This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy

Journal of Volcanology and Geothermal Research 182 (2009) 23–33

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research

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

Characterising unrest during the reawakening of the central volcanic complex on Tenerife, Canary Islands, 2004–2005, and implications for assessing hazards and risk mitigation

J. Martí a,⁎, R. Ortiz b, J. Gottsmann c, A. Garcia b, S. De La Cruz-Reyna d a Institute of Earth Sciences “Jaume Almera”, CSIC, Lluís Solé Sabarís s/n, Barcelona 08028, Spain b Department of Volcanology, Museo Nacional de Ciencias Naturales, CSIC, C/ José Gutiérrez Abascal, 2, 28006 Madrid, Spain c Department of Earth Sciences, University of Bristol, Wills Memorial Building, Queens Road, Bristol BS8 1RJ, United Kingdom d Instituto de Geofísica, Universidad Nacional Autónoma de México, Ciudad Universitaria, México 04510 D.F., México article info abstract

Article history: Increased onshore seismic activity in April 2004 marked the first documented renewal of tectonic unrest on Received 13 June 2007 Tenerife, Canary Islands, Spain, since the island's last volcanic eruption in 1909. Events included tremors, felt Accepted 30 January 2009 earthquakes and the occasional emission of a visible gas plume from the central 3718 m high Teide volcano, Available online 15 February 2009 and an increased diffuse emission of CO2. Here, we evaluate results from seismic and microgravimetric observations in addition to other available data obtained between April 2004 and July 2005, in order to shed Keywords: light on the source of these events. We discuss the information to assess whether collectively the phenomena Tenerife “ ” central complex qualify to be termed volcanic unrest , and the socio-economic implications of the phenomena, and critically fi unrest examine the ensuing scienti c response. We also evaluate the potential volcanic-eruption precursory volcano monitoring character of the data. Suggestions for the establishment of improved volcano monitoring programmes, early hazard assessment warning systems and civil response protocols for volcanic crises on Tenerife are proposed. © 2009 Elsevier B.V. All rights reserved.

1. Introduction thermal system may equally produce ground deformation and seismi- city (Sturtevant et al., 1996; Bianco et al., 2004; Tikku et al., 2006; Episodes of unrest are inherent components of the lifecycle of a Gottsmann et al., 2007). As a consequence, the beginning of unrest at a volcano. In a number of recent cases including Soufrière Hills (Sparks dormant volcano presents investigators with the intrinsic dilemma as to and Young, 2002) and Mt. St. Helens (Endo et al., 1981), unrest whether the unrest will culminate in an eruptive phase, hence posing a preceded eruptions and must hence be seen as an important eruption direct threat to life and property around a volcano, or whether the precursor (Sandri et al., 2004). However, there are also examples of “unusual” behaviour will eventually dissipate, causing little disruption unrest waning-off after months or years of restlessness without any to communities and hence little socio-economic damage. This uncer- eruptive volcanic activity. Most prominent cases include the volcanic tainty is even more challenging and extends to the basic question of how calderas of Long Valley (Battaglia et al., 2003a,b) and the Campi to define unrest when the dormant volcano has not erupted in historical Flegrei (Dvorak, and Berrino, 1991). times or has not previously shown signs of unrest, either witnessed by a Volcanic unrest is the manifestation of complex sub-surface local population or recorded by a monitoring network. This situation processes leading to detectable signals at the ground surface. Processes was dramatically illustrated, for example, by the eruptions of El Chichón such as magma migration and emplacement, tectonic and hydrothermal (1982) and Popocatépetl, (1994) (México) (De La Cruz-Reyna and activity can trigger seismicity, ground deformation, thermal variations Tilling, 2007), Pinatubo (Philippines) (1991) (Newhall and Punongba- and changes in the potential fields around a volcano. Seismicity and yan, 1996), or Soufrière Hills (Montserrat) (1995–present) (Kilburn and ground deformation may be induced by brittle failure of surrounding Voight, 1998; Sparks and Young, 2002). rocks due to the pressure increase accompanying the replenishment of The Spanish volcanological community was confronted with this magma reservoirs and the exsolution of a gas phase, a process regarded dilemma in early 2004, during a period of increased seismic activity as a key trigger for volcanic eruptions (Murphy et al., 1998). and manifestations of activity that was potentially volcanic began on Alternatively, volume/pressure increases within a sub-surface hydro- the volcanic island after a quiescence of almost 100 years. The lack of both previous data and a volcano monitoring network, created a particular situation in which the available information was assessed ⁎ Corresponding author. and interpreted based on comparison to other volcanic systems, not E-mail address: [email protected] (J. Martí). necessarily to Tenerife, and on the expertise of the scientists involved

0377-0273/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jvolgeores.2009.01.028 Author's personal copy

24 J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 in the investigation. However, there was no general consensus on observed phenomena by various authors with respect to the abnormal either the volcanic evolution of Tenerife or on its current state of activity on Tenerife. activity to use as a basis for judging the future outlook. This gave rise to a rather confusing situation in which some scientists defined the 2. Volcanological background information anomalous behaviour as volcanic unrest (García et al., 2006; Gottsmann et al., 2006), while others claimed that there was no The Canary Islands form a volcanic archipelago with a long-standing clear indication for unrest (Carracedo and Troll, 2006; Carracedo et al., history of volcanic activity that began more than 40 million years ago 2006). This controversy, which is in part due to the lack of knowledge (Araña and Ortiz, 1991; Anguita and Hernan, 2000). More than a dozen on the Tenerife volcanology, was particularly dramatic in the light of eruptions have occurred on the islands of Tenerife, Lanzarote, and La a possible reawakening of the Teide-Pico Viejo volcanic complex, as Palma since the 16th century. Tenerife, the largest of the Canary Islands, there is clear disagreement concerning whether or not this is still has an eruptive history of over 12 million years including a shield an active volcanic complex (Ablay and Marti, 2000; Carracedo building phase followed by the construction of a central volcanic et al., 2003; García et al., 2006; Carracedo et al., 2007; Marti et al., structure, the Las Cañadas edifice (Marti et al., 1994)(Fig. 1). The 2008). volcanic evolution of Tenerife comprises both constructive and Unfortunately, the situation experienced in Tenerife is not ex- destructive phases including vertical and lateral collapses on the order clusive to this region, but has occurred previously in other volcanic of several km3 (Marti et al., 1997). At least three vertical collapses areas (e.g., the Guadeloupe crisis in 1976, Tazieff, 1979) and it will resulted in the formation of the 16 km-wide Las Cañadas caldera, into most certainly occur again somewhere else in the future. In an attempt which the prominent Teide-Pico Viejo volcanic complex was emplaced to address the hazards and risk implications of a poorly known during predominantly effusive and also occasional explosive activity volcanic system, we analyse the particular case of Tenerife. We present over the past 170–190 ka (Marti et al., 1994; Marti and Gudmundsson, data collected prior to and (with an emphasis) during the crisis from 2000; Ablay and Marti, 2000). This complex appears to be fed by both early 2004 to late 2005 from geophysical investigations including shallow-level (b5 km) phonolitic magma reservoirs and deeper-seated seismic, gravimetric and geodetic observations, and discuss whether basaltic magma patches (Ablay et al.,1998; Ablay and Marti, 2000; Martí or not these data allow to suggest a change in the behaviour of et al., 2008). Recent (b0.5 ka) volcanic activity was located on the Teide- volcanic system on Tenerife, i.e. the occurrence of volcanic unrest. We Pico Viejo complex (explosive and effusive phonolitic eruptions) as well also discuss historic volcanic activity on Tenerife in the light of theses as along a NW–SE and NE–SW oriented extensional structural new investigations and examine whether or not Teide should still be lineaments, referred to as the Santiago del Teide Rift and the Dorsal considered an active volcano. Finally, we discuss implications for risk rift, respectively (dominantly monogenetic mafic eruptions) (Fig. 1). mitigation on the island given the different interpretations of the Historic eruptions from the Teide-Pico Viejo complex and the rifts

Fig. 1. Simplified geological map of Tenerife (after Ablay and Marti, 2000) indicating the location of the IGN seismic station CCAN and the benchmarks used in microgravimetic surveys. T, Teide; PV, Pico Viejo; MB, Montaña Blanca; RG, Roques de García; G, Guajara; SRZ, Santiago del Teide rift zone; DRZ, Dorsal rift zone; SVZ, Southern volcanic zone. Vents: black symbols: mafic and intermediate vents; white symbols, felsic vents; stars: historic and sub-historic; circles: other vents; Other symbols: solid square, CCAN; open squares, benchmarks. Author's personal copy

J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 25

Fig. 2. Seismic network of the Canary Islands. White symbols: seismic stations installed before 2004. Black symbols: seismic stations installed after May 2004. Diamonds: Short- period seismic stations from CSIC. Circles: Short-period seismic stations from IGN. Squares: Broad Band seismic stations from IGN. complex were of basaltic composition and occurred in 1704, 1706, 1798, monitoring network (Fig. 2) can be obtained on-line from the official and 1909. seismic catalogue of the National Geographic Institute (IGN) at www. ign.es. This catalogue has been used as the main data source in this 3. Background volcanic activity and chronology of events (April paper. Earthquake epicentres of background seismicity clustered in 2004–July 2005) two off-shore areas located to the north and to the south of Tenerife (Figs. 2–5). These (deep) earthquakes were most likely associated The background level of volcanic activity on Tenerife has been with tectonic movement along crustal heterogeneities. Onshore manifested as weak fumarolic activity, with average temperatures of seismicity also occurred, but was much less frequent. around 86 °C on the summit of Teide, as well as gas emissions along a “Unusual” manifestations commenced in early 2004, when a series spur known as the Roques del Garcia, which divides the Las Cañadas of earthquakes marked a change from the usual “background” activity caldera into a western and an eastern sector (Hernández et al., 1998, on Tenerife. This phenomena was unusual insofar, as: (i) the epi- 2000; Galindo, 2005)(Fig. 1). In addition, there were diffuse gas centres were located mainly onland (Fig. 4), (ii) the number of events emissions (CO2, H) above background level (as defined by average was significantly higher than during background activity (Fig. 5), values elsewhere on the island) around the Teide crater, the caldera (iii) most of the events where located at shallow depth and (iv) four border and the rift zones (Hernández et al., 1998, 2000; Galindo, earthquakes were felt, representing the first account of such phe- 2005). Epicentres location of seismic events recorded by the regional nomena since the last eruption on the island in 1909. During 2004

Fig. 3. Seismicity in and around Canary Island from 1993 to 2005. Data from the public catalogue of the Instituto Geográfico Nacional (IGN, www.ign.es). The activity is concentrated around Tenerife and after 2004 in the North-West sector of the island. Author's personal copy

26 J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33

Fig. 4. Evolution of seismicity in and around Tenerife from 1993 to 2005. Data from the public catalogue of the Instituto Geográfico Nacional (IGN, www.ign.es). The high level of activity appear in the first moths of 2004. epicentres propagated with time from the northern and north- 4. Scientific responses to unrest and data analysis western part of the island (Fig. 6) southward, towards the Teie-Pico Viejo stratovolcanoes into the westernmost depression (Llano de Seismic monitoring has been performed on Tenerife for ca. 20 years Ucanca) of the Las Canadas caldera. Inland seismicity continued by the National Geographical Institute of Spain (IGN), with a 7 station throughout 2005, 2006, and 2007 albeit at a lower rate (249 in 2004, seismic network in the Canary Islands originally designed to monitor 207 in 2005, 138 in 2006 and 141 in 2007) up to the time of this regional tectonic seismicity, with similar characteristics to those of the submission. National Seismic Network deployed in the mainland. At present, the Simultaneous with the seismic events, an increase in the flux of increment in the monitoring tasks conducted by several national and

CO2 in diffuse degassing was detected along the Santiago del Teide Rift local institutions (IGN, CSIC, UCA, ITER) have resulted in a significant (Galindo, 2005). A fumarolic plume at the summit of Teide (3718 m a. increase of the number of seismic stations on Tenerife (Fig. 2), as well m.s.l) was visible to the naked eye for a few hours on the 20 October as the installation of permanent GPS and magnetic networks. Gas 2004, in addition to an increase in the seismic noise (tremor) at Teide monitoring has been conducted since 1997 by the Tenerife-based (García et al., 2006). On the 5 December 2004, a new fracture (few Institute of Technology and Renewable Energies (ITER) (www.iter.es). tens of meters long) appeared in the Orotava valley, with associated With the advent of unusual seismic activity, a scientific discussion gas emission. It was preceded nine days before by a significant ensued as to whether Tenerife is preparing for a renewal of eruptive increase in low frequency seismic energy events and associated with volcanic activity. This was mirrored by a public and political debate as to some seismic event in the area (1, 8, and 9 December, 2004) (García the associated risks and the consequences for the socio-economic et al., 2006). stability of the region. The lack of an integrated monitoring database

Fig. 5. Number of seismic events recorded from 1999 and 2005 in and around Tenerife at the CCAN (IGN) reference seismic station, unchanged from 1992. The figure shows the high increment of activity in April–May 2004. Author's personal copy

J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 27

Fig. 6. Located seismic events during 2004 in the northwest sector of Tenerife (IGN 2004 catalogue). The epicentres show a moderate migration from north to south during this period. for Tenerife had important consequences for regional hazard assessment, An additional complexity in the scientific response to the emerging risk mitigation, scenario planning and the development of response unrest was the existence of three different levels of political protocols: administration (federal, regional, local) with different responsibilities for civil protection and consequently differences in the response to a i.) The assessment of the status quo of the volcanic system was potential volcanic crisis, and an unclear line of command. difficult as crucial background data was either missing or not An initial response to the developments in early 2004 was the available. installation of a joint deformation and micro-gravity network (Gottsmann ii.) Changes in geodynamic processes were hence difficult to et al., 2006), the deployment of additional seismic stations by CSIC and IGN interpret in terms of the degree of deviation from background and an increase in the frequency of gas sampling by ITER. Other research behaviour. projects carried out by the CSIC and supported by Spanish Ministry of iii.) The reliability of existing data, though of crucial importance, Education and Science, started one year later (2005) in order to investigate was difficult to assess. the origin and temporal evolution of the “anomalous” situation. The

Fig. 7. Characteristic waveform and its spectrogram of the tremor recorded on Tenerife: examples from 23 May 2004. Note the two frequencies of the tremor. Tremor was recorded at, CSIC short period seismic station, BODE, placed at the northern side of Tenerife in the Icod Valley. Author's personal copy

28 J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33

Fig. 8. During 2004 some seismic events preceded tremor by some minutes. The plot shows a one hour waveform and its spectrogram on 7th of October 2004. Arrow S indicates the occurrence of the seismic event and arrow T the tremor. motivation to respond on a scientific level originated from the indication deep (several tens of km depth) seismicity generally concentrated that epicentre locations appeared to migrate from the northern coastline offshore, both to the north and to the south of the island. This pattern towards the central Teide-Pico Viejo complex. Combining these observa- of seimicity has been traditionally attributed to a regional tectonic tions with results from petrological investigations (Ablay et al., 1998; origin (Mezcua et al., 1992). Activity has been continuously monitored Triebold et al., 2006; Andujar, 2007), which indicate the existence of on Tenerife since 1992 with an IGN permanent seismic station, CCAN, shallow phonolitic magma reservoirs beneath the complex, evoked the an analogue short period seismometer with UHF telemetry located in possibility that migrating magma or hot fluids might disrupt the ther- the Las Cañadas caldera. Another two short period seismic stations on modynamic balance of these reservoirs. Early warning of mass and density Tenerife, located in the Güimar Valley (S) and Anaga (NE) together changes at depth coupled with seismic information on source location and with CCAN and few other stations on the remaining islands comprised evolution were therefore seen as essential means to assess the potential the regional IGN seismic network. for hazards associated with the unusual situation on the island. In 2000, IGN replaced some of their seismic stations with broad band instruments while at the same time expanding the existing 4.1. Seismicity network from 7 seismometers to a total of 12. By 2004 the IGN operated a total of three broad-band and five new short period seismic Prior to the increase in seismic activity in 2004, most of the stations. Before, during and after network amendments, station CCAN seismicity recorded on and around Tenerife corresponded to relatively acted as the control reference (Fig. 2). As a result of the denser network

Fig. 9. Example of a seismogram from CCAN showing the three components recorded on the 29th of August 2004. The amplitude of P and S waves indicates that the incidence is nearly horizontal, i.e. the waves come from shallow depths. This fact could be also results from a particular relations between the seismic station and the focal mechanism of the seismic event. Anyway, the time difference between the arrivals of P and S waves is of the order of 1 s (shown by the double arrow). This suggest a depth of less than 7 km even in the case that the seismic event was located exactly below the station. Author's personal copy

J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 29

Fig. 10. Time interval histograms showing the different statistical behaviour of two populations of earthquakes: tectonic occurring offshore and those mainly concentrating onland. Two radii of 25 km and 75 km, respectively, have been fixed arbitrarily around Teide, although the results do not show significant variations between 20 and 30 km for the inner limit and between 60 and 80 km for the outer. 60% of the seismic events occurring onland are clustered in the following 24 h, while the earthquakes from offshore occur isolated except for the appearance of pairs (two seismic events in the same zone within a few hours of each other), which contribute to the value of 25% in the next 24 h. with higher precision instruments, an increase in seismicity was (Fig. 4). In addition, a lineament of epicentre locations oriented NW–SE detected along the entire Canarian Archipelago and also on and around was clearly detectable, joining the two offshore zones by crossing the Tenerife. There, in addition to the previously known seismo-tectonic La Orotava and Güimar valleys on the island (Fig. 4). zone located offshore to the southeast of the island, a previously Since April 2004, several seismic swarms occurred (including four undetected seismo-tectonic zone (offshore) to the north was revealed felt earthquakes with magnitudes of 2.5 to 3.5 Mw) and continued to

Fig. 11. Gutenberg-Richter b values of the seismic events recorded on Tenerife for the period 2004–2005. Dots indicate earthquakes located at less than 25 km, diamonds represent earthquakes located between 25 and 75 km. The histograms showing the magnitude distribution of earthquakes located at less than 25 km (left) and between 25 and 75 km (right), respectively, are used to define the minimum magnitude in the compute the b value. The b-histograms show the bimodal character of the distribution of the seismic event near Teide volcano, and unimodal for the seismicity around the island. Author's personal copy

30 J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 occur at the time of writing, though with less intensity. This increase those from between 25 and 75 km (Fig. 11). The b-values obtained in the number of seismic events is evident if we only consider the clearly indicate that by October 2004 seismicity changed from an records from the CCAN station, which has remained invariable since earlier, mainly tectonic origin to bi-modal b-value seismicity, largely 1992 (Fig. 5). This activity included the well-known deep offshore influenced by a preponderance of volcano-seismic events (Fig. 11). tectonic seismicity as well as previously known seismic sources whose epicentres concentrated mostly onland. These new events included 4.2. Time-lapse microgravity and ground deformation volcanotectonic and purely volcanic signals such as harmonic tremor and few long period events (Tárraga et al., 2006; Almendros et al., As a further response to the unusual activity the first joint ground 2007; Tárraga, 2007)(Figs. 7 and 8). Compared to the tectonic seis- deformation/microgravity network was installed on the island in May micity, this new volcano-tectonic and volcanic activity was charac- 2004, approximately two weeks after the start of increased seismicity terised by shallow-seated (few km depth) hypocentres, as shown in (Gottsmann et al., 2006). Ground deformation data allow the quanti- Fig. 9. It is important to note that the new seismicity was initially fication of sub-surface volume changes which, combined with subsur- located beneath the northern and north-western flanks of the Teide- face mass variations obtained from micro-gravity data, enables an Pico Viejo complex along the western portion of the Icod valley but assessment to be made of the nature of the source causing the unrest, be subsequently migrated towards the interior of the island along the it magma, hydrothermal fluids or a mixture of both. Prior to 2004 a western sector of Las Cañadas caldera. comprehensive network around the Teide-Pico Viejo complex was It is also worth mentioning that the statistical behaviour of the two lacking. The new network aimed at providing i) important data on the populations of earthquakes (off-shore tectonic as opposed to onshore sub-surface dynamics and ii) critical baseline data for future volcanic and volcano-tectonic) is markedly different (Fig. 10). There developments. appears to be a relationship between the offshore and onshore seismicity The network consists of 15 benchmarks, which are positioned to in the sense that the tectonic and volcanic seismic events, respectively, provide coverage of the central volcanic system, including the Teide- show predictive patterns (Tárraga et al., 2006; Tárraga, 2007). Pico Viejo complex, the Las Cañadas caldera, as well as the Santiago Finally, many onshore events are followed by periods of tremor Rift (Fig. 12). The first reoccupation of the network was performed in that in some cases exceed one hour in duration. (Carniel et al., 2005; July 2004, followed by campaigns in April 2005 and July 2005. Tárraga et al., 2006; Vila et al., 2006; Carniel et al., 2008). Detailed Detailed descriptions of the network and resulting data are given in analysis of seismic signals show that N45% of the events recorded Gottsmann et al. (2006). Here, we summarise the key findings. during the period of maximum activity (from July to September of Gravity changes across the area under investigation were smallest 2004), were related to a tremor-type seismic signal (Tárraga et al., in the central and eastern depression of Las Cañadas caldera, where 2006). Moreover, the study of the system memory from the seismic cumulative changes over the 14-month period where only slightly signal (Carniel et al., 2008) shows the presence of long duration higher than the instrument precision level (±0.015 mGal on average; tremor periods associated with the occurrence of seismic events. In 1 mGal=10 µm/s2) average. A marked gravity anomaly with a general, the low amplitude of this signal makes it difficult to detect maximum gravity increase of around 0.4 mGal was found in the directly from the seismogram (Carniel et al., 2008) Also, the existence of some LP events has been reported in the area of Pico Viejo (Almendros et al., 2007).

4.1.1. Temporal variation of the Gutenberg-Richter b parameter In order to demonstrate the difference between purely tectonic and volcanic or volcano-tectonic seismicity we calculated the Gutenberg-Richter b-value from the IGN seismic catalogue consider- ing the data available from 2000. The Gutenberg-Richter law states that the distribution of seismic events of a certain magnitude occurs in a particular region during a particular time interval corresponding to: log ðÞN = b4M + a where N is the cumulative number of events exceeding a given magnitude M, and a, b are constants. The b-value strongly depends on the properties of the seismic medium and on the nature of the stress regime in which the earthquakes occur. Tectonic areas where seismicity is caused by regional stresses usually show b-values near 1.0 (Frohlich and Davis, 1993). Stress concentration leading to clustering of seismicity and fracture over distances of a few kilometers seems to have a strong effect on the value of b (Ogata and Katsura, 1993; Wiemer and Mc Nutt, 1997; Wiemer and Wyss, 1997, Novelo- Casanovaa et al., 2006). We have used the maximum likelihood method to calculate the coefficient b by means of the following equation:

ðÞ log e Fig. 12. Residual gravity changes between May 2004 and July 2005, corrected for the b = b N − M Mmin effect of ground deformation and water table changes (from Gottsmann et al., 2006). Residual changes are draped over a DEM of the central volcanic complex (CVC) of where bMN is the average magnitude of 20 consecutive events and Tenerife. Uncertainty in gravity residuals are on average ±0.015 mGal (1 mGal=10 µm/s2). Stars represent epicentres of seismic events recorded between Mmin is the minimum magnitude that the monitoring network can May 2004 and July 2005. Both gravity increase and seismicity appear to be spatially and detect in the whole area (Fig. 10). We deduce Mmin from the temporally correlated. Locations of boreholes from which water table data was used for histograms of events located at less than 25 km from Teide and data reduction are shown as white circles. See Gottsmann et al. (2006) for details. Author's personal copy

J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 31 northwestern part of the covered area. With time, this anomaly clearly indicates events related to volcanic processes became a appeared to “migrate” in a southward direction, reaching the western major feature of seismicity from October 2004 onwards. part of Las Cañadas caldera between July 2004 and April 2005 c) Thus, starting from early 2005 onwards, onshore seismicity (Fig. 12). The gravity increase along the Santiago del Teide Rift, noted appears to correspond to a combination of tectonic and volcanic between the first two campaigns, had disappeared by April 2005. At processes. However, offshore seismicity kept nearly constant with the same time, though, gravity, increased significantly along the values of b≈1. The limited detection capability of the seismic northern slopes of Teide, adding to the impression of a spatio- network at these distances and the relatively small number of temporal migration of the causative source (Fig. 12). It is thus events would explain the observed variations. pertinent that in this same vicinity on 5 December 2004 a new fissure d) The pronounced increase in b values for close events (less than with fumarole emission appeared in the Orotava valley (see below). A 25 km from Teide) occurred at the end of October 2004, coinciding gas plume emanating from the summit fumaroles of Teide was with the increase of the fumarolic activity observed on 20 October. particularly noticeable during October 2004 (García et al., 2006), e) The fact that some of the tremor episodes can be predicted by the between surveys 2 and 3. In summary, significant gravity changes analysis of purely tectonic signals suggests that there is some occurred mainly along the northern flanks of the Teide-Pico Viejo coupling of the Tenerife volcanic or hydrothermal system to the complex and along a ca. 3 km wide zone along the western side of the regional tectonic activity. volcanic complex extending into the westernmost sector of the Las f) The evaluation of all seismic data presented here is exclusively Cañadas caldera between May 2004 and July 2005 (Fig. 12). At the based on data recorded by seismic station (CCAN) that has been same time ground deformation over the investigation period operating for more than 20 years. Therefore, the differences remained largely within measurement errors (2.5 cm vertical). The observed between the seismicity in the interval considered in our significant perturbation of the gravity field was hence not induced by study and the previous background period, including a significant widespread ground deformation but rather by residual sub-surface increase in the number of the total events recorded, cannot be mass/density changes. attributed to an increased sensitivity of the seismic network since 2000. 4.3. Other observations g) Then, there is the perturbation of the gravity field, which occurred simultaneously with the elevation in seismicity. Both spatial and In addition to the increase in seismicity and the microgravity temporal variations were detected indicating residual mass changes described above other signs of anomalous behaviour on changes beneath the PV–PT complex and extending into the Tenerife were also reported. These include a significant increase in the western sector of the Las Cañadas caldera over the 14 months diffuse emission of carbon dioxide along the Santiago del Teide Ridge observation time. The observed changes are most likely due to and around the Las Cañadas caldera (Galindo, 2005; Pérez et al., shallow (few kms depth) hydrothermal fluid migrations and not

2005), an increase in the CO2 flux of the Teide fumaroles, the exclusively due to magma movement (Gottsmann et al., 2006). appearance of traces of SO2 in the Teide fumaroles (M. Martini, personal communication), an increase in the fumarolic emission from Moreover, two other observations made during the studied period Teide (García et al., 2006), and the opening of a new fissure with gas have to be added to the previous list: the increase of fumarolic activity emission in La Orotava valley (García et al., 2006). at Teide's crater on the morning of 20 October 2004, and the opening of a new fracture with gas emission in the Benijos area on the 5 5. Discussion December 2004. One of the authors of this paper (JM) witnessed the increased The data and information presented in the previous sections are in fumarolic activity at Teide crater from the highest point (Guajara our opinion sufficient and reliable enough to regard the unusual peak) of the Las Cañadas caldera wall. He observed a pulsating (with a behaviour starting in spring 2004 as unequivocal evidence of volcanic period of approximately half a minute) emission of a dilute, low unrest (see below). However, the existence of unrest has been strictly column of steam at the western side of the crater. This column formed ruled out by Carracedo and Troll (2006) and Carracedo et al. (2006), in the absence of any weather cloud or other atmospheric phenomena who suggest that Tenerife has never abandoned its equilibrium state and was rapidly dispersed by westerly winds. The phenomenon was and argue that the manifestations of geophysical activity can be clearly observed from 8.30 h to 10.30 h, local time, when clouds attributed to changes in the sensitivity of the monitoring devices and/ started to appear and masked the steam emission from the crater. or misinterpretation of the observations made. While a diversity of These characteristics do not accord with any meteorological phenom- opinions motivates good scientific debate, in the particular case of enon usually observed on Teide (e.g., el sombrerito del Teide). Also, Tenerife (and perhaps in fact in all cases of the reawakening of the atmospheric soundings made by the Tenerife meteorological volcanoes after significant repose periods) discrepancies in the station during October 2004 do not show any significant variation in interpretation of such signals can have important socio-economic the atmospheric conditions that could justify any anomalous implications. It is therefore of paramount importance to clarify the meteorological events on top of Teide that day. Moreover, the facts on which any scientific interpretation is based. characteristics of the seismic tremor changed in the period during Regarding whether or not volcanic unrest occurred or is which the fumarolic event was observed (Fig. 13). We therefore perhaps still occurring, the data presented in this paper strongly propose that this episode of increased fumarolic activity was support the existence of a change from the “normal” background associated with the disturbance of the background activity of the volcanic activity on Tenerife. In this regard the following facts are Teide-Pico Viejo complex. important: Anomalous gas emission was detected along a new fracture that opened in the area of Benijos in the Orotava valley on the 5 December a) The significant increase in seismicity as well as the location of a 2004. Carracedo and Troll (2006) proposed that this gas emission was considerable number of earthquakes onshore with hypocentres at the residual vapour emissions of a cheese factory located in the shallow depths, contrasts with the previous seismicity pattern vicinity, while ITER reported the presence of a magmatic component characterised by deep-seated earthquakes mostly located offshore. in the Benijos gas emission (as published in the local magazine “El b) The presence of volcanic tremor and long period signals suggest Baleo”, Feb. 2005, 25, 9–11). The Benijos' fracture is superimposed on that, at least, part of the new seismicity is non-tectonic in origin. the tectonic lineament defined by epicentre locations between the This is also confirmed by the Gutenberg-Richter b parameter that Guimar and Orotava valleys, which suggests a tectonic control for the Author's personal copy

32 J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 fracture location. In this context it is perhaps important to note that one eruption per 100 years. No historical records exist on eruptions the gas emission appeared between two earthquakes occurring on 29 from the central phonolitic system, but the available geochronological November at La Guancha and 9 December, northwest of Puerto de la data (Carracedo et al., 2003, 2007) suggest inter-eruption intervals for Cruz. These locations mark the extreme ends of the La Orotava fracture the Teide-Pico Viejo system of the order of hundreds of years to one to system. The tremor signal started to increase in intensity on 29 two thousand years, having occurred the last eruption about November and disappeared on the day of the Benijos gas emission 1000 years ago (Carracedo et al., 2003, 2007). (García et al., 2006). An important yet perhaps neglected topic relates to coupling In summary, we propose that there is enough evidence to support between deep basaltic injections and shallow phonolitic reservoirs the argument for an episode of volcanic unrest in Tenerife starting in (Marti et al., 2008). Although not yet fully investigated and spring 2004. Whether this unrest is precursory to an eruptive process understood, the existing petrological data indicate that the mixing cannot be unambigiously answered here. Deep intrusion of magma of the two magmas is a consistent signature of the Teide system into the Santiago del Teide rift and a subsequent migration of (Ablay et al., 1998; Ablay and Marti, 2000; Triebold et al., 2006; hydrothermal fluids have been proposed to explain the main changes Andujar, 2007). We have to conclude that basaltic magma migration observed with different monitoring methods (Gottsmann et al., 2006; and injection can be regarded as a main trigger of phonolitic Almendros et al., 2007). The probability for an individual intrusion of volcanism on Tenerife. As such, unrest episodes triggered by the magma to reach the surface and to cause an eruption is generally very deep intrusion and ascent of basaltic magmas should be assessed in low (Gudmundsson et al., 1999), and as such unrest episodes cannot, the context of the possible (re)-activation of the phonolitic Teide- on their own, be regarded as unequivocal evidence for an impending Pico Viejo magmatic system. Therefore, care is required in identify- eruption. However, unrest should remind people that the region is ing and interpreting unrest signals at a reawakening volcano, such volcanically active and that measures to mitigate risks have to be as the Teide-Pico Viejo complex, where reliable information from considered. Ignoring unrest as a possible sign of incipient active previous unrest and eruption episodes is lacking and knowledge volcanism leads to the underestimation of risk from volcanic activity. on the evolution and dynamics of its magmatic system is poorly In fact, simply dismissing such signs inevitably increases the risk to life constrained. by curtailing preparedness and reduces recognition of threat to property, giving rise to a false perception of safety in the region. 6. Conclusions For the particular case of Tenerife there is the added complexity of the volcanic systems compared to other active volcanic areas where A comprehensive federal volcano monitoring programme, compar- information on background volcanic activity is more abundant and able to those of other European nations with active volcanic areas such enables a better identification of unrest phenomena. In addition to a as Portugal, Italy or France, is absent in the Canary Islands including complete lack of monitoring data from previous unrest and/or Tenerife. eruptive episodes, there is still little knowledge on the properties The societal impact of a renewal of volcanic activity on Tenerife and dynamics of Tenerife's active volcanic system. There is certainly may be severe. Tenerife has a population of ca. 1 million, increasing the possibility for a new eruption along the rift zones or the central significantly during the summer months due to tourism, which volcanic complex. Eruptions over the last 5000 years (Carracedo et al., represents the major economic asset on the island. Population and 2003, 2007; Marti et al., 2008) have occurred from both settings, with tourist centres are predominantly spread along the southern and a higher frequency in the former. Basaltic eruptions through the active northern coastlines, but also around the southern and western flanks rift zones have occurred in historical times with an average of about of the island. The island also hosts two major airports representing vital lifelines for prosperity on the island. Signs of volcanic unrest or for the renewal of eruptive activity on Tenerife need to be assessed in the light of their potential effects on socio-economic structures in the area. A centrally administered database including complementary geophysical, geological, geodetic, and geochemical investigations, which could contribute to a comprehensive assessment of the current state of the volcanic system, is absent. Such a data pool administered, for example by a governmental body, be it on a regional or federal level, should be a prerequisite for effective volcano monitoring, reliable hazard assessment and risk mitigation purposes. Two important lessons that have been learned during the recent episode of unrest on Tenerife should be considered:

1) Tenerife, as with the rest of the Canary Islands, is an active volcanic area and as such its volcanism requires scientific, public and governmental attention. 2) The effective reduction of volcanic risk requires i) the development of mitigation programmes including a significant improvement of the knowledge on the volcanological and geological evolution of the island, ii) a designated volcano monitoring programme capable of detect changes in the activity of the volcanic system, iii) public education programmes on the issue of living in an active volcanic environment, iv) a territorial infrastructure planning programme Fig. 13. Variations of parameters from Teide Information Seismic System before and based on the knowledge and assessment of present volcanic risk, and after the appearance of fumarolic emissions at the summit crater of Teide on 20 October v) effective emergency and crisis management plans. The role of the 2004. Each plot corresponds to a daily lower envelope of all 60-minute Power Spectral scientific community is crucial in helping all members of the society Density curves from the 16 to the 22 October 2004. The fumarole emission was to appreciate the strengths and weaknesses of scientific insights, as preceeded by a significant increase in seismic energy for spectral components above 1 Hz and followed by a predominant frequency of 2.4 Hz, with a return to normal levels well as the opportunities and threats of living in an active volcanic within a few hours after the emission. area. In giving scientific advice it is therefore advisable to critically Author's personal copy

J. Martí et al. / Journal of Volcanology and Geothermal Research 182 (2009) 23–33 33

and objectively review all available data, and not just present or Frohlich, C., Davis, S., 1993. Teleseismic b-values: or, much ado about 1.0. Journal of Geophysical Research 98, 631–644. discuss selective data taken out of the relevant context. Galindo, I., 2005. Estructura volcano-tectónica y emisión difusa de gases de Tenerife (Islas Canarias). PhD Thesis, University of Barcelona, 350 pp. Acknowledgments García, A., Vila, J., Ortiz, R., Macià, R., Sleeman, R., Marrero, J.M., Sánchez, N., Tárraga, M., Correig, A.M., 2006. Monitoring the reawakening of Canary Island's Teide volcano. EOS Trans. AGU., 87 (6), 61, 65 This research has been funded by the EC EXPLORIS (EVR1-2001- Gottsmann, J., Wooller, L., Marti, Fernandez, J., Camacho, A., González, P., García, A., 00047) and MEC TEGETEIDE (CGL2004-21643-E) projects. JM is grateful Rymer, H., 2006. New evidence for the reawakening of Teide volcano. Geophysical for the MEC grant PR-2006-0499. JG acknowledges support from an MEC Research Letters 33, L20311. doi:10.1029/2006GL027523. Gottsmann, J., Carniel, R., Coppo, N., Wooller, L., Hautmann, S., Rymer, H., 2007. “Ramon y Cajal” grant, a Royal Society University Research Fellowship Oscillations in hydrothermal systems as a source of periodic unrest at caldera and a Royal Society International Joint Project Grant. We thank the volcanoes: multiparameter insights from Nisyros, Greece. Geophysical Research Instituto Geográfico Nacional (IGN) for allowing us to use their data Letters 34, L07307. doi:10.1029/2007GL029594. Gudmundsson, A., Marinoni, L., Martí, J.,1999. Injection and arrest of dykes: implications from the seismic catalogue and in particular Carmen López and María for volcanic hazards. Journal of Volcanology and Geothermal Research 88, 1–13. José Blanco (IGN) for their helpful discussions on the seismic activity in Hernández, P.A., Pérez, N., Salazar, J.M.L., Nakai, S., Notsu, K., Wakita, H., 1998. Diffuse Tenerife. We thank Willy Aspinall and an anonymous reviewer for their emission of carbon dioxide, methane and helium-3 from Teide volcano, Tenerife, Canary Islands. Geophysical Research Letters 25 (17), 3311–3314. helpful and constructive reviews. Hernández, P.A., Pérez, N., Salazar, J.M.L., Sato, M., Notsu, K., Wakita, H., 2000. Soil gas

CO2,CH2, and H2 distribution in and around Las Cañadas caldera, Tenerife, Canary Islands, Spain. Journal of Volcanology and Geothermal Research 103, 425–438. References Kilburn, C.R.J., Voight, B., 1998. Slow rock fracture as eruption precursor at Soufriere Hills volcano, Montserrat. Geophysical Research Letters 98GL01609, 25,19, p. 3665. Ablay, G.J., Marti, J., 2000. Stratigraphy, structure, and volcanic evolution of the Pico Marti, J., Gudmundsson, A., 2000. The Las Cañadas caldera (Tenerife, Canary Islands): an Teide-Pico Viejo formation, Tenerife, Canary Islands. Journal of Volcanology and overlapping collapse calderas generated by magma chamber migration. Journal of – Geothermal Research 103, 175–208. Volcanology and Geothermal Research 103, 161 173. Ablay, G.J., Carroll, M.R., Palmer, M.R., Martí, J., Sparks, R.S.J., 1998. Basanite–phonolite Martí, J., Mitjavila, J., Araña, V.,1994. Stratigraphy, structure and geochronology of the Las – lineages of the Teide-Pico Viejo volcanic complex, Tenerife, Canary Islands. Journal Cañadas caldera (Tenerife, Canary Islands). Geological Magazine 131 (6), 715 727. of Petrology 39, 905–936. Marti, J., Hurlimann, M., Ablay, G.J., Gudmundsson, A., 1997. Vertical and lateral – Almendros, A., Ibañez, J.M., Carmona, E., Zandomeneghi, D., 2007. Array análisis of collapses in Tenerife and other ocean volcanic islands. Geology 25, 879 882. volcanic earthquakes and tremor recorded at Las Cañadas caldera (Tenerife Island, Martí, J., Geyer, A., Andujar, J., Teixido, F., Costa, F., 2008. Assessing the potential for Spain) during the 2004 seismic activation of Teide volcano. Journal of Volcanology future explosive activity from Teide-Pico Viejo stratovolcanoes (Tenerife, Canary – and Geothermal Research 160, 285–299. Islands). Journal of Volcanology and Geothermal Research 178 (6), 529 542. Andujar, J., 2007. Application of experimental petrology to the characterisation of Mezcua, J., Buforn, E., Udias, A., Rueda, J., 1992. Seismotectonics of the Canary Islands. – phonolitic magmas from Tenerife, Canary Islands. PhD Thesis, University of Barcelona, Tectonophysics 208, 447 452. Spain, 191 pp. Murphy, M.D., Sparks, R.S.J., Barclay, J., Carroll, M.R., Lejeune, A.-M., Brewer, T.S., Anguita, F., Hernan, F., 2000. The Canary Islands origin: a unifying model. Journal of Macdonald, R., Black, S., Young, S., 1998. The role of magma mixing in triggering the Volcanology and Geothermal Research 103, 1–26. current eruption at the Soufriere Hills volcano, Montserrat, West Indies. Geophysical – Araña, V., Ortiz, R., 1991. The Canary Islands: Tectonics, magmatism and geodynamic Research Letters 25, 3433 3436. framework. In: Kampunzu, A.B., Lubala, R.T. (Eds.), Magmatism in Extensional Structural Newhall, C.G., Punongbayan, R.S. (Eds.), 1996. Fire and mud: eruptions and lahars of Settings: the Phanerozoic African Plate,. Springer-Verlag, New York, pp. 209–249. mount Pinatubo, Philippines. University of Washington Press. 1126 pp. Battaglia, M., Segall, P., Murray, J., Cervelli, P., Langbein, J., 2003a. The mechanics of unrest Novelo-Casanovaa, D.A., Martínez-Bringasb, A., Valdés-Gonzáleza, C., 2006. Temporal – at Long Valley caldera, California; 1. modeling the geometry of the source using GPS, variations of Qc-1 and b-values associated to the December 2000 January 2001 leveling and two-color EDM data. In: Sorey, M.L., McConnell, V.S., Roeloffs, E. (Eds.), volcanic activity at the Popocatepetl volcano, Mexico. Journal of Volcanology and – Crustal unrest in Long Valley Caldera, California; new interpretations from geo- Geothermal Research 152, 347 358. physical and hydrologic monitoring and deep drilling. Journal of Volcanology and Ogata, Y., Katsura, K., 1993. Analysis of temporal and spatial heterogeneity of – Geothermal Research, vol. 127, pp. 195–217. magnitude frequency distribution inferred from earthquake catalogues. Geophy- – Battaglia, M., Segall, P., Roberts, C., 2003b. The mechanics of unrest at Long Valley sical Journal International 113, 727 738. caldera, California; 2. constraining the nature of the source using geodetic and Pérez, N.M., Hernández, P.A., Melián, G., Galindo, I., Padrón, E., Nolasco, D., Salazar, P., micro-gravity data. In: Sorey, M.L., McConnell, V.S., Roeloffs, E. (Eds.), Crustal unrest Pérez, V., Marrero, R., Gonzalez, P., Coello, C., Barrancos, J., Gonzalez, J., 2005. in Long Valley Caldera, California; new interpretations from geophysical and Premonitory geochemical and geophysical signatures of volcanic unrest at Tenerife, hydrologic monitoring and deep drilling. Journal of Volcanology and Geothermal Canary Islands. European Geophysical Union, Geophysical Research Abstracts, Research, vol. 127, pp. 219–245. abstract n: EGU05-A-09993. Bianco, F., Del Pezzo, E., Saccorotti, G., Ventura, G., 2004. The role of hydrothermal fluids Sandri, L., Marzocchi, W., Zaccarelli, L., 2004. A new perspective in identifying the – in triggering the July–August 2000 seismic swarm at Campi Flegrei, Italy: evidence precursory patterns of eruptions. Bulletin of Volcanology 66, 263 275. from seismological and mesostructural data. Journal of Volcanology and Geother- Sparks, R.S.J., Young, S.R., 2002. The eruption of Soufrière Hills Volcano, Montserrat – fi mal Research 133, 229–246. (1995 1999): overview of scienti c results Geological Society, London. Memoirs 21, – Carniel, R., Tárraga, M., Ortiz, R., Marrero, J.M., Garcia, A., 2005. Interactions between 45 69. fi volcano-tectonic events and volcanic tremor at Teide-Pico Viejo volcanic complex, Sturtevant, B., Kanamori, H., Brodsky, E.E., 1996. Seismic triggering by recti ed diffusion Canary Islands. European Geosciences Union (EGU) General Assembly 2005. in geothermal systems. Journal of Geophysical Research 101, 25,269. Geophysical Research Abstracts 7, 03366. Tárraga, M., 2007. Análisis dinámicos de series sismo-volcánicas. Estudio de los Carniel, R., Tárraga, M., Jaquet, O., Ortiz, R., García, A., 2008. The seismic noise at Las volcanes Villarrica, Tungurahua, Stromboli y Teide. PhD Thesis, Universidad Cañadas caldera, Tenerife, Spain. Persistence characterization, and possible Complutense de Madrid, Spain, 205 pp. relationship with regional tectonic events. Journal of Volcanology and Geothermal Tárraga, M., Carniel, R., Ortiz, R., Marrero, J.M., García, A., 2006. On the predictability of Research 173, 157–164. volcano-tectonic events by low frequency seismic noise analysis at Teide-Pico Viejo Carracedo, J.C., Troll, V.R., 2006. Seismicity and gas emissions on Tenerife: a real cause volcanic complex, Canary Islands. Natural Hazards and Earth Systems Science 6, – for alarm? Geology Toda 22, 138–141. 365 376. Carracedo, J.C., Paterne, M., Guillou, H., Pérez-Torrado, F.J., Paris, R., Rodríguez-Badiola, Tazieff, H., 1979. What is to be forecast: outbreak of eruption or possible paroxysm? The – E., Hansen, A., 2003. Dataciones radiométricas (14C y K/Ar) del Teide y el rift example of the Guadeloupe Soufrière. Journal of the Geological Society 136, 327 329. noroeste, Tenerife, Islas Canarias. Estudios Geológicos 59, 15–29. Tikku, A.A., McAdoo, D.C., Schenewerk, M.S., Willoughby, E.C., 2006. Temporal fl Carracedo, J.C., Troll, V.R., Pérez-Torrado, F., 2006. Recent unrest at Canary Islands' Teide uctuations of microseismic noise in Yellowstone's Upper Geyser Basin from a volcano. EOS Trans. AGU., 87 (43), 462, 465. continuous gravity observation. Geophysical Research Letters 33, L11306. Carracedo, J.C., Rodríguez-Badiola, E., Guillou, H., Paterne, M., Scaillet, S., Pérez-Torrado, doi:10.1029/2006GL026113. fi F.J., Paris, R., Fra-Paleo, U., Hansen, A., 2007. Eruptive and structural history of Teide Triebold, S., Kronz, A., Wörner, G., 2006. Anorthite-calibrated backscattered electron pro les, volcano and rift zones of Tenerife, Canary Islands. Geological Society of America trace elements, and growth textures in feldspards from Teide Pico-Viejo volcanic – Bulletin 119, 1027–1051. complex (Tenerife). Journal of Volcanology and Geothermal Research 154, 117 130. De La Cruz-Reyna, S., Tilling, R.I., 2007. Scientific and public responses to the ongoing Vila, J., Macia, R., Kumar, D., Ortiz, R., Moreno, H., Correig, A.M., 2006. Analysis of the volcanic crisis at Popocatépetl volcano, Mexico: importance of an effective hazard- unrest of active volcanoes using variations of the base level noise seismic spectrum. – warning system. Journal of Volcanology and Geothermal Research 170, 121–134. Journal of Volcanology and Geothermal Research 128, 11 20. – Dvorak, J.J., Berrino, G., 1991. Recent ground movement and seismic activity in Campi Wiemer, S., Mc Nutt, S.R., 1997. Variations in frequency magnitude distribution with Flegrei, southern Italy: Episodic growth of a resurgent dome. Journal of Geophysical depth in two volcanic areas: Mount St. Helens, Washington, and Mt. Spurr, Alaska. – Research 96 (B2), 2309–2323. Geophysical Research Letters 24, 189 192. Endo, E.T., Malone, S.D., Noson, L.L., Weaver, C.S., 1981. Locations, magnitudes, and Wiemer, S., Wyss, M., 1997. Mapping the frequency magnitude distribution in aspe- statistics of the March 20–May 18 earthquake sequence. U.S. Geological Survey rities: an improved technique to calculate recurrence times? Journal of Geophysical – Professional Paper 1250, 93–107. Research 102, 15115 15128.