Cumbre Vieja Volcano -- Potential collapse and tsunami at La Palma, Canary Islands

Geological evidence suggests that during a future eruption, Cumbre Vieja Volcano on the Island of La Palma may experience a catastrophic failure of its west flank, dropping 150 to 500 km of rock into the sea. Using a geologically reasonable estimate of landslide motion, we model tsunami waves produced by such a collapse. Waves generated by the run-out of a 500 km (150 km) slide block at 100 m/s could transit the entire Atlantic Basin and arrive on the coasts of the Americas with 10-25 m (3-8 m) height. 1. Lateral collapse of island volcanoes -A tsunami wave source Lateral collapses of oceanic island volcanoes rank amongst the most spectacular natural events on Earth. Although no such lateral collapse punctuates the historical past, residual debris found on the seafloor evidence their abundance in recent geological time. Moore (1964) first identified the remains of lateral collapses off the flanks of Hawaii. Since then, dozens have been recognized adjacent to island volcanoes in nearly every ocean (Moore et al. 1994; Keating and McGuire, 2000). These observations constrain not only the geography and frequency of lateral collapses, but also their magnitude (up to 5000 km of material), extent (to 300 km length) and ferocity (underwater speeds to 140 m/s). Tsunami list among the many hazards associated with lateral collapses. Admittedly, direct geological evidence of tsunami in identifiable deposits (Moore and Moore, 1984) or coastal erosion features (Young and Bryant, 1992) are controversial. Still, history has documented large and damaging tsunami from far smaller lateral collapses of stratovolcanoes in island arc environments (Johnson, 1987; Satake and Kato, 2001). Prudence dictates that the lack of abundant wavecaused signatures associated with collapses of oceanic island volcanoes be viewed as more a function of scanty preservation than evidence that these events do not produce tsunami. Hazard from collapse tsunami may be particularly important in the Atlantic Ocean because of both the number of active oceanic islands there and the recent proposals (Day et al., 1999a, 1999b) that at least two of these volcanoes show signs of incipient instability. It seems timely then, for this paper to investigate the consequences of tsunami waves induced by a collapse of one of these unstable volcanoes -Cumbre Vieja on the island of La Palma, Canary Islands (Figure 1). 2. Geological evidence for a future collapse of the Cumbre Vieja During most if not all of the past 125ka, Cumbre Vieja has been the most active volcano in the Canary Islands (Carracedo et al., 1999). Subaerial Cumbre Vieja forms the southern third of the island of La Palma (Figure 2), rising 2 km above sea level with average slopes of 15° to 20°. The early Holocene has seen major changes in Cumbre Vieja. Day et al. (1999a) observe that, over the last several thousand years, the distribution and orientation of vents and feeder dykes within the mountain have shifted from a triple rift system (typical of most oceanic island volcanoes) to one consisting of a single N-S rift with westward extending vent arrays. They argue that these structural re-organizations are in response to evolving stress patterns associated with the growth of a detachment fault under the volcano's west flank. Coincident with the most recent eruption of the Cumbre Vieja in 1949 (Bonelli Rubio, 1950), the steeply inclined headwall section of this detachment surfaced as a west-dipping normal fault along the crest of the volcano (see Figure 2). The scarp extended 4 km with a maximum offset of 4 m. The appearance of surface rupture is ominous because: (1) initial subsurface development of a detachment fault, (2) its later propagation to the surface, and (3) ultimate slide block failure, typically sequences landslide development (Martel and Muller, 2000). Detailed examination of the 1949 rupture and geodetic measurements in the period 1994-1998 (Moss et al., 1999) indicate that the fault has been inactive since 1949. Inactivity is not unexpected however, because the triggering of flank instability on steep volcanoes generally requires additional destabilizing influences such as dyke emplacement or pressurization of trapped groundwater (Elsworth and Voight, 1995). These events often accompany a volcano's eruptive phases. This line of reasoning leads us to believe that a future eruption near the summit of the Cumbre Vieja will likely trigger a flank failure. To estimate the surface extent, subsurface geometry and total volume of such a failure we turn to geological evidence and comparisons with existing lateral collapse scars. Because the breadth of the Holocene structural changes in Cumbre Vieja appears to have affected the entire subaerial edifice, Day et al. (1999a) conclude that the developing detachment now underlies most if not all of the western flank of Figure 1. Inset. Canary Island chain off the western coast of Africa. Above. Location of La Palma Island, home to Cumbre Vieja volcano. As evidenced by the abundant landslide deposits strewn about their bases, the Canary Island volcanoes have experienced at least a dozen major collapses in the past several million years. WARD AND DAY: LA PALMA COLLAPSE AND TSUANMI 2 the volcano. The unstable block above the detachment extends to the north and south at least 15 km, however its length may be greater as these edges lack surface expression. The 1949 fault break skirts the crest of the volcano about 8 km inland and it marks the eastern boundary of the presently unstable zone. Because lateral collapses typically cut across the crest of volcanoes and into their reverse slopes (e.g. Mount St. Helens, Voight et al., 1983), we place the head of the future La Palma collapse 2 to 3 km east of the 1949 rupture (Figure 2). The western boundary of the unstable block lies hidden underwater. Bathymetric and imaging sonar surveys of older collapses at La Palma and elsewhere (Watts and Masson, 1995; Urgeles et. al., 1999) suggest that the toe of the block surfaces in 1 to 3 km water depth -about 5 to 10 km offshore. The best geological evidence that we have paints a Cumbre Vieja collapse sending down a slide block 15-20 km wide and 15-25 km long. The thickness of the slide block is not easily fixed. Mapping the depth to the detachment surface by locating earthquakes that occur on it has not been possible. No records exist of seismic activity associated with the 1949 eruption, or the subsequent 1971 eruption at the island's southern tip. No other tectonic earthquakes of consequence have struck under La Palma in the last three decades either. Nevertheless, characteristics of past collapses point to a listric detachment 2 to 3 km below the summit of the volcano. Toward the west, the surface dips seawards at a shallow angle to intersect the offshore toe. Toward the east, the detachment steepens sharply to intersect the surface within a few km of the mountain's crest. In consideration of everything, we believe that a future flank failure of Cumbre Vieja volcano will dislodge a broadly wedge-shaped slide block as cartooned at the bottom of Figure 2. The volume and mean thickness of rock participating in a flank failure depends upon the detailed shape of the basal surface, but they should fall in the range of 150 to 500 km and 1 to 2 km respectively. The inferred geometry and volume of the expected failure coincide closely with features of the previous La Palma collapse (~566 ka), remains of which are still visible to the north on Cumbre Nueva (Day et al., 1999a). 3. Landslide Tsunami Model Generalities The section above provides a feeling for the size and shape of the block that may slide into the sea during a lateral collapse of Cumbre Vieja. What magnitude of tsunami might this collapse induce? One straightforward means to address this question employs classical, linear wave theory. Consider a uniform ocean of depth h. Under this theory, a general vertical bottom disturbance (i.e. the landslide) u z bot (r,t) starting at t=0 stimulates surface tsunami waveforms (vertical component) at observation point r of (Ward, 2001)