Introduction

Located in the remote reaches of northern Chad, Africa, the Tibesti massif (Fig. 1a, b) has long been neglected by researchers due to the region’s relative inaccessibility and history of political instability. Though largely absent from scientific inquiry for several decades, it is arguably one of the world’s most significant examples of intracontinental volcanism. As much of our understanding of mantle hot spots has been obtained from observations of active volcanoes located in oceanic settings, the Tibesti massif and its related hot spot volcanics afford a valuable opportunity to explore the dynamics and properties of continental hot spot volcanism.

Fig. 1
figure 1

a Location of the Tibesti massif (black) and the Tibesti Volcanic Province (grey), Chad, northern Africa; b International Space Station oblique photo of the Tibesti massif (Photo ID: ISS002-E-7327; 11 June 2001). The TVP can be seen in the central to lower portions of the massif (outlined with dashed line). The oblique nature of the photograph is such that the surface area increases to the lower right (south) within the figure. The northwest portion of the massif lies outside the photograph frame. (Image courtesy Earth Sciences and Image Analysis, NASA-Johnson Space Center. 27 June 2003)

Full size image

In the early 20th Century, the region that includes the Tibesti massif was a contentious territory among the colonial powers of Africa. Due in part to the prospect of significant uranium deposits (Goodell 1992), conflict intensified in 1973 between the independent countries of Libya and Chad. Only in recent years has this 20-year-old border dispute been resolved by international agreement (Cohen 1994). Travel to the region is nonetheless problematic; estimates indicate that perhaps tens of thousands of land mines (International Campaign to Ban Landmines 2001) were planted in the Borkou–Ennedi–Tibesti area of northern Chad, and subsequently abandoned, before fighting was brought to an end. Landmine reports pertaining to Chad (Survey Action Center 2002) indicate that the Tibesti region remains most heavily contaminated with both mines and unexploded ordinance.

The most scientifically fruitful expeditions to the Tibesti massif were conducted between 1920 and 1970 by several teams, notably Tilho (1920), Gèze et al. (1959), Grove (1960), and Vincent (1970). Soon after, however, contest for control of the region made further geological research nearly impossible. General sketches of the geology and petrology of the Tibesti were provided by Gèze et al. (1959) and Vincent (1970). Malin (1977) investigated satellite images of large-scale Martian volcanism as analogues for the volcanism within the Tibesti massif. Apart from the geochemical and petrological work conducted by Gourgaud and Vincent (2004) on Emi Koussi (southeast Tibesti massif), we are unaware of any research published since these early works.

Although ground-based access to the Tibesti massif remains challenging, advances in satellite remote sensing do provide a basis for a modern synoptic survey of the volcanology of the region. Our aim here, in particular, is to exploit the sophisticated sensing capabilities of the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) launched in December 1999 as part of NASA’s Earth Observing System (EOS). This instrument provides not only multispectral imagery spanning the visible and infrared electromagnetic bands (with multiple channels in the shortwave and thermal infrared regions of the spectrum) but also permits retrieval of precision digital elevation data via its stereo imaging capability. A satellite-based analysis such as this is greatly enhanced by the excellent preservation of geologic features due to the aridity of the region. We evaluate aspects of the Tibesti massif in the broader context of global hot spot volcanism. This systematic survey is also intended to provide a coherent framework of the Tibesti volcanism for any future studies of the region, including reconnaissance-scale field research.

Much of the toponymy of the Tibesti massif is derived from the Arabic and Teda-daza languages, and is used throughout the region. The term ehi refers to peaks or rocky hills, emi to larger mountains or mountainous regions, and tarso to high plateaux or gently-sloping mountainsides. Standard catalogues of terrestrial volcanoes (e.g. Simkin and Siebert 1994) list very few of the most identifiable Tibesti structures (limited solely to Emi Koussi, Tarso Toussidé, and Tarso Voon). In some cases, more than one place name may be found in the existing literature for the same feature within the Tibesti (e.g. Gèze et al. 1959; Grove 1960; Vincent 1970); for the sake of clarity and consistency, we choose here to use the most unambiguous and recognisable names. Additionally, we use here the terms dome and plateaux to refer to general topographic relief, and intend no reference to genetic or structural constraints unless otherwise indicated.

Location and geological setting

The Tibesti massif (see Fig. 1b) extends from approximately 19 to 23° N latitude and 16 to 19° E longitude (area ∼100,000 km2), and rises up to a maximum elevation of 3,394 km above the Saharan desert. For reference, the East African Rift system is located 1,900 km to the east, and the Cameroon hot spot is 1,800 km to the southeast. Though the political borders of the region have shifted repeatedly in the past, all of the volcanic portions of the Tibesti massif are now located within northernmost Chad, in the Borkou–Ennedi–Tibesti province.

The Tibesti massif is one of six major exposures of Precambrian crystalline rocks found in northern Africa. It consists of a core of intrusive and metamorphic rocks surrounded by Palaeozoic and younger sedimentary sequences, all of which are partly capped by Tertiary volcanic rocks (Ghuma and Rogers 1978). The ages of the basement rocks are not well known, but were radiometrically dated by Vachette (1964) as between 500–600 Myr. The Tibesti is separated from other massifs to the east and west by deep basins of Palaeozoic and younger sedimentary cover. To the north, it is bounded by a northward-thickening wedge of Palaeozoic, Mesozoic, and Tertiary sedimentary rocks. Southward, the massif connects with the older Precambrian terranes of central Africa, which are overlain by thin, mostly post-Palaeozoic sequences (Ghuma and Rogers 1978). From previous field campaign observations, the TVP has been generally assumed to be comprised of basalts, basanites, dacites, and youthful tephra deposits and ignimbrite sheets.

The term ‘hot spot’, in reference to the Tibesti region, reflects the broad concept as described by Morgan (1972) and Crough (1978) as a region of intra-plate or anomalous ridge crest volcanism that is either persistent or accompanied by a broad topographic swell. These characteristics have been noted for various oceanic swells and linear tracks of oceanic volcanism as partial evidence of upwelling plumes. Hot spots are typically associated with long-lived active volcanism, and are characterized by the advection of high heat flux to the earth’s surface (Smith and Braile 1994). Mantle plumes are interpreted to begin as instabilities near one of two major physical boundaries within the mantle: the upper/lower mantle boundary (∼670-km deep) and the core-mantle boundary (∼2,900 km depth) (Perfit and Davidson 2000). Low-density material travels buoyantly from depth through the mantle, displacing denser aesthenosphere, and ultimately heating the base of the lithosphere. This process ostensibly results in regional topographic uplift, lithospheric thinning, and magmatism. The mantle plume hypothesis is now widely accepted to explain the presence of hot spot volcanoes, but direct evidence for actual plumes is relatively weak; recent studies (Zhao 2001, 2004) have better explored the issues of delimiting and characterising mantle plumes via seismic modelling, and the locations of presumed hot spots have been correlated with a rise in the gravitational potential of the region (Perfit and Davidson 2000). Models for the distribution of volcanism in the context of plumes diverge, however. For example, White and McKenzie (1995) envisaged penetration of the lithosphere by numerous individual mantle plumes, while Ebinger and Sleep (1998) proposed that sub-lithospheric channelling of magma from a single, giant plume can laterally feed geographically discrete volcanic provinces (e.g. the Afar plume). The nature of hot spot volcanism in northern Africa has been comparatively little studied with respect to many other hot spot provinces.

Regional faults striking NNE–SSW follow Precambrian trends throughout the entire Tibesti region (El Makhrouf 1988), but are generally obscured within the TVP by the volcanic products. Malin (1977) recognised these faults as well, and suggested that subsequent uplift of the regional basement along a NNW axis was followed by the overall tilting to the NNE. Furon (1963) suggested that the northeast–southwest volcanic alignments of the northern Tibesti massif are an extension of those of the Cameroon Trough to the southwest, on the western continental margin of Africa. Guiraud et al. (2000) provided evidence for a 6,000 km, northwest–southeast striking lineament in northern Africa, extending from the volcanism of the East African Rift system, through the Tibesti massif, to the northwest coast of the African continent. The relationships between this lineament and the volcanism of the Tibesti, if any, have not yet been determined.

Centroid-Moment Tensor (CMT) solutions from the Harvard University Seismology group suggest that the Tibesti region has been effectively aseismic, at least above Mw ∼5.5, since data have been recorded in northern Africa from 1976 (cf. Dziewonski et al. 1981).

Our preliminary inspection of the EGM96 spherical harmonic model of the Earth’s gravitational potential indicates a significant positive gravity anomaly beneath the TVP, though wavelength limitations of current gravity model data for the region limit interpretations of sub-surface lithosphere–aesthenosphere interactions.

Remote sensing and data processing

Due to the aridity of the Sahara, the Tibesti massif is nearly free of accumulated vegetal soil, and robust plant life is deficient in almost every area throughout the year. Conditions highly favour exposure of the region’s geological features, particularly when viewed using remote sensing data. The utility of remote sensing techniques in studying volcanic regions on Earth is well documented (e.g. de Silva and Francis 1990; Okada and Ishii 1993; Kahle et al. 1995; Oppenheimer 1998; Ramsey et al. 2004; Wiart and Oppenheimer 2005), and, of course, research on planetary volcanism has depended almost exclusively on remote sensing (e.g. Malin 1977; Mouginis-Mark and Robinson 1992; Davies et al. 2001). Both ground-based and spaceborne remote sensing data provide a unique perspective on large- and small-scale volcanic processes. With careful analysis and interpretation of spectral reflectance, image texture, and topographical data, rich palaeovolcanological, lithological and related information can be extracted.

Our primary source of remote sensing data used in this study is the multispectral Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER; Table 1). ASTER is an imaging instrument carried on Terra, a satellite launched in December 1999 as part of NASA’s Earth Observing System (EOS). The characteristics and capabilities of the ASTER instrument have been discussed elsewhere in detail (e.g. Kahle et al. 1991; Yamaguchi et al. 1998; Pieri and Abrams 2004), and are summarised here in Table 2. The multispectral data were digitally processed using the ENVI (v. 3.6 and 4.0) imaging software (Research Systems, Inc.) on a PC platform. Each ASTER scene containing volcanic features was treated using well-established image processing techniques, all designed to enhance the spectral, spatial, and textural characteristics of the imagery. In total, we have utilized 51 individual scenes (Table 3) to cover the entire volcanic portion of the Tibesti massif in Chad (one scene covers approximately 60 × 60 km).

Table 1 Spectral wavelength resolution characteristics of ASTER
Full size table
Table 2 ASTER instrument capabilities
Full size table
Table 3 Selected ASTER satellite imagery covering the Tibesti Volcanic Province, northern Chad
Full size table

Digital elevation models (DEMs) were produced for each of the volcanic features by making use of the stereo capabilities of the ASTER instrument. Separate sensors exist for ASTER band 3, each viewing the Earth’s surface at different look-angles (nadir and aft) to allow stereoscopic analyses. Much of the description and analysis presented within this investigation was based on study of interactive perspective views of the Tibesti volcanism in various user-defined combinations of spectral channels (Fig. 2).

Fig. 2
figure 2

Digital elevation model (DEM) of the Emi Koussi volcano (q.v.), eastern Tibesti, with the corresponding ASTER bands 3, 2, and 1 (as R, G, B) draped on top (total relief = 2,894 m; vertical exaggeration = 4×). The DEM was created using the nadir- and backward-looking capabilities of the ASTER imaging system for stereoscopic viewing. In plan view, the dimensions of this scene are approximately 60 × 60 km. Image ID pg-PR1B0000-2002012902_105_001; bands 3,2,1 as R,G,B

Full size image

In addition to its high spatial resolution, the ASTER instrument was deliberately designed for enhanced spectral resolution in the short-wave and thermal infrared channels (SWIR and TIR, respectively) with geologists in mind; these additional spectral bands facilitate better discrimination of mineralogy and lithology in satellite imagery. The wide range of wavelengths covered by ASTER (see Table 1) allows good quality distinction between iron oxide minerals, clay-bearing minerals, sulphate minerals, ammonia minerals, siliceous rocks, and carbonates (e.g. Kahle and Goetz 1983). Other researchers (e.g. Abrams and Hook 1995; Hubbard et al. 2003) have used simulated ASTER data to produce geologic maps of unaltered and hydrothermally altered volcanic areas that are as accurate as published maps made by traditional field methods.

In this study, we have relied solely upon the use of several combinations of ASTER spectral channels to help resolve basic lithologies of the TVP (translated to R G, B for image processing). Fundamental distinctions between tephra deposits, basalts, and sandstone regions of the Tibesti region were made possible by the use of decorrelation contrast stretches within the TIR wavelengths of the ASTER imagery, and were compared on a first-order basis to existing, yet limited, studies of the Tibesti lithology (e.g., Vincent 1970; Gourgaud and Vincent 2004) as ad hoc field data. Stretching the spectral data by this process produces spectral differences between surface units to be displayed as colour differences in R, G, B, while most of the temperature variation prevalent in thermal wavelengths is displayed as differences in intensity. We recognize that limitations abound in detailing the Tibesti volcanic lithologies, however, due to an obvious paucity of ground-based sampling or comparison. For this reason, we do not present extensive lithologic mapping of the Tibesti massif within this particular study.

Important supplementary databases used were Landsat Thematic Mapper (TM) images, and 70-mm astronaut photography from the Space Shuttle (STS087-717-075, 19 November–5 December 1997; STS102-717-60, 19 March 2001; STS108-701-008, 5–17 December 2001; Earth Sciences and Image Analysis 2003). The latter provides natural colour images at variable illuminations and look angles with a nominal ground resolution of ∼20 m under optimal circumstances.

Volcanoes of the Tibesti massif

Though the areal extent of the Tibesti massif is some 100,000 km2, most of the volcanism is confined to a region in the southern section, which we here refer to as the Tibesti Volcanic Province (TVP), covering approximately one-third of the massif’s total area. Loosely defined geographical boundaries within the TVP provide a basis for dividing the TVP into three sub-provinces: the Eastern, Central, and Western regions; this classification does not suggest genetic or structural significance, and is used within this study for sake of ease. Detailed and extensive descriptions of the primary TVP volcanic structures, including three-dimensional animations, can be found online as Electronic supplementary material. A summary of the characteristics of the volcanoes and associated features of the Tibesti can be found in Table 4. For the general location map of the main volcanic centres and other features, refer to Fig. 3.

Table 4 Characteristics of the prominent volcanic features within the Tibesti Volcanic Province, northern Chad
Full size table
Fig. 3
figure 3

Principal features and characteristics of the Tibesti Volcanic Province, Tibesti Massif, northern Chad

Full size image

Evolution of the Tibesti volcanic province

Concrete evidence of the chronology of Tibestian eruptive activity is very sparse, though Quaternary volcanic ash deposits were emplaced into standing bodies of water (e.g. Trou au Natron) where diatoms flourished (Furon 1963). Field samples from the Tibesti with any useful locality information have proved equally difficult to acquire. At present, the most productive means for determining the volcanic history of the Tibesti lies with interpretation of the satellite imagery. Due to the areal extent of the TVP, age relations amongst the volcanoes are difficult to discern; in the absence of quantitative fieldwork, however, remote sensing imagery does shed some light on the sequence of events. Given the nature of the available data, we stress that the age relationships described for the volcanic features are based here on superposition and embayment relationships of the volcanic units currently exposed at the surface, which are assumed to be predominantly the youngest materials associated with each feature.

Some relative dating information and other basics have been established over many decades, however. The TVP volcanics directly overlie Precambrian basement granites, diorites, and crystalline schists (Gèze et al. 1959). The basement rock is, in turn, completely surrounded and overlapped by flat-lying sediments of the Sahara Desert (Cahen et al. 1984). The intense activity of the TVP began perhaps as early as the Oligocene, though the major products that mark its surface occurred during the Quaternary (Furon 1963; Gourgaud and Vincent 2004). As reference, we note that Gourgaud and Vincent (2004) suggest an age of Lower Miocene for the initial thrust of volcanic activity and lava effusion in the Tibesti Volcanic Province, with a somewhat younger estimate for the majority of Emi Koussi products of 2.42 to 1.32 Myr. Here, we present the succession of volcanism within the Tibesti, based primarily on interpretation of the satellite imagery (Fig. 4). We stress that these divisions (Phase 1, Phase 2, etc.) do not necessarily denote changes in volcanic styles or significant breaks in volcanic or tectonic activity, but represent convenient groupings of the observed volcanic history of the Tibesti volcanism. Additionally, we do not infer details of the subsurface magmatic plumbing system. A short description of the main features within each phase in the TVP volcanic evolution is summarized in Table 5.

Fig. 4
figure 4

Simplified interpretation of six evolutionary phases of volcanism within the TVP

Full size image
Table 5 Phase evolution summary of the volcanism, structures, and products of the Tibesti Volcanic Province, northern Chad
Full size table

Phase 1

The initial phase of volcanism began in the Central TVP, following an inferred uplift and extension of the Precambrian basement. The relative dating of these volcanic products is based on the close proximity of the structures and their reasonably distinct deposits. Based on the ostensibly high degree of weathering of its central and flank regions, we suggest that the development of Tarso Abeki and its lava flows began early in the TVP history. Formation and growth of Tarso Tamertiou, Tarso Tieroko, Tarso Yega, Tarso Toon, and Ehi Yéy followed. Any deposits from their early activity have been obscured by subsequent eruptions.

Phase 2

The second phase consists of activity spreading north and east from the Central TVP. This migration formed not only Tarso Ourari, but the initial lava emplacement of the Eastern TVP as well. Volcanism also spread to the southeast, with the formation of Emi Koussi.

Gèze et al. (1959) indicated ancient ignimbrite deposits hidden beneath many portions of the Eastern TVP (the source as yet undetermined), including Emi Koussi to the southeast. These deposits are difficult to distinguish in satellite imagery, but if present would suggest earlier activity in that region contemporaneous to, or just after, the full-scale volcanism of the Central TVP.

Phase 3

It is during this middle stage that the eruptions and deposition of extensive lavas and tephra deposits from Tarso Yega, Tarso Toon, Tarso Tieroko, and Ehi Yéy took place in the Central TVP. Associated with Tarsos Yega and Tieroko were the effusive eruptions of lavas; large-scale, explosive eruptions leaving tephra deposits; and subsidence of the summit regions and caldera formation. The lack of a well-defined topographic rim or caldera floor at Tieroko might suggest structural variations in the style of magma chamber tapping, or differences in magma composition and flux from that at Tarso Yega. Also seen were further caldera collapse and associated widespread tephra deposits from Tarso Yega, as well as the appearance of lavas that make up the darker-toned Ehi Yéy, Tarso Tieroko, and Bounaï structures.

Several stages of lava emplacement have occurred within the large plateaux regions of the Eastern TVP, and individual phases are thought here to be distributed sporadically throughout the Tibesti Volcanic Province’s timeline of activity; the emplacement of lava flows from numerous vents formed the large volcanic plateaux of Tarsos Emi Chi, Mohi, and Ahon. These regions have been somewhat subjectively delineated in previous literature, and have no obvious relationships to the surrounding volcanics, apart from their relative position within the Eastern TVP. Furthermore, this phase of the TVP volcanism included activity that continued the initial construction of the Emi Koussi composite volcano to the south.

Based on stratigraphic relationships, textures of its products, and characteristics of its caldera structure in satellite imagery, we suggest that the full appearance of Tarso Voon within the Central TVP took place during this phase.

Phase 4

Lava emplacement within the Eastern TVP continued, though, according to Gèze et al. (1959), at a lower rate. We suggest that during this phase the emplacement of lavas associated with the Eastern TVP activity continued filling the geographical boundary between the Eastern and Central provinces; as stated, however, the timing is uncertain. It is likely that the sustained build-up of lavas and tephra at Emi Koussi also characterises this phase, though solid evidence for the timing of Koussi’s activity is difficult to discern due to its relative detachment from potential stratigraphic relationships within the rest of the TVP.

Within Phase 4, we also see the formation of Tarso Toussidé in the Western TVP. Contemporaneously, there was caldera collapse at the summit of Tarso Voon, and deposition of its associated extensive tephra took place, which mantled a large portion of the Central TVP.

Elsewhere, to the northwest of the Western TVP, the far-reaching lava flows associated with Tarso Tôh seem to be intermediate in both texture and weathering in satellite imagery, and are similar in those respects to the lava plateaux of the Eastern TVP, perhaps suggesting that the emplacement for at least a portion of these flows is broadly contemporaneous. Sediments within the Begour maar of Tarso Tôh have been radiocarbon dated at 8,300 ± 300 years (Hagedorn and Jakel 1969).

Phase 5

During Phase 5, lava production in the Eastern TVP wanes further still, and most of the large-scale activity throughout the TVP had abated. The most striking activity during this stage, however, was the development of the nested caldera systems of both Tarso Toussidé (Western TVP) and Emi Koussi (Eastern TVP). Compared to the extensive tephra deposits left by the collapse at Tarso Toussidé, Emi Koussi shows less evidence of widespread tephra deposition on its flanks, though the effects of aeolian erosion and perhaps precipitation should not be overlooked. Following tephra deposition on Tarso Toussidé, Ehi Sosso developed near the east boundary of the Western TVP. Finally, we see the extrusion of the lavas at Ehi Mousgou appearing to the northwest of Tarso Voon, on top of the most recent Voon tephra deposits.

Phase 6

The final phase of the TVP volcanism was characterized by the formation of Pic Toussidé on the west rim of the pre-Toussidé calderas, as well as several distinct, associated lava flows. The most prominent characteristic of Pic Toussidé is its obvious youth, based on the low visible and IR reflectance and clear definition of the lava flows, as well as the preservation of its cone structure and flanks. The appearance of Timi, in the northern portion of Tarso Toussidé, appears contemporaneous with the formation of Pic Toussidé, and has a fresh, youthful appearance as well, suggesting that weathering has had little time to affect the lava ramparts that make up its flanks. Relatively fresh lava flows and what we interpret as minor pyroclastic deposits on Emi Koussi are apparent in satellite imagery, and we place their development in this later phase as well. Several small, youthful cinder cones appeared on Emi Koussi’s flanks, particularly near or within the caldera region.

Most recently, the formation of Trou au Natron took place, which dissected the pre-Toussidé calderas; the Doon Kidimi crater, just east of the pre-Toussidé calderas; and the Era Kohor crater within the caldera system at the summit of Emi Koussi.

Present activity

There have been reports of manifestations of hydrothermal activity associated with the Tibesti massif, such as that at the Soborom Dome (Central TVP), and fumaroles near Emi Koussi at Yi Yerra (Eastern TVP), and at Pic Toussidé (Gèze et al. 1959). Deposition of carbonates at both Trou au Natron and Era Kohor have taken place relatively recently, as has the formation of the volcanic centres on the floor of Trou au Natron. Tarso Tôh, Tarso Toussidé and Emi Koussi are all listed as Holocene in age by the Smithsonian Institution’s Global Volcanism Program.

Discussion

Origin and development of the TVP

To account for the origin of the Tibesti massif, Gèze et al. (1959) suggested an upwarping of the basement due to the intrusion of a laccolith. A more contemporary interpretation infers the presence of a mantle plume beneath the cratonic African lithosphere (∼130–140 km thick; Ebinger and Sleep 1998). The buoyancy of the plume provided, and sustains, dynamic uplift of the Tibesti massif; early volcanism is likely to have been basaltic (Saunders et al. 1997). Geochemical evidence for the presence of a mantle plume has since been identified by Gourgaud and Vincent (2004). Similar origins have been proposed for other volcanic regions of Africa, including the Cameroon Line, Cameroon (Deruelle et al. 1991); the Hoggar, Algeria (Dautria and Lesquer 1989); and the Darfur Dome, Sudan (Franz et al. 1994).

Waning magma supply from the mantle source would likely result in the compositional evolution of the magma feeding portions of the Tibesti volcanism, ultimately producing late-stage viscous flows and domes (Keddie and Head 1994). At the youthful structure of Emi Koussi, slightly more felsic lavas (Gourgaud and Vincent 2004) sit atop several older, basaltic lapilli tuffs and lava flows. Vincent (1970) originally interpreted the observation of bimodal compositions for lavas within the Tibesti Volcanic Province as the result of two independent magmatic sources. Compositional bimodality such as this could also reflect assimilation of crustal material and magma mixing (Grove 2000). For Emi Koussi, Gourgaud and Vincent (2004) have shown that fractional crystallisation is the main differentiation process for the two mineralogical suites, though for the oversaturated series, crustal contamination probably took place as well. To date, no substantial geochemical or mineralogical data have been reported for any other of the Tibesti’s volcanoes.

The volcanoes of the TVP do not show any obvious relationship between age and relative position, as seen, for example, in the case of the Hawaii-Emperor and Cook-Austral volcanic chains. The size, elevations and geographic distribution of the Tibesti volcanoes presumably reflect the comparatively slow rate of motion of the African Plate with respect to the hot spot reference frame. Burke (1996) describes the African plate as effectively immobile for the past 35 Myr, allowing approximately 40 mantle plumes to penetrate the plate without the creation of any significant hot spot tracks. Likewise, O’Connor et al. (1999) restrict the absolute motion of the African plate to be no more than 20 ± 1 mm/year for at least the past 19 Myr. The prominent linear alignment of Cameroon volcanism (western Africa) is not attributed to hot spot-related tracks above a mobile African plate, but rather as a response to a concentration of extension normal to the Cameroon Line (Burke 2001). The volcanism itself is likely due to pressure relief beneath a line of extended lithosphere, with fairly regular spacing a result of convection (Burke 2001), or is localised by shear zones within the continental lithosphere (Deruelle et al. 1991). Furthermore, Burke (1996) refuted the possibility of plate motion as the mechanism for the observed linearity by postulating an age of <30 Myr for the onset of Cameroon volcanism, just after steady rotation of the African plate had effectively ceased.

Comparisons with other hot spot regions

Pike (1978) makes detailed quantitative comparisons of the gross form of 668 terrestrial and planetary volcanoes, resulting in a statistical analysis using five topographic measurements: height, flank width, and the diameter, depth, and circularity of the summit depression. The morphology of the major structures of Emi Koussi, Tarso Toon, Tarso Voon, Tarso Yega, and Tarso Toussidé fit statistically within what Pike (1978) categorizes as ‘alkalic stratocones with summit calderas’. One clear obstacle in otherwise comparing the Tibesti volcanics to other hot spot regions is its relatively unique continental placement. Additionally, the aridity and nature of the Sahara region affords a well-preserved view of the TVP structures from a remote sensing standpoint; compared with some other continental hot spot regions which have been significantly weathered or vegetated (e.g. the Snake River Plain in the western United States). With this in mind, we present straightforward comparisons of the Tibesti volcanic structures to a handful of volcanic hot spot regions on both Earth and Mars. We have chosen a relatively small sample set in an attempt to highlight first-order differences and similarities between more ‘conventional’ hot spot volcanism and that found within the Tibesti massif. Here, we limit our comparisons to the Hawaiian volcanic chain (Pacific plate), the Galápagos Islands (Nazca plate), the Canary and Cape Verdes archipelagos (African plate), the relatively nearby Jebel Marra in the Darfur region of western Sudan (African Plate), and to large-scale hot spot features on Mars. These regions were singled out for their diversities in location, regional lithospheric thickness, eruption styles, rate of plate motion with respect to an underlying melt source region, and general volcanic geomorphology. Comparisons of the Tibesti Volcanic Province to these locations are summarized in Table 6.

Table 6 Comparisons of tectonic and morphological characteristics amongst several terrestrial hot spot regions, the Tharsis Region of Mars, and the Tibesti Volcanic Province, northern Chad
Full size table

Lithospheric thickness

The Tibesti massif is situated on a stable cratonic region in the centre of the African plate. Ebinger and Sleep (1998) indicated an average lithospheric thickness of ∼130–140 km in the proximity of the Tibesti massif. Though similar, it is slightly less at the location of the 2,992-m elevation Jebel Marra volcano in western Sudan (∼110 km thickness; Ebinger and Sleep 1998). In comparison, the lithospheric thickness at Hawaii is approximately 80 km (Ribe and Christensen 1999). The lithosphere beneath the Galápagos is relatively young and thin (∼15 km; Feighner and Richards 1995). Filmer and McNutt (1989) suggested a thickness of approximately 48 km for the lithosphere beneath the Canary Islands, though Stein and Stein (1992) indicated a much greater thickness of 95 ± 15 km. Data for the Cape Verdes archipelago are sparse, as are data for the volcanic regions of Mars (e.g., Tharsis Bulge). Estimates of the Martian lithospheric thicknesses range from 100 to 500 km (Banerdt et al. 1996) and depend largely on assumed (and as yet unconstrained) values of its lithospheric rigidity and thermal history calculations. The geochemical effects of a thickened crust on rising magma sources (e.g. variations in trace element proportions, silica enrichment due to crustal contamination, etc.) have been investigated in regions such as the central Andes (e.g. Davidson et al. 1991) and Honshu, Japan (e.g. Kersting et al. 1996). The tectonic basis of volcanism throughout these regions is, of course, different from the manifestly intracontinental Tibesti massif, though associations between crustal contamination of rising magma, fractionation of the magma source, and volcanic products could be made with prospective sampling and geochemical analysis of the Tibesti region. The effects on Tibesti volcanism of a thick, cratonic lithosphere above a mantle source are as yet unexplored. In any case, the volcanism in regions most morphologically comparable to the TVP typically lies directly upon oceanic rather than continental crust.

Absolute plate velocity

Due to the slow-moving African plate, the Tibesti Volcanic Province shows no signs of linear tracks of volcanism or age-successive migration of the volcanic centres. The immense sizes (up to ∼600 km diameter) of several central Martian volcanoes (e.g. Olympus Mons and the Tharsis Montes) suggest that they, too, reflect at least in part a lack of plate motion relative to a magma source (Zimbelman 2000). Malin (1977) found remarkable similarities in summit caldera and flank morphologies between some of the volcanoes of the Tibesti (Emi Koussi, in particular) and Elysium Mons, Mars. His conclusions indicated that the source of magma for both regions has been essentially stationary with respect to the surface.

Although the Canary Islands do appear to age roughly eastwards (Morgan 1971; Dañobeitia and Canales 2000), historic volcanic activity occurring on several of the islands, including Lanzarote, Gran Canaria, and Tenerife (Simkin and Siebert 1994), complicate the picture of a simple age progression of the archipelago (e.g. Hoernle and Schmincke 1993). Likewise, the Cape Verdes Islands show a broad age progression from east (older) to west (younger) (Phipps Morgan et al. 1995). Nine of the western Galápagos volcanoes have erupted within the past 200 years (McClelland et al. 1989; Global Volcanism Network 1991a,b; Simkin and Siebert 1994), with no clear age progression. Though the rate of motion of the Nazca Plate above the Galápagos magma source (51 mm/year; Gripp and Gordon 1990) is significantly higher than that of the African plate, the focusing of magma supply below the Galápagos appears not simply to be controlled by plate motion (Rowland et al. 1994). The linear and age-related Hawaiian hot spot trend differs from all of these hot spots due to the high velocity of the Pacific plate above the established mantle plume.

Flank profile

In terms of gross morphology, the base and mid-flank regions of Tibesti volcanoes share similar slopes (∼4–12°) to those in Hawaii (3–13°; Mark and Moore 1987), as well as to many in the Galápagos. The major Hawaiian volcanoes tend to retain their gentle slopes at their summit regions, whereas several Tibesti volcanoes (Emi Koussi, Pic Toussidé, Tarsos Toon and Yega, in particular) exhibit steepened slopes as their summits are approached. The summit regions of both the Tibesti and Galápagos volcanoes are similarly steep, though in the case of the Galápagos, slopes are suggested to be due primarily to prolonged wave erosion of the outer flanks, modifying the gradients at lower levels (Rowland et al. 1994). The topographic profiles of shield volcanoes within the Galápagos archipelago range widely from values similar to Hawaiian volcanoes, up to 20–26° in some areas (Nordlie 1973). The summit and basal flanks of Jebel Marra, in Sudan, have morphological similarities to those of Emi Koussi and to those of the Galápagos, in particular, with gradients upwards of 13° at the central crater rim and a relatively abrupt levelling off of approximately 3° towards the mid-flank and base regions.

Rift zones/ fissures

Based on overt resemblances to the Galápagos volcanoes and Jebel Marra in gross morphology, we suggest that there may be similar development of the Tibesti Volcanic Province, the volcanism in the Galápagos, and western Sudan, despite differences in setting and scale.

The volcanoes of the Galápagos Islands are principally basaltic shields with relatively large summit calderas (∼5 km diameter). In contrast to what we infer about the nature of the Tibesti Volcanic Province, most eruptions of Galápagos volcanoes are from either radial or circumferential eruptive fissures. These fissures are the surface expression of dikes that have reached the surface (Chadwick and Howard 1991). In the case of the Hawaiian Islands, long, narrow rift zones have developed as a result of dike intrusion (e.g. Swanson et al. 1976; Cervelli et al. 2002). Effects of gravity on both the volcanic edifice and magmatic system have also resulted in the development of a prominent rift system in the Canary Islands (Ablay and Marti 2000), which possess the highest proportion of fissure vent volcanoes as primary features (Simkin and Siebert 1994). The existence of eruptive fissures or rift zones in association with the Tibesti massif has not been previously reported, and is not apparent in satellite imagery we have examined.

Caldera formation and morphology

Variations in subsidence geometries with respect to widespread caldera collapse events are discussed by Lipman (1997) and Walker (1984); the collapse style of a caldera can typically be inferred from the size of the existing caldera structure. Comparisons amongst several hot spot-related calderas (Fig. 5) indicate that the areas and dimensions of Tibesti calderas fall in the moderate- to high-end of the spectrum (∼12 km in diameter). Indeed, the above-average diameters of Tibesti calderas may correlate with the size of underlying magma reservoirs (cf. Lipman 1997). Based on the scale and characteristics of the Tibesti calderas, we postulate that the underlying mechanism for caldera collapse in the TVP has largely involved plate (piston) subsidence of a relatively coherent floor. Subordinate downsag could be a possible explanation for the slightly depressed floors of Tarso Voon, Tarso Yega, and the pre-Toussidé calderas. Clear evidence of piecemeal or multicyclic subsidence (e.g. stepped block faulting of the caldera floor, arcuate growth faults) has not been observed in most of the Tibesti volcanoes, though the stepped summit of Emi Koussi could potentially suggest multi-stage development. The presence of relatively well-defined rims and planar floors within the majority of Tibesti calderas exclude the likelihood that asymmetric subsidence has taken place. Future fieldwork is needed to determine the style of caldera formation for each of these centres.

Fig. 5
figure 5

Comparison amongst several hot spot caldera system dimensions, including both oceanic and continental regimes. Calderas associated with the Tibesti volcanoes are significantly larger than most other examples seen here. Caldera profiles for all structures were created using ASTER-based digital elevation models. Vertical exaggeration = 2×

Full size image

Calderas associated with Hawaiian volcanoes have developed by the loading of dense cumulates beneath the volcanic edifice, promoted by thermal weakening of the volcanic pile (e.g. Peterson and Moore 1987). Evidence for this type of mechanism includes positive gravity anomalies over the calderas, a stepped, funnel-shaped deformation profile, inward-dipping flows, and a large discrepancy in erupted volume versus subsidence volume (Walker 1984, 1988). Morphological characteristics such as these are not evident in the satellite imagery we have analysed of the TVP structures, though that does not completely preclude their presence. Data required to provide estimates of erupted and subsidence volumes of the TVP volcanoes have not yet been produced. Additionally, gravimetric data are not yet available at sufficient resolution for the Tibesti region to distinguish small-scale anomalies for better comparison.

The Las Cañadas nested collapse caldera (Tenerife, Canary Islands) is more comparable in scale to many calderas within the Tibesti Volcanic Province (see Fig. 5). Its present morphology reflects the successive migration and collapse of shallow magmatic chambers (Marti and Gudmundsson 2000), and has gradually enlarged over >2.3 Myr to a 25 km diameter due to successive flank failures (Cantagrel et al. 1999). Interestingly, evidence of significant flank failures is not apparent in satellite imagery of the TVP, however, and none has been reported previously. Several volcanoes within the Galápagos display near-vertical ring faults circumscribing their summits (e.g. Simkin and Howard 1970; Reynolds et al. 1995). A few related features (e.g. small volcanic extrusions, maars, cinder cones) can be seen in proximity to some Tibesti volcanoes (Emi Koussi and Tarso Voon, in particular) that may provide evidence for caldera-related ring faults, though no ring faulting is readily apparent in satellite imagery.

Based on these first-order comparisons, we suggest that the major Tibesti calderas appear to have formed by events similar to those found in the Galápagos (cf. Simkin and Howard 1970). The nearby Jebel Marra main volcanic centre shows similarities to Emi Koussi in satellite imagery as well, and the tectonic setting is, of course, similar.

Conclusions

Based on detailed analyses of satellite imagery, mostly from the high-resolution ASTER instrument, we have updated the only previous synoptic study of the volcanology of the Tibesti massif by Gèze et al. (1959). The relative inaccessibility of the Tibesti massif for most researchers establishes a prime example of the utility of remote sensing as an effective and efficient tool for volcanological investigation under such circumstances.

We propose that the TVP lacks any clear spatial progression of volcanic centres, though volcanism seems to have initiated in the Central TVP (e.g. Tarsos Voon and Toon). Volcanism subsequently migrated to the Eastern and Western TVP regions, developing substantial lava plateaux and additional volcanic features (including both Emi Koussi and Pic Toussidé). Geochronological investigations will be needed to further establish the details of TVP development, though estimates place the Tibesti volcanism to Late Miocene and younger (Gourgaud and Vincent 2004).

Comparisons with other hot spot volcanoes (e.g. the Hawaiian hot spot track, the Galápagos Islands, the Canary and Cape Verdes archipelagos, Jebel Marra, and Martian volcanoes) point to distinctions of the TVP, including the thickness and slow velocity of the lithosphere beneath the Tibesti region, its intracontinental tectonic setting, and its extent and diversity of volcanic features. The TVP does not present a linear hot spot track as seen, for example, in the case of the Hawaii-Emperor volcanic chain. The large scale of TVP calderas and the inferred plate subsidence and downsag as their primary collapse mechanisms point to further differences from other hot spot volcanoes. Future studies of the TVP are clearly warranted to further our understanding of the causes of these distinctions, including more detailed radiometric dating and reconnaissance-scale petrological and geochemical investigations. Relative volumes of the erupted vs. subsided material for each of the TVP volcanoes (and therefore the whole of the TVP) have not as-yet been constrained using remote sensing techniques. The lack of data concerning the elevation of the sedimentary basement, make volumetric estimates of the TVP volcanics difficult. Further, values could vary by orders of magnitude depending upon speculated thicknesses of erupted lava and/or tephra deposits. Again, future field investigations will be needed to provide robust estimates of volumes of products. A systematic, reconnaissance-scale field survey would greatly advance our understanding of the Tibesti Volcanic Province, and continental hot spot volcanism more generally, and we hope that this contribution stimulates further research of this fascinating region.