Journal of Volcanology and Geothermal Research 179 (2009) 206–216

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

j ourna l homepage: www.e lsev ie r.com/ locate / jvo lgeores

Eruptive and earthquake activities related to the 2000 eruption of Mount Cameroonvolcano (West Africa)

B. Ateba a, C. Dorbath b,e,⁎, L. Dorbath b,e, N. Ntepe a, M. Frogneux b, F.T. Aka a, J.V. Hell c,J.C. Delmond d, D. Manguelle f

a Institut de Recherches Géologiques et Minières, Branch for Geophysical and Volcanological Research, P.O. Box 370, Buea, Cameroonb École et Observatoire des Sciences de la Terre, IPGS, 5 rue René Descartes, 67084 Strasbourg Cedex, Francec Institut de Recherches Géologiques et Minières, P.O. Box 4110, Yaounde, Cameroond Institut de Physique du Globe (IPGP), 4 Place Jussieu, 75252 Paris Cedex 05, Francee IRD, 213, rue La Fayette, 75480 Paris Cedex 10, Francef Faculté des Sciences, Département de Physique, UYI, B.P. 812 Yaounde, Cameroon

⁎ Corresponding author. École et Observatoire des ScRené Descartes, 67084 Strasbourg Cedex, France.

E-mail address: catherine.dorbath@eost.u-strasbg.fr

0377-0273/$ – see front matter © 2008 Elsevier B.V. Aldoi:10.1016/j.jvolgeores.2008.11.021

a b s t r a c t

a r t i c l e i n f o

Article history:

Mount Cameroon is an activ Received 6 February 2007Accepted 4 November 2008Available online 3 December 2008

Keywords:Mount Cameroonvolcanoeruptionseismicityearthquake swarms

e volcano located in the Gulf of Guinea, west of Central Africa. After the March–April 1999 eruption on the SW flank, another eruption of the volcano occurred in 2000. It took place fromthree sites on the southwest flank and near the summit. The first eruptive site was located 500 m to thesouthwest of the summit, at 3900 m altitude. Activity on this site was mainly explosive with no lava flow. Thesecond site was located between 3220 and 3470 m altitude. Lava was emitted along NNE–SSE fissures fromthis site and flew towards Buea, the main city of the area, stopping ~4 km from the first houses. The last sitewas located in the south western flank at 2750 m altitude. The lava ejected from an old cone near the first1999 eruptive site was divided into two branches, for a total length of around 1 km. The location of activevolcanic cones in 1999 and 2000 seems to be linked to the local tectonics. The pre-eruptive period wascharacterized by a seismic swarm which may be a precursor recorded in March 2000 by an analogue seismicstation. The main shock was a magnitude 3.2 event, and was felt by the population in Ekona town located onthe eastern flank. It had a Modified Mercalli intensity of III–IV. When the eruption started, a temporarynetwork of short period 3-component seismic stations was set up around the volcano to improve themonitoring of seismic activity. The co-eruptive period from late May to September was characterized bysequences of earthquake swarms, volcanic tremor and a family of earthquakes having similar waveform andappearing regularly in August and early September. Some of the earthquakes were felt by the population inBuea and its environments. The largest seismic event recorded had a magnitude of 4. During the post-eruptive period from mid-September to December, seismicity returned to its background level of 1–3earthquakes per 3 days. Hypocenter locations reveal a linear narrow structure under the summit zone whichcould represent the magmatic conduit of the volcano. The frequency/magnitude relationship revealed a b-value of 1.43 higher than those previously determined, but more representative of volcanic media. Seismicenergy release was gradual after the 2000 eruption started.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Mount Cameroon is situated in the middle of the CameroonVolcanic Line (CVL) and separates its oceanic from its continental parts(Fig. 1). The CVL is an alignment striking N30°E, about 1600 km longand is formed by the volcanic islands of the Gulf of Guinea, continentalvolcanoes and plutonic complexes. The volcanic islands of the oceanicsector are (from south to north): Annobon (or Pagalu), São Tome,

iences de la Terre, IPGS, 5 rue

(C. Dorbath).

l rights reserved.

Principe and Bioko. The continental volcanic massifs are MountCameroon (4095 m), Mount Manengouba (2420 m), Mount Bambouto(2679m) andMount Oku (3011 m). The northern extension of the CVLis divided into two branches: one of which stretches to the Biu Plateauin north-eastern Nigeria, while the other passes through theAdamawa Plateau (Fitton, 1980; Halliday et al., 1988; Lee et al., 1994;Marzoli et al., 2000; Aka et al., 2004). Two major directions mark thetectonics along the CVL, namely: N20–40°E and N70°E. The N20–40°Edirection is that of the volcanic line along which several parallel andtransversal fractures are buried by lava flows (Moreau et al., 1987). TheN70°E direction corresponds to a major Precambrian shear zoneconsidered as the continuation of the Pernambuco lineament in

Fig. 1. Location of Mount Cameroon and of the main volcanic centers belonging to the Cameroon Volcanic Line.

207B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

northeast of Brazil, prior to the continental separation (Burke et al.,1971; Sibuet and Mascle, 1978; Cornacchia and Dars, 1983; Browneand Fairhead, 1983; Ngako et al., 2003).

Mount Cameroon is a stratovolcanomade of a piling of basaltic lavaflows, pyroclastic deposits and lahars (Fig. 2). It has a southwest–northeast elongated shape that is a little oblique to the generaldirection of the CVL. It measures 50 by 35 km at its base. Gèze (1943)counted more than one hundred volcanic cones, most of which arealigned along its northeast–southwest axis. However, isolated conesalso exist on its flanks.

Mount Cameroon rises between two sedimentary basins: theDouala basin, to the east and the Rio del Rey basin, to the west. Itssymmetry is truncated by a secondary peak, Mount Etinde (1715 m)located on its south western flank. It is considered as a stepped horstparallel to the direction of the CVL (Vincent, 1971; Gouhier et al., 1974)and, in that horst/graben structure, the central uprising part iscompensated by the collapse of lateral sedimentary basins (Déruelleet al., 1987). Twomain faults oriented SSW–NNEweremapped aroundMount Cameroon by Hedberg (1968): the Tiko fault to the East andthe Boa fault to the West (Fig. 2). Zogning (1988) also identified theLikomba, Esuke, Ombé and Tole–Ekona faults on the Eastflank. TheNWflank is marked by a circular steep structure (the NW cirque) and the

Bokosso fault system of direction NNW–SSE. The ‘Fako’, the highestpeak on the summit area, is situated at the intersection of two bigstructural directions namely, the axis of the volcano, N45–50°E, andthe direction that is almost perpendicular (Gèze,1953; Zogning,1988).

The age of Mount Cameroon is not well known. Samples collectedon and around the volcano give K–Ar ages which are less than 1 Ma(Piper and Richardson, 1972). Samples of Mount Etinde do not exceed0.7 Ma (Nkoumbou et al., 1995).

Records of past volcanic eruptions onMount Cameroon are given byGèze (1953), Fitton et al. (1983) and Déruelle et al. (1983, 1987). Seveneruptions were reported in the 20th century: in 1909 (NW flank), in1922 (SW flank and summit area), in 1954 (summit area), in 1959(E flank), and in 1982 (SW flank). In 1999 and 2000, eruptions also tookplace on the summit and SW flank (Déruelle et al., 2000; Suh et al.,2003).

The first permanent network of seismic stations to monitor MountCameroon activity was installed in 1984–1985 (Fairhead, 1985). It hasbeen reinforced by temporary seismic stations between 1984 and1987 (Ambeh et al., 1989). Due to break downs, only one sensor wasoperating during the 1999 and 2000 eruptions. However when the2000 eruption started, a temporary network of digital seismic stationswas set up tomonitor the activity of the volcano. This study focuses on

Fig. 2. Geological sketch map of Mount Cameroon modified from an unpublished geological map (Vincent, pers. com.).

208 B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

the 2000 eruptive activity of the volcano and its associated seismicityas monitored by the permanent and the temporary seismic stations.

2. Eruptive activity

The 1999 eruption took place on the SW flank of the volcano attwo sites, at 2650 m and 1400 m altitude. Activity at the lower sitewas effusive and released a 14 km long lava flow while the highersite had a strombolian activity characterized by small explosions andshort lava flows. Air France pilots first reported the resumption oferuptive activity at the summit of Mount Cameroon on May 28,2000. The Scientific Committee set up in 1999 was reactivated tomonitor the eruption and propose security and eruption disastermitigation measures to the civil protection authorities. A firstskimming by helicopter supplied by the Cameroon National AirForce on June 1st, 2000 located eruptions in the summit zone and onthe SW flank, not far from the 1999 eruptive site. We carried outfield investigations from June 3rd, 2000 and three main eruptivesites were identified (Fig. 3). The first site was located 500 msouthwest of the summit. The eruption here reactivated two cratersaligned SE–NW. The western crater at 4000 m altitude, was 10 mwide, and emitted white dust and ash for a short time. The secondcrater was located at 3975 m altitude, was 250 m in diameter and100 m high. It did not emit any lava. The eruption here was

intermittent. Paroxysmal periods were characterized by violentexplosions extruding blocks, fragments of juvenile materials, dust,ash, black smoke and gases. Heavy materials felt on and close to thevent while light material was carried by south-westerly winds.Paroxysmal periods were followed by quiet periods during whichonly white fumaroles escaped from the crater. The last explosions onthis site were reported in November 2000.

The second eruptive site was located at the flank of Fako (summit)between 3470 and 3220 m altitude. It was divided into 3 series ofvolcanic features aligned along the NNE–SSW fissures. The first serieshas about twenty craters on a distance of ~200 m between 3470 and3440m altitude. The biggest of these craters was 20m in diameter and5 m in height. The second series had 3 cones and an important lavatunnel opening to the west at 3360 m altitude. The last series had 4cones of which 3were regrouped at 3220m high. These last two serieswere connected by a fracture, onewall of which collapsed by 10m. Thesecond site produced the greatest part of lava flows in 2000, estimatedat ~6×106m3 (Suh et al., 2003). The zone of emission had aweak slopeof about 5°, and lava flows first stagnated, giving the impression fromfar of a lava lake as was reported by some scientists. Channelled by thetopography of the volcano, lava flew towards Buea and threatenedpopulations. The dynamism of the second site was mostly effusive.Explosions were weak; those observed mid-slope (~2000 m altitude)were due to the interaction between tunnelled hot lava and

Fig. 3.Map of recent lava flows onMount Cameroon on a topographicmap. I, II and III are relative to the 2000 eruptive sites. The contour lines are at 500m intervals. Also indicated arethe sites of temporary seismic stations operating in 2000 and the permanent stations falling on the same area. The Kumba station (not on the map) is located far north.

209B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

underground running water. They produced columns of black smokethat seriously threatened the population and worried local authoritiesbecause they thought that a new crater had opened close to Buea. Theactivity of the second site would have stopped on June 8, approxi-mately two weeks after the beginning of the eruption. From thatday the lava front did not move any further. Like in the 1999 eruption,both aa and pahoehoe lava were produced. The lava attained a maxi-mum length of ~6 km (Fig. 3), and stopped 4.25 km from Bokwango inBuea.

The third and last eruptive site was located at 2750 m altitudeapproximately 1 km to the east of the first 1999 eruptive site. Lavawasemitted on the first days of the eruption from an old crater, anddivided into two branches, in the direction of the SW flank. The totallength was about 1 km. The 1999 and 2000 eruptions producedhawaiites, basalts and basanites (Déruelle et al. 2000; Suh et al., 2003;Yokoyama et al., 2007).

3. Seismic data acquisition and analysis

3.1. Data acquisition — the seismic network

Following the 1982 eruption, a permanent network of shortperiod 1 Hz sensors was set up around the Mount Cameroon region

in 1984–1985 (Ambeh et al., 1989; Ateba and Ntepe, 1997). Seismicsignals from 6 well-distributed stations around the volcano weretelemetered and recorded on 24 hours ink paper drum recorders(Portacorder manufactured by Teledyne Geotech). After 10 years offunctioning, breakdowns and acts of vandalism reduced this networkto only one station (EKT), located in the premises of the observatoryin Ekona (Fig. 3). This station was used to monitor the seismicityassociated with the 1999 and 2000 eruptions. It had an Omega clocksystem that was hardly synchronized so the data cannot beintegrated with that of other networks. When the eruption startedin 2000, a temporary network of seismological stations was set upwith technical support from the French–Cameroon Cooperation,through IPGP (Paris Earth Physics Institute) and the EOST (Schooland Observatory of Earth Sciences) of Strasbourg. The temporarynetwork consisted of 3-component portable digital stations, oneMars-88 FD and 4 Geostar.

The Mars-88 FD station is an autonomous 4-channel portabledigital seismic data acquisition system, manufactured by LennartzElectronic (Germany). It was operated in trigged mode with datastored in 1 MB internal RAM and then written on two 1.44 MB floppydisks for subsequent analysis. It is linked to a 1 Hz natural frequency3 component Lennartz Electronic 3D seismometer. The Geostar sta-tion is a portable 4-channel digital seismic data acquisition system,

Fig. 4. Histogram of the bi-monthly frequency of earthquakes on the Mount Cameroonregion in year 2000. Three main periods are identified: pre-eruptive (1), co-eruptive(2) and post-eruptive (3).

Table 1P-wave velocity model

Depth Velocity

(km) (km/s)

a) This study0 3.70.2 4.22.0 4.54.2 5.010.0 6.220.0 7.9VP /VS=1.75

b) Dorbath and Dorbath (1984); Stuart et al. (1985)0 5.95.0 6.215.0 6.433.0 8.0VP /VS=1.75

210 B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

manufactured by EOST (University Louis Pasteur, Strasbourg, France).It was linked to a 2-Hz natural frequency, 3 component MarkProduct L22 sensor. The station was operated in continuous modesampling seismic signals at 75 Hz and storing the data on a 2.1 Gointernal hard disk drive. The data was retrieved through a laplinkwith an external PC. There was a possibility of radio links, in whichcase a PC in the observatory could access and download selectedparts of data stored in the distant hard disk drive located in thefield.

The installation of temporary seismic stations started on June 13,2000. The eruptive sites were surrounded by four stations (Fig. 3). Thefifth station at 2000 m altitude was the nearest to the summit. TheMars-88 FD and the Portacorder recorders were operating in theobservatory at Ekona while the four Geostar stations were spreadon the field. Two stations HUT and LEL had a radio link with theobservatory.

Since 1984, the permanent and temporary seismic stations set upin the Mount Cameroon region helped to study and characterize itslocal seismicity. Results show that one or two local events with amagnitude less than 3 are recorded every three days with a depthrange from 0 to 55 km (Ambeh et al., 1989, 1992). Periods of seismicswarms are regularly observed during which the frequency of eventsincreases and some strong events, however with a magnitude lessthan 5 are felt (Ubangoh et al., 1997; Ateba and Ntepe, 1997). Themaximum Modified Mercalli (MM) intensity observed is around VI–VIII for shallow events. From 1984 to 1994 three main seismicswarms have been recorded, on the northwest flank (Bokosso area),on the South flank (Bimbia area) and on the east flank (Tiko area).The first two swarms were located in the crust (Ambeh et al., 1992),while the third corresponds to subcrustal (down to 50–60 km)events (Ateba and Ntepe, 1997). Moreover, the summit region showsa permanent low activity. Fairhead (1985) pointed out that eruptionson Mount Cameroon are usually associated with increases of seismicactivity. This was verified in March 1999, when an eruption startedin the SW flank of the volcano following a 48-hour seismic swarmduring which about 50 earthquakes were felt in Buea and itsneighbourhood. Only five of the 50 events were recorded worldwideand the largest magnitude reported by USGS was 4.9 on March 28,1999 the day the eruption started.

3.2. Data processing and hypocenter location

The first step in processing the temporary array digital dataset isthe scanning of CD ROM archive files in order to make a referencetable of seismic events for each station. For events recorded bymore than 3 sensors, P and S wave arrival times, the maximumamplitude, the corresponding period and the coda duration are

read. The smallest reading errors are around 1/75 s (time betweentwo samples) for noise-free clear P wave onsets, and 0.1 to 0.3 s forS wave arrival times embedded in the coda of the P waves. Therecorded signals are classified into 3 categories based on their waveforms. Volcano-tectonic (VT) signals are tectonic-like earthquakes,located near the volcanoes (Latter, 1981). They have clear P and Swave onsets, but there are cases where S waves are difficult toidentify. The second category is long period (LP) signals. They haveemergent onsets of P waves but S waves are not identified. Theywould correspond to non double-couple mechanisms (Latter, 1981;Schick, 1981). The last categories, the volcanic tremor, due toharmonic vibrations observed in volcanic media, linked to fluidmovements in conduits and magmatic chambers (Chouet, 1996). Inthis study, only the first category of earthquakes was located usinga modified version of HYPOINVERSE (Klein, 1978), in whichlocations are done using different trial focal depths and altitudesof stations are taken into account. The solution retained gives thesmallest errors in hypocentral parameters and root mean square(rms) of the difference between observed and calculated traveltimes (residuals), while using the largest number of phases. Theresults depend on the 1-D velocity model of superimposed layerswhich is not well known around Mount Cameroon. In all recentstudies (Ambeh et al., 1989; Ateba and Ntepe, 1997) the velocitymodel used was the one determined by seismic refraction studiesin the Adamawa region, situated more than 500 km to thenortheast of Mount Cameroon (Dorbath and Dorbath, 1984; Stuartet al., 1985). Gravity studies suggest that the crust is thin along thecontinental part of the CVL (Poudjom et al., 1995). In the oceanicsector, Meyers and Rosendhal (1991) established that the top of theMoho is located at a maximal depth of 20 km on the Cameroon lineaxis. We tested another velocity model for which we chose thecrust to be 20 instead of 33 km thick (Table 1). The upper part(10 km) is the model used for routine localization of earthquakes atthe Piton de la Fournaise Volcano (Ferrazzini, pers. com.). Thismodel gives better results for events located in the volcano (mean

Fig. 5. A sample of a seismogram of the March 31, 2000 seismic swarm at 06:22 LT. The tic marks are at one minute intervals.

211B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

rms ~0.04 against 0.05 for the former model). Nevertheless,refraction and reflection experiments will be needed to improvethe definition of the velocity structure of the Mt. Cameroon area.Duration magnitudes were estimated using the empirical formulaestablished by Ambeh et al. (1989).

Fig. 6. Amplitude on paper chart records of

3.3. Seismic activity associated with the 2000 eruption

Seismic activity associated with the 2000 eruption of MountCameroon can be divided into three periods (Fig. 4). The pre-eruptiveperiod from March to May, marked by the March seismic swarm, the

volcanic tremor at Ekona station (EKT).

212 B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

co-eruptive period from late May to mid-September marked bysequences of seismic swarms, and the post-eruptive period aftermid-September when the seismicity returned to its background levelof one to two events per three days.

3.3.1. Pre-eruptive period — the March 2000 seismic swarmSix months after the March–April 1999 eruption, small seismic

swarms were recorded and small earthquakes were felt on the eastflank of the volcano, far from the eruptive sites. The seismicity fromJanuary to March 2000, recorded on paper charts, was at a normal(background) level of 1 to 3 earthquakes per three days until March31, when 50 events were recorded in one day. The two largestearthquakes were felt and had duration magnitudes of 3.2 and 2.7and occurred respectively at 06:22 LT and 14:46 LT (LT=GMT+1). Asample seismogram of the tectonic-like character of March 31earthquakes is shown on Fig. 5. From April 1st, the daily frequency ofearthquakes decreased and during the following week, nearly 160seismic events were recorded. The last felt earthquake of the serieswith a magnitude of 2.8, was recorded on April 7, 2000. The level oflocal seismicity remained high in April and May with sporadic eventsrecorded by the Ekona (ETK) station, with waveforms similar toearthquakes generated in the 1999 eruptive sites. We considered thislevel of activity to be ‘anomalous’ and so alerted the Governor of theMount Cameroon region (chairperson of the volcanic eruptiondisaster mitigation committee) on the abnormal activity of thevolcano.

Fig. 7. Samples of digital seismograms at Eon station. VT event (a)

Field macroseismic investigations were conducted to determinethe nature and the extent of damage, the strength of earthquakes andthe epicenters.Wenoticed that traditionalwoodhouses better resistedground shaking. On the other hand,fissureswere noticed onwalls builtwith concrete blocks. Close examination showed that some of thefissures were new while others were reactivated. Destruction of stonefoundationswas also noticed. In Ekonavillagemore than 30 requests ofdamaged houses were made. The observatory building at Ekona wasalso fissured. The fissures occurred parallel to pillars or beams. Peoplesleeping woke up and hanging objects fell. Earthquakes were felt on avery small area of 5 km radius with Ekona as the epicenter. The effectsof these earthquakes correspond to a maximumMM intensity of V–VIand may reflect the tectonic activity of the Ekona fault.

3.3.2. Co-eruptive period — the May–September 2000 seismic swarmThe co-eruptive periodwas associatedwith a seismic swarmwhich

started in late May and was characterized by sequences of increasedseismicity. The first sequence started with one week of paroxysmalphases manifested by an increase in amplitude of volcanic tremorfromMay 25 at 21:40 LT, saturating analogue paper records during 6 h(Fig. 6). Such saturations were regularly observed until May 30.Volcano-tectonic earthquakes were also recorded during this phasewith the highest daily frequency of 20 onMay 27, the day duringwhichthe inhabitants of Buea city felt some earthquakes. The highestmagnitude of the earthquakes was 4.2 on May 30th, at 19:40 LT. Afterthis phase, in June and July, episodes of volcanic tremor and volcano-

, volcanic tremor event (b) and a family of earthquakes (c, d).

213B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

tectonic events were followed by quiet periods ranging from a fewhours to two weeks. The last event felt in Buea had a magnitude of 2.8on June 10, at 17:01 LT. The first sequence, fromMay to the end of July,was characterized by a net increase of the local seismicity, feltearthquakes and episodic periods of volcanic tremor. The seismicitylevel decreased to less than 10 events during the two mid-July weeksand high amplitude volcanic tremor disappeared. Lava fountains andlava flows on the flanks were observed during the first two weeks ofthe eruption while intermittent gas explosions were observed in thesummit region.

The second sequence covers the period from August to September2000. It started with the appearance of a family of earthquakes onAugust 5, at 00:54 LT. Their waveforms were almost the same (Fig. 7),with S–P propagation time differences around 3 to 4 s at the Ekona(EKT) station where their coda duration varied between 10 and 30 s.The time interval between two successive events was several dayswhen the series started. It decreased to aminimumof 3 h onAugust 30,and started increasing again in early September. At least 30 of themwere detected and the last event of the series was recorded onSeptember 10, at 03:31 LT. During the second sequence, analoguerecordings were not saturated by the amplitude of volcanic tremor.Explosive activity was still going on in the summit area, whereintermittent black smoke and white fumaroles could still be observedfrom Buea.

From June to September, we located 64 earthquakes recorded bythe temporary array and finally kept only thosewith rms less than 0.25and horizontal and vertical errors less than 2 kmand 3 km respectively.These criteria allowed rejecting 5 locations. The first striking result isthe clustering of epicenters in the vicinity of 4.2°N and 9.2°E that isclose to the summit (Fig. 8), a regionwhere eruptionswere observed. Itcan be noticed that all magnitudes are less than 3 because the seismic

Fig. 8. Seismicity map of Mount Cameroon and volcanic cones associated with the 1999 andlava flows and seismic stations. EKT was the only permanent station working during that p

network was installed after the largest earthquakes had shaken thearea. To study the variation of hypocentral distributions with depth ofthe main cluster, two profiles 10 kmwide, and centered on the cluster,were plotted. One is along the great axis of the volcano N40°E (Fig. 9a)and the other is perpendicular to it (Fig. 9b). Both profiles depict avertical linear feature from the surface to a depth of 20 km under thesummit region. All the events of the August–September earthquakefamily are located within this feature. Although the lack of resolutionwith depth is a well-known problem in locating earthquakes, webelieve that this result likely suggests the existence of averticalmagmaconduit under the central axis of the volcano from the surface to abouta 20 km depth.

3.3.3. Post-eruptive periodThe post-eruptive period began in mid-September. During this

phase, only white fumaroles escaped from craters near the summit.No earthquakes were felt, and almost no volcanic tremor was re-corded. The number of earthquakes was decreasing with time. From15 to 30 September, 11 earthquakes were recorded. The monthlyfrequency of seismic events dropped to less than 20 in October–December 2000.

3.4. The frequency–magnitude relation

The frequency of earthquake occurrence as a function of mag-nitude is usually modeled as

log10 Nc = a−bM

This is a slightly modified form of the Gutenberg–Richter relation-ship (Gutenberg and Richter, 1954), where Nc is the cumulative

2000 eruptions. Error of epicenter location is 2 km. Also indicated are the sites of recenteriod, together with the temporary array.

Fig. 9. Vertical profiles of hypocenters along N40°E, the axis of the volcano (a); and along N130°E, the direction perpendicular to it (b). Positions of profiles are given onFig. 8.

214 B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

number of earthquakes of magnitude greater or equal to M, and aand b are constants. The constant b (known as the b-value) is theslope of the best fit line and determines the relative number oflarge versus small events. It has been used as an indicator of theheterogeneity of the material (Mogi, 1963), or of the stress regime(Scholz, 1968). Utsu (1971) pointed out that b-values are usuallyhigher for swarms than for discrete earthquakes in general but,they can also be influenced by an increase in the thermal gradient(Warren and Lathan, 1970). Data for the year 2000 of tectonicand volcano-tectonic events for the Mount Cameroon region give a

Fig. 10. The frequency–magnitude relation for year 2000 of the Mount Cameroonseismicity.

b-value of 1.43±0.02, using standard linear regression (Fig. 10). Thisvalue, which is between 1 and 3, agrees with observations done onother volcanic areas (Power et al., 1998; Iannaccone et al., 2001).The first b-values for the Mount Cameroon region reported byAmbeh et al. (1992) based on data collected between 1985 and1987 are lower (0.18–0.86) than we estimated here from the 2000dataset.

Using the 1999 dataset, Ateba et al. (in press) have deter-mined a b-value of 0.55±0.01, a rather low value. Provided thecompleteness of the seismic catalogue in 1999 and 2000, one canspeculate on the increase of the b-value from 1999 to 2000. If assuggested by Scholz (1968), the b-value is inversely proportionalto the state of stress of the source region, one can suggest thatthe 1999 eruption was associated with higher stress conditionsthan was the case in 2000. This is supported by the occurrenceof many felt events with a 4–5 magnitude in 1999 and just onemagnitude-4 event in 2000. The intrusion of magma like in Hawaii(Klein et al., 1987) or the movement of lithospheric plates assuggested by Ambeh et al. (1992) could be responsible for thisstress concentration.

3.5. The energy release

To calculate the energy or more precisely variations of energyreleased by earthquakes, we used the empirical relationship estab-lished by Richter (1958)

log10 E = 9:9 + 1:9M− 0:024M2

where E stands for the energy expressed in Ergs, and M stands forthe local magnitude. The results for the year 2000 (Fig. 11), showthat the essential part of the annual seismic energy was released inMay. From June, the monthly energy release appeared to begradual. The monthly frequency of earthquakes presents twomaxima: the first in May and the second in September. The firstcorresponds to a sharp increase in the energy release while thesecond has a more regular growth rate. This result contrasts with

Fig. 11. Seismic energy release of the Mount Cameroon area in 2000.

215B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

the 1999 observations where the quasi-totality of the annualenergy was released in some 3 days (Ateba et al., in press).

4. Discussion

During the year 2000, Mount Cameroon erupted simultaneouslyat 3 different sites. In the past, two sites at the most were involvedat the same time. The task of the monitoring program that is toforecast the time and the sites of future eruptions is then moredifficult. The 2000 eruption shows a good correlation between theseismicity and the magmatic activity but there is variability on thebehavior of the volcano because in 1999, the seismic swarmpreceding the eruption started 2 days before, while in 2000, it was2 months. If the network of sensors was installed just after the1999 eruptions, we could have been able to locate the events inorder to study the potential migration of earthquake foci towardsthe surface.

Within a short timescale of one month, the March 2000 seismicswarm could be interpreted as a tectonic event linked to the Ekonafault. But for a longer timescale, it could be seen as a premonitorysign of the impending eruption. Similarly, Hurst and McGinty (1999)analyzing earthquake swarms preceding the Mt. Ruapehu eruptionin 1995 pointed out that precursors should not be expected onlyunder active vents of volcanoes. In the past, seismic activity on theEkona fault has not been linked to any volcanic eruption. The seismicswarm preceding the 1999 eruption was strongly felt in Buea, andslightly felt in Ekona, therefore it can not be linked to the activity onthis Ekona fault.

In Hawaii, swarms of volcanic earthquakes have been associatedwith a narrow and nearly vertical pipe between 7 and 20 kmidentified as a magma conduit (Klein et al., 1987). During theeruptive period of Mount Cameroon in 2000, a swarm of earth-quakes located under the summit region between 0 and 20 km couldbe linked to a linear structure in the crust, under the central axis ofthe volcano. This could be a magma conduit, which feeds eruptionsof Mount Cameroon and through which magmatic fluids escape inthe form of volatiles. The location of a magma chamber is not yetwell determined. Its top could be at the 20 km depth or even deeperas suggested by Suh et al. (2003).

The recent 1999 and 2000 eruptive sites are distributed along therift axis of the volcano,which is the general direction of the CVL. On thesummit region, the superposition of epicenters and cones is striking.According to Zogning (1988) two axes would intersect at that level andmake it the most active volcanic site of the volcano. All these resultstend to confirm the tectonic control of eruptions on Mount Cameroon.If so, the hypothesis suggested by Meyers and Rosendahl (1991) onpreferential volcanism at crossed fracture zones along the CVL seemsto be verified in the Mount Cameroon area.

Ubangoh et al. (1997) suggested that seismic swarms or eruptionson Mount Cameroon start at the beginning or during the rainy season.The reason for this is still not clear. One speculation could be thatunderground water infiltration probably provokes sliding on faults ormass movements as was observed in Limbe in 2001 (Ayonghe et al.,2004).

Factors which can account for an increase in the b-value include therelatively small number of large events. For example, of all the eventsrecorded during the 2000 eruption, only one had a magnitude as highas 4. Other factors could be a greater heterogeneity of the mediumwhich fractured in 1999 eruption, a thermal gradient due to fumarolicactivity as observed throughout the 2000 eruption and/or, a stressregime due to the injection of new magma (e.g., Power et al., 1998).Results should be takenwith care due to poor coverage of the volcanoby seismic.

5. Conclusion

After the March/April 1999 eruption, Mount Cameroon eruptedagain in 2000 from three sites. The first site near the summit wasexplosive with no lava flow. The second site on the south flank waseffusive. Lava flew from NNE trending fissures and small cones werebuilt. The flows had a total length of about 6 km, and stopped 4.25 kmfrom Buea, twoweeks after the beginning of the eruption. The last sitewas effusive. An old cone in the southwest flank was reactivated andproduced a 1-km long lava flow that divided into two branches. Themain precursor of the 2000 eruptionmay be a seismic swarm recordedtwo months earlier on the east flank. The biggest event in the swarmhad a magnitude of 3.2, and was felt on the eastern flank at Ekonawith a MM intensity of III–VI. After the resumption of the eruptiveactivity in May, a temporary digital seismic network of 5 short period3-component sensors was installed in June, improving the monitoringof the local seismicity. The co-eruptive period from late May to mid-September was characterized by episodic sequences of earthquakeswarms and volcanic tremor. A few earthquakes were felt, the largestmagnitude recorded being 4.2. The locations show an alignment ofhypocenters just below the central axis of the volcano which could beinterpreted as a magmatic conduit linking the surface to a deep(N20 km) magma chamber. Some epicenters were also distributed onthe western and northwestern flanks. The frequency–magnituderelation for the year 2000 revealed a b-value of 1.43 which is higherthan previously calculated. Release of seismic energy increased whenthe eruption started and was gradual afterwards.

Acknowledgements

Many thanks to late J.L. Cheminée former President of WOVOwho facilitated the installation of temporary stations and a newpermanent digital seismic network around Mt. Cameroon. H. Ebonghelped for GIS maps, M. Edoa for reading the manuscript, H. Haesslerduringdata processing. Officials of the crisis committee encouragedus:N. Aboubakar, M. Mengue, P.C. Acham. Field works were carried outwith our colleagues of the scientific committee with whom wediscussed extensively about the recent activity of the volcano; to namea few: J. Nni, R. Ubangoh, S. Ayonghe, C.E. Suh, L. Ngardi, J. Lissom. Twoanonymous reviewersmade valuable suggestionswhich improved themanuscript.

216 B. Ateba et al. / Journal of Volcanology and Geothermal Research 179 (2009) 206–216

References

Aka, F.T., Nagao, K., Kusakabe, M., Sumino, H., Tanyileke, G., Ateba, B., Hell, J., 2004.Symmetrical helium isotope distribution on the Cameroon Volcanic Line, WestAfrica. Chem. Geol. 203, 205–223.

Ambeh, W.B., Fairhead, J.D., Francis, D.J., Nnange, J.M., Soba, D., 1989. Seismicity of theMount Cameroon region, West Africa. J. Afr. Earth Sci. 9, 1–7.

Ambeh, W.B., Fairhead, J.D., Stuart, G.W., 1992. Seismotectonics of the Mt. Cameroonvolcanic region, West Africa. In: Gasparini, P., Scarpa, R., Aki, K. (Eds.), VolcanicSeismology. IAVCEI Proceedings in Volcanology, vol. 3. Springer, Berlin, pp. 45–61.

Ateba, B., Ntepe, N., 1997. Post-eruptive seismic activity of Mount Cameroon (Cameroon,West Africa): a statistical analysis. J. Volcanol. Geotherm. Res. 79, 25–45.

Ateba, B., Ntepe, N., Ubangoh, R., Nni, J., Hell, J.V., Nnange, J.M., in press. La sismicitéassociée à l'éruption du Mont Cameroun de 1999: résultats préliminaires. Geocam.

Ayonghe, S.N., Ntasin, E.B., Samalang, P., Suh, C.E., 2004. The June 27, landslide onvolcanic cones in Limbe, Mount Cameroon, West Africa. J. Afr. Earth Sci. 39,435–439.

Browne, S.E., Fairhead, J.D., 1983. Gravity study of the Central African rift system:a model of continental disruption. 1: The Ngaoundere and Abu Gabra Rifts.Tectonophysics 94, 187–203.

Burke, K.C., Dessauvagie, T.F.J., Whiteman, A.J., 1971. The opening of the Gulf of Guineaand the geological history of the Benue Trough and Niger Delta. Nature 233, 51–55.

Chouet, B.A., 1996. Long-period volcano seismicity: its source and use in eruptionforecasting. Nature 380, 309–316.

Cornacchia, M., Dars, R., 1983. Un trait structural majeur du continent africain. Leslinéaments centrafricains du Cameroun au Golfe d'Aden. Bull. Soc. Geol. Fr. 25,101–109.

Déruelle, B., Moreau, C., Nkonguin, N.E., 1983. Sur la récente éruption du MontCameroun (16 octobre–12 novembre 1982). C.R. Acad. Sci. Paris 296, 807–812.

Déruelle, B., N'ni, J., Kambou, R., 1987. Mount Cameroon: an active volcano of theCameroon Line. J. Afr. Earth Sci. 6, 197–214.

Déruelle, B., Bardintzeff, J.M., Cheminée, J.L., Ngounouno, I., Lissom, J., Nkoumbou, C.,Etamé, J., Hell, J.V., Tanyileke, G., Nni, J., Ateba, B., Ntepe, N., Nono, A., Wandji, P.,Fosso, J., Nkouathio, D.G., 2000. Eruptions simultanées de basalte alcalin et dehawaiite au Mont Cameroun (28 mars–17 avril 1999). C.R. Acad. Sci. Paris 331,525–531.

Dorbath, C., Dorbath, L., 1984. Approche sismologique de la structure de la lithosphèreen Afrique de l'Ouest. Thèse Doctorat d'État. Université Paris VI.

Fairhead, J.D., 1985. Preliminary study of the seismicity associated with the Cameroonvolcanic province during the volcanic eruption of Mount Cameroon in 1982. J. Afr.Earth Sci. 3, 197–301.

Fitton, J.G., 1980. The Benue trough and the Cameroon line — a migrating rift system inWest Africa. Earth Planet. Sci. Lett. 51, 132–138.

Fitton, J.G., Kilburn, C.R.J., Jhirwall, M.F., Hughes, D.J., 1983. 1982 eruption ofMt Cameroon, West Africa. Nature 306, 327–332.

Gèze, B., 1943. Géographie physique et géologie du Cameroun occidental. Mém.Museum Nation. Hist. Natur. Nouv. Sér. 17, 1–272.

Gèze, B., 1953. Les volcans du Cameroun Occidental. Bull. Volc. 13, 63–92.Gouhier, J., Nougier, J., Nougier, D.,1974. Contribution à l'étude volcanique du Cameroun:

Ligne du Cameroun - Adamaoua. Ann. Fac. Sci. Yaoundé 17, 3–48.Gutenberg, B., Richter, C.F., 1954. Seismicity of the earth and associated phenomena.

Princetown Univ. Press, Princetown.Halliday, A.N., Dickin, A.P., Fallick, A.E., Fitton, J.G., 1988. Mantle dynamics: a Nd, Sr, Pb

and O isotopic study of the Cameroon Line volcanic chain. J. Petrol. 29, 181–211.Hedberg J.D., 1968. A geological analysis of the Cameroon trend. Ph.D. Thesis, Princeton

Univ., Princeton.Hurst, A.W., McGinty, P.J., 1999. Earthquake swarms to the west of Mt Ruapehu

preceding its 1995 eruption. J. Volcanol. Geotherm. Res. 90, 19–28.Iannaccone, G., Alessio, G., Borriello, G., Cusano, P., Petrosino, S., Ricciolino, P., Talarico,

G., Torello, V., 2001. Characteristics of the seismicity of Vesuvius and Campi Flegreiduring the year 2000. Ann. Geofis 44, 1075–1091.

Klein F.W., 1978. Hypocenter location program HYPOINVERSE Part 1: users guide toversions 1, 2, 3 and 4, U.S. Geological Survey Open-file Rep. 78-694, 113 pp.

Klein, F.W., Koyanagi, R.Y., Tanigawa, W.R., 1987. The seismicity of the Kilauea's magmasystem. In: Decker, R.W., Wright, T., Stauffer, P.H. (Eds.), Volcanism in Hawaii. USGeol. Surv. Prof. Paper, vol. 1350, pp. 1019–1185.

Latter, J.H., 1981. Volcanic earthquakes and their relationship to eruptions in Ruapehuand N'gauruhoe volcanoes. J. Volcanol. Geotherm. Res. 9, 293–309.

Lee, D.C., Halliday, A.N., Fitton, J.G., Poli, G., 1994. Isotopic variations with distance andtime in the volcanic islands of the Cameroon line: evidence for a mantle plumeorigin. Earth Planet. Sci. Lett. 123, 119–138.

Marzoli, A., Piccirillo, E.M., Renne, P.R., Bellieni, G., Iacumin, M., Nyobe, J.B., Aka, F.T.,2000. The Cameroon Volcanic Line revisited: petrogenesis of continental basalticmagmas from lithospheric and asthenospheric mantle sources. J. Petrol. 41, 87–109.

Meyers, J.B., Rosendahl, B.R., 1991. Seismic reflection character of the Cameroon volcanicline: evidence for uplifted oceanic crust. Geology 19, 1072–1076.

Mogi, K., 1963. Some discussions on aftershocks, foreshocks and earthquake swarms.The fracture of semi-infinite body caused by an inner stress origin and its relation toearthquake phenomena, 3. Bull. Earthq. Res. Inst. Univ. Tokyo 41, 615–658.

Moreau, C., Regnoult, J.M., Déruelle, B., Robineau, B., 1987. A new tectonic model for theCameroon Line, Central Africa. Tectonophysics 141, 317–334.

Ngako, V., Affaton, P., Nnange, J.M., Njanko, Th., 2003. Pan-African tectonic evolution incentral and southern Cameroon: transpression and transtension during sinistralshear movements. J. Afr. Earth Sci. 36, 207–214.

Nkoumbou, C., Déruelle, B., Velde, D., 1995. Petrology of Mt Etinde nephlinite series.J. Petrol. 36, 373–395.

Piper, J.D.A., Richardson, A., 1972. The paleomagnetism of the Gulf of Guinea volcanicprovince, West Africa. Géophys. J.R. Astrom. Soc. 29, 147–171.

Poudjom, D.Y.H., Nnange, J.M., Diament, M., Ebinger, C.J., 1995. Effective elastic thicknessand the crustal thickness variations in the west central Africa inferred from gravitydata. J. Geophys. Res. 100, 22047–22070.

Power, J.A., Wyss, M., Latchman, J.L., 1998. Spatial variations in the frequency–magnitudedistribution of earthquakes at Soufriere Hills Volcano, Montserrat, West Indies.Geophys. Res. Lett. 25, 3653–3656.

Richter, C.F., 1958. Elementary seismology. W.H. Freeman and Co., San Francisco.Schick, R., 1981. Source mechanism of volcanic earthquakes. Bull. Volcanol. 44, 491–497.Scholz, C., 1968. The frequency–magnitude relation of microfracturing in rock and its

relation to earthquakes. Bull. Seismol. Soc. Am. 58, 339–415.Sibuet, J.C., Mascle, J., 1978. Plate kinematic implications of Atlantic Equatorial fracture

zone trends. J. Geophys. Res. 83, 3401–3421.Stuart, G.W., Fairhead, J.D., Dorbath, L., Dorbath, C., 1985. A seismic refraction study of

the crustal structure associated with the Adamawa plateau and the Garoua Rift,Cameroun, West Africa. Geophys. J. R. Astr. Soc. 81, 1–12.

Suh, C.E., Sparks, R.S., Fitton, J.G., Ayonghe, S.N., Annen, C., Nana, R., Luckman, A., 2003.The 1999 and 2000 eruptions of Mount Cameroon: eruption behavior andpetrochemistry of lava. Bull. Volcanol. 65, 267–281.

Ubangoh, R.U., Ateba, B., Ayonghe, S.N., Ekodeck, G.E., 1997. Earthquake swarms ofMount Cameroon, West Africa. J. Afr. Earth Sci. 24, 413–424.

Utsu, T., 1971. Aftershocks and earthquake statistics III. J. Fac. Sci., Hokkaido Univ.series VII 3, 379–441.

Vincent, P., 1971. Données nouvelles sur le volcan Mont Cameroun. Sixth Coll. on Afr.Geol., Leicester, Abstract.

Warren, N.W., Latham, G.V., 1970. An experimental study of thermally inducedmicrofracturing and its relation to volcanic activity. J. Geophys. Res. 75, 4455–4464.

Yokoyama, T., Aka, F.T., Kusakabe, M., Nakamura, E., 2007. Geochim. Cosmochim. Acta 71,1835–1854.

Zogning A., 1988. Le Mont Cameroun, un volcan actif: contribution à l'étude degéographie physique appliquée. Thèse Doc. 3C, Univ. Yaoundé, 447 pp.