, Yaoundé affinité transitionnelle du plateau Bamoun, témoins d'un réservoir mantellique Cameroun, pp. 3–48. enrichi sous la ligne volcanique du Cameroun. C. R. Géosciences 339, 831–837. Gountié Dedzo, M., Nono, A., Njonfang, E., Zangmo Tefogoum, G., Kagou Dongmo, A., Ngako, V., Njonfang, E., Tongwa Aka, F., Affaton, P., Metuk Nnange, J., 2006. The North– Nkouathio, D.G., 2009. Le volcanisme ignimbritique de la Ligne du Cameroun South Paleozoic to Quaternary trend of alkaline magmatism from Niger-Nigeria to (Mounts Bambouto et Bamenda, Ouest-Cameroun): signification dans la genèse Cameroon: complex interaction between hotspots and Precambrian faults. J Afr des caldeiras. Int. Coll. 3MA Beni Mellal, April 23–24. Morocco, pp. 74–75. Earth Sci 45, 241–256. Grégoire, V., Darrozes, J., Gaillot, P., Nédélec, A., Launeau, P., 1998. Magnetite grain Njonfang, E., Nono, A., 2003. Clinopyroxene from some felsic alkaline rocks of the shape fabric and distribution anisotropy versus rock magnetic fabric: a 3D-case Cameroon Line, Central Africa: petrological implications. Eur. J. Mineral. 15, 527–542. study. J. Struct. Geol. 20, 937–944. Nkouathio, D.G., Ménard, J.-J., Wandji, P., Bardintzeff, J.-M., 2002. The Tombel graben Hart, K., Carey, S., Sigurdsson, H., Sparks, R.S.J., Robertson, R.E.A., 2004. Discharge of (West Cameroon): a recent monogenetic volcanic field of the Cameroon Line. J. Afr. pyroclastic flows into the sea during the 1996–1998 eruptions of the Soufriere Hills Earth Sci. 35, 285–300. volcano. Montserrat. Bull Volc. 66, 599–614. Nkouathio, D.G., Kagou Dongmo, A., Bardintzeff, J.-M., Wandji, P., Bellon, H., Pouclet, A., Hillhouse, J.W., Wells, R.E., 1991. Magnetic fabric, flow directions, and source area of the 2008. Evolution of volcanism in graben and horst structures along the Cenozoic lower Miocene Peach Springs Tuff in Arizona, California, and Nevada. J. Geophys. Cameroon Line (Africa): implications for tectonic evolution and mantle source Res. 96, 12443–12460. composition. . 94, 287–303. Incoronato, A., Addison, F.T., Tarling, D.H., Nardi, G., Pescatore, T., 1983. Magnetic fabric Nono, A., Nkouathio, D.G., Gountié Dedzo, M., Njonfang, E., Kagou Dongmo, A., Tchoua, investigations of pyroclastic deposits from Phlegrean Fields, southern Italy. Nature F.M., 2003. Zonal and vertical variations in welding rate and composition of 306, 461–463. ignimbrites in the Bambouto volcano (Cameroon Line, Central Africa): volcano- Itiga, Z., Chakam Tagheu, P.J., Wotchoko, P., Wandji, P., Bardintzeff, J.M., Bellon, H., 2004. logical importance. Geophys. Res. Abstr. 5, 14541. La Ligne du Cameroun: volcanologie et géochronologie de trois régions (Mount Nono, A., Njonfang, E., Kagou Dongmo, A., Nkouathio, D.G., Tchoua, F.M., 2004. Manengouba, plaine du Noun et Tchabal Gangdaba). Géochronique 91, 13–16. Pyroclastic deposits of the Bambouto Volcano (Cameroon Line, Central Africa): Jackson, M., Rochette, P., Filon, G., Banerjee, S., Marvin, J., 1993. Rock magnetism of evidence of a strombolian initial phase. J. Afr. Earth. Sci. 39, 409–414. remagnetized Paleozoic carbonates: low-temperature behavior and susceptibility Nzolang, C., Kagami, H., Nzenti, P., Hotz, F., 2003. Geochemistry and preliminary Sr–Nd characteristics. J. Geophys. Res. 98, 6217–6225. isotopic data on the Neoproterozoic granitoids from the Bantoum Area, West 132 M. Gountié Dedzo et al. / Journal of Volcanology and Geothermal Research 203 (2011) 113–132

Cameroon: evidence for a derivation from a Paleoproterozoic to Archean crust. Saviulo, G., Secco, L., Marzoli, A., Piccirillo, E.M., Nyobe, J.B., 2000. Ca-rich pyroxene from Polar Geoscience 16, 196–226. basic and silicic volcanic rocks from the Cameroon Volcanic Line (West-Africa): Ort, M.H., 1993. Eruptive processes of caldera formation in a nested downsag-collapse crystal chemistry and petrological relationships. . 70, 73–88. caldera: Cerro Panizos, central Andes Mountains. J. Volcanol. Geotherm. Res. 56, Seaman, S.J., McIntosh, W.C., Geissman, J.W., Williams, M.L., Elston, W.E., 1991. 221–252. Magnetic fabrics of the Bloodgood Canyon and Shelley Peak Tuffs, southwestern Ort, M.H., Orsi, G., Pappalardo, L., 2003. Anisotropy of magnetic susceptibility studies of New Mexico: implications for emplacement and alteration processes. Bull. depositional processes in the Campanian Ignimbrite. Italy. Bull. Volcanol. 65, 55–72. Volcanol. 53, 460–476. doi:10.1007/s00445-002-0241-2. Seaman, S.J., Williams, M.L., 1992. Center-to-center analysis and flow fabric character- Ort, M.H., Rosi, M., Anderson, C.D., 1999. Correlation of deposits and vent locations of ization in ash-flow tuffs. Bull. Volcanol. 54, 319–328. the proximal Campanian Ignimbrite deposits, Campi Flegrei, Itali, based on natural Suzuki, K., Ui, T., 1982. Grain orientation and depositional ramps as flow direction remanent magnetization and anisotropy of magnetic susceptibility characteristics. indicators of large-scale pyroclastic flow deposits, Japan. Geology 10, 429–432. J. Volcanol. Geotherm. Res. 91, 167–178. Suzuki, K., Ui, T., 1983. Factors governing the flow lineation of a large-scale pyroclastic Palmer, H.C., Macdonald, W.D., Hayatsu, A., 1991. Magnetic, structural and geochro- flow an example in the Ata pyroclastic flow deposit. Japan. Bull. Volcanol. 46, 71–81. nologic evidence bearing on volcanic sources and Oligocene deformation of ash Suzuki, K., Ui, T., 1988. Depositional ramps: asymmetrical distribution structure in the flow tuffs, northeast Nevada. J. Geophys. Res. 96, 2185–2202. Ata pyroclastic flow deposit. Japan. Bull. Volcanol. 50, 26–34. Palmer, H.C., MacDonald, W.D., 1999. Anisotropy of magnetic susceptibility in relation Taira, A., Scholle, P.A., 1979. Deposition of resedimented sandstone beds in the Pico to source vents of ignimbrites: empirical observations. Tectonophysic 307, formation, Ventura Basin, California, as interpreted from magnetic fabrics 207–218. measurements. Geol. Soc. Am. Bull. 90, 952–962. Petronis, M.S., Geissman, J.W., 2008. Anisotropy of magnetic susceptibility data bearing Tarling, D.H., Hrouda, F., 1993. The magnetic anisotropy of rocks. Capman and Hall, on the transport direction of mid-tertiary regional ignimbrites, Candelaria Hills London. 217p. area, West-Central Nevada. Bull. Volcanol. 69, 930–951. doi:10.1007/s00445-008- Tchoua, F.M., 1973. Sur l'existence d'une phase initiale ignimbritique dans le 0212-3. volcanisme des Mounts Bambouto (Cameroun). C.R. Acad. Sci. Paris 276, Porreca, M., Mattei, M., Giordano, G., De Rita, D., Funiciello, R., 2003. Magnetic fabric and 2863–2866. implications for pyroclastic flow and lahar emplacement, Albano maar. Italy. J. Toteu, S.F., Van Schmus, W.R., Penaye, J., Michard, A., 2001. New U–Pb and Sm–Nd data Geophys. Res. B 108, 2264. doi:10.1029/2002JB002102. from north Cameroon and its bearing on the pre-Pan African history of central Potter, D.B., Oberthal, C.M., 1987. Vent sites and flow directions of the Otowi ash flows Africa. Precamb. Res. 108, 45–73. (lower Bandelier Tuff), New Mexico. Geol. Soc. Amer. Bull. 98, 66–76. Ui, T., Suzuki-Kamata, K., Matsusue, R., Fujita, K., Metsugi, H., Araki, M., 1989. Flow Rochette, P., 1987. Magnetic susceptibility of the rock matrix related to magnetic fabric behavior of large-scale flows. Evidence obtained from petrofabric analysis. Bull. studies. J. Struct. Geol. 9, 1015–1020. Volcanol. 51, 115–122. Rochette, P., 1988. Inverse magnetic fabrics in carbonate-bearing rocks. Earth Planet. Wang, X., Roberts, J., Schmidt, P., 2001. Flow directions of carboniferous ignimbrites, Sci. Lett. 90, 229–237. southern New England Oregon, Australia, using anisotropy of magnetic suscepti- Rochette, P., Aubourg, C., Perrin, M., 1999. Is this magnetic fabric normal? A review and bility. J. Volcanol. Geotherm. Res. 110, 1–25. case studies in volcanic formations. Tectonophysics 307, 219–234. Wolff, J.A., Ellwood, B.B., Sachs, S.D., 1989. Anisotropy of magnetic susceptibility in Robertson, D.J., France, D.E., 1994. Discrimination of remanence-carrying minerals in welded tuffs: application to a welded-tuff dyke in the Tertiary Trans-Pecos Texas mixtures, using isothermal remanent magnetization acquisition curves. Phys. Earth volcanic province. USA. Bull. Volcanol. 51, 299–310. planet. Inter. 82, 223–234. Youmen, D., Schmincke, H.-U., Lissom, J., Etame, J., 2005. Données géochronologiques : Robin, P.Y., 2002. Determination of fabric and strain ellipsoids from measured sectional mise en évidence des différentes phases volcaniques au Miocène dans les monts ellipses — theory. Journ. Struct. Geol. 24, 531–544. Bambouto (Ligne du Cameroun). Sci. Technol. Dev. 11, 49–57.