|
|
|
| |
|
|
|
|
| |
|
( rubrique Poubelle ) 473- De Mylène_coquine, à Paris, 25 ans :
Aidez moi s'il vous plait
| ( 4 réponses ) ( 357 lectures )
|
|
|
| |
| 1 - | | réponse de Mylène_coquine, à Paris, 25 ans : |  | |
|
|
invité(e) | | réponse initiale de Mylène_coquine, à Paris, 25 ans:
Je 15 Sept 05, 21:12 
bonjour, allez voir ma page pour un concours
http://sweeping13.free.fr
cliquez sur l'image puis le lien. Si vous voulez un cadeau gratos laissez un msg a ce message. Le 100eme qui met un msg et qui a cliqué je prend on email et je le contacterai pour lui envoyer une clé usb sony 256Mo valeur : 50 euros. Merci à tous ceux qui m'aideront et qui cliqueront
|
2 - réponse de kiko, à Paris, 00 ans:
Ve 16 Sept 05, 18:43  
Assez de pub!!! non?????????
3 - réponse de marlonne, à ici, 20 ans:
Ma 20 Sept 05, 1:42
yen a ki ont vraiment rien a faire!
4 - réponse de d, à d, d ans:
Ma 4 Oct 05, 18:19
Island Edifice Failures and Associated Tsunami Hazards
Abstract-Volcanic ocean islands are prone to structural failure of the edifice that result in landslides that can generate destructive tsunamis. These island landslides range enormously in size, varying from small rock falls to giant sector failures involving tens of cubic kilometers of debris. A survey of literature has allowed us to identify twenty-three processes that contribute to edifice collapse. These have been divided into endogenetic and exogenetic sources of edifice failure. Endogenetic sources of instability and failure include unstable foundations, volcanic intrusions, thermal alteration, edifice pore pressures, unbuttressed structures, and buried faults. Exogenetic sources of instability and failure include collapse of subaerial or submarine deposits, endo-upwelling, karst megaporosity, fractures, overstcepcning, overloading, sea-level change, marine erosion, weathering including hurricanes, glacial response, volcanic activity, regional uplift or subsidence, tectonic seismicity and anthropogenic agents. While the endogenetic sources dominate during periods of active volcanism and**** building, the exogenetic sources may cause failure at any time. Tsunamis, both small and large, are associated with these edifice failures.
Key words: Island, edifice failure, landslide, tsunami.
Introduction
Ocean islands are prone to structural failures that result in landslides that are known to produce tsunamis. Island landslides range from minor rock falls to giant landslides (MOORE, 1964; VOGT and SMOOT, 1984; MOORE. et al., 1994a,b; CARRACEDO, 1994; DuPON, 1984; JONES et al., 1984; TAYLOR el al., 1980; KODAGALI et al., 1998; HoLCOMB and SEARLE, 1991; SMOOT, 1985) with the submarine landslides around the Hawaiian islands some of the largest on Earth (LIPMAN et al., 1988; NORMARK et al., 1993a). Multiple destabilization processes operate simultaneously, producing events of varying magnitudes and frequencies. Some volcanoes appear to have a far greater potential for failure than others.
Volcanic edifice collapse has been identified as the primary cause of instability on islands and atolls within the geologic record (FAIRBRIDGE, 1950) and at many currently active volcanoes (SIEBE:RT, 1984; UI, 1983). SIEBERT (1992) estimates that there have been four structural failures of volcanic edifices each century during the past 500 years. This is probably an underestimate since three major sector collapses have occurred this century in the Kurile-Kamchatka region alone (BELousov, 1994) and open amphitheater-shaped depressions arc evidence of additional failures on 22 Kamchatkan volcanoes (LEONOV, 1995).
About 5% of all tsunamis are estimated to have been formed by volcanic activity, and at least one fifth of these result from volcanically induced landslides (8mm) and SHEPHERD, 1996). For brevity's sake, readers are referred to tsunami texts such as MYLES (1985). DUDLEY and LEE (1988), and SIMPKIN and FISKE (1983) for additional information. Three examples are summarized here as examples of catastrophic tsunamis generated by landslides. At Mt. Unzen volcano in Japan, during 1792, a small subaerial landslide (estimated at 0.34 cubic kms) generated destructive waves when it poured into Ariake Bay (HAYASHI and SELF, 1992). The landslide (not connected with any volcanic activity per se) generated a tsunami that killed 14,500 people.
After Krakatoa erupted in a violent eruption-column collapse in 1883, only one-third of the original island (5 by 9 kms) remained. While the Krakatoa eruption was violent and probably involved a complex combination of caldera collapse, pyroclastic flows, landslides etc., LATTER (1981) argues that submarine landslides generated during the volcanic disruption produced tsunami waves 40 m high. LATTER (1981) used the phrase volcanic tsunami to describe the tsunami that arrived 15 minutes before the corresponding air waves. It is estimated that over 36,000 people lost their lives in this volcanic collapse which began with a vertical caldera collapse and was accompanied by lateral collapse of the disrupted material (SIMPKIN and FISKE, 1983) and associated tsunami destruction.
As recently as the fall of 1998 a tsunami struck within 17 minutes of an earthquake off the coast of Papua New Guinea (PNG). The first of three waves reached heights of 6 tol5 m and washed 2 km inland along the coastal zone near Aitape. Seven coastal towns were destroyed on the seaward side of Sissano Lagoon and upwards of 2,000 people lost their lives. Initial analysis indicates that the wave height was far higher than predicted from the earthquake source and that modeling shows the tsunami had very high amplitudes (SYNOLAKis et al., 1998; IMAMURA et al., 1998; LETZ et al., 1998; Tn ov and GONZALEZ, 1998). TAN1OKA (this issue) discusses the causes of early arrivals of the deadly Great Kuril Tsunami of 1994. It seems likely the earthquake (not associated with any volcanic activity) near Aitape triggered a submarine landslide on the steep island slope and that the landslide created the tsunami (TANIOKA and RUFF, 1998; IMAMURA et al., 1998). Recent surveys have documented many rock falls on the submarine slopes of Aitape but have not located one large debris avalanche of sufficient size to generate the observed tsunami (CROOK and SAKATE, personal comm., 1999). The Papua New Guinea region has experienced 65 historical earthquakes between the years 1768 to 1985 (Reference: http://omzg.sscc.ru/tsulab/tsun19980717. html and V. K. Gu-SOAKOv, personal comm., 1998). Since landslide related tsunamis are capable of great destruction and loss of life, the nature of their destructiveness within the ocean island context is well worth study and review. Currently our understanding of oceanic island failures and associated tsunamis is very limited.
Volcanic Instability
Volcanic instability has been defined by MCGUIRE (1996), "as the condition within which a volcanic edifice has been destabilized to a degree sufficient to increase the likelihood of structural failure of all or part of the edifice." Failures can occur over a long period of time or instantaneously and can range in size from small rock falls and slumps with volumes of several hundred cubic meters, to slides a few thousand cubic meters in volume (e.g., ROWLAND and MUNRO, 1992; MCGUIRE et al., 1991, 1993; MUNRO and ROWLAND, 1996). Giant lateral col-lapses. with volumes in excess of 2,000 cubic km, have occurred on the flanks of large ocean-island volcanoes (MOORE et al., 1994a,ó; CARRACEDO, 1994; Hot. COMB and SEARLE, 1991). Some Hawaiian landslides exceed 300 km in length making them the largest such structures on the planet (MooRE. et al., 1994a,b). FILMER et al. (1994) and SMOOT (1995) suggest that up to half of the volume of the Hawaiian island chain consists of material displaced by mass wasting (Fig. 1).
Small volume collapse events probably occur at one volcano or another every few days to years, the frequency of giant ocean-island landslides is unknown. Giant flank collapses of the Hawaiian volcanoes may occur every 25-100 Ka (LIPMAN et al., 1988; NORMARK et al., 1993a) but the uncertainty of these age estimates is large, and further study is needed.
Structural failures result from any one of a large number of destabilization processes, twenty-three processes that contribute to collapse are identified in Figures 2 and 3 and are divided into endogenetic (intra-edifice) and exogenetic (extra-edifice) processes (Tables 1 and 2). The endogenetic processes dominate during times of active volcanism and**** building while exogenetic processes may cause failure at any time. In the discussions that follow, the observations leading to the identification of a destabilization process, the collapse mechanisms, and specific examples are cited based upon historical data and geologic records.
For the ocean-island setting, little quantitative data are available which describe the nature of the edifice destabilization, collapse, or the resulting tsunamis. Complexity and infrequency of observation forestall the development of process-based models. This synthesis of the historic record is an effort to identify destabilization processes involved in edifice collapse and encourage research into the threshold conditions governing destructive ocean island landslides.
Endogenetic sources of failure
Figure 2
Schematic profile of an oceanic volcano illustrating the distribution of endogenetic sources of failure.
Symbols used in the illustration include: vertical lines indicate dykes, the double dash pattern indicates
thermally altered zones, the v-pattern indicates a magma chamber, dotted pattern indicates subaerial and
submarine debris, and the patterns of lines and ovals illustrate lava flows and pillow basalts. This sketch
is not to scale, the vertical scale being exaggerated for illustrative purposes. Endogenetic sources of
failure include (1) unstable pelagic clay foundation, (2) unstable volcanic foundations, (3) zones of
thermal alteration, (4) cumulates, (5) dykes, (6) positive edifice pore pressures, (7) unbuttressed flanks,
and (5) buried faults. These topics are discussed in the text in the order shown here.
Endogenetic Failure Processes
Foundation Slip Surfaces
Since volcanic foundations strongly affect the overall geologic stability of the volcanic edifice, we begin with a discussion of ocean island foundations. Most mid-plate volcanic island chains (with notable exceptions) were formed by hot spot activity on ancient sea floor covered with poorly consolidated sediments, mostly pelagic clays accumulated in an aqueous environment and are generally imcompetent and easily deformable under excess load (of the volcano and ocean). DIETERICH (1988) examined the friction coefficients assuming normal hydrostatic pore pressure and concluded that the slip was compatible with the friction measurements of clay-rich material, but that slip on faults within the volcanic material would likely require overpressuring of the pore fluids. He notes that because of their low permeability, clay sediments are likely to develop excess hydrostatic pressures during compaction in response to the weight of the volcano.
Figure 3
Schematic profile of an oceanic island indicating the distribution of exogenetic sources of failure.
Symbols used in the illustration include: vertical lines indicate dykes, double dash pattern indicates
thermally altered zones, the v-pattern indicates a magma chamber, the dotted pattern indicates subaerial
and submarine debris, and the patterns of lines and ovals to illustrate lava flows and pillow basalts. This
sketch is not to scale, and the vertical scale is exaggerated for illustrative purposes. The exogenetic
sources of failure are numbered here in the order they are discussed in the text.
Decoupling can also occur in low-strength layers (e.g., rubble zones or along buried faults) within the subvolcanic sequences, not necessarily immediately under-lying the volcanic pile (VAN WYK DE VRIES and BORGIA, 1996; COLINI and BORGIA, 1997). Weak materials within the volcano also provide zones of effective decoupling that aid gravitational collapse or edifice spreading (NAKAMURA, 1982, 1980; SWANSON et al., 1976). CROSSON and ENDO (1982) show evidence that nonvolcanic earthquakes in Hawaii are associated with slip planes coincident with zones of weakness, e.g., the pelagic sediment layer above the "old oceanic crust" that extends as deep as 10 km (LIPMAN et al., 1985; LIPMAN, 1995).
BORGIA (1994) also suggests that unstable clay-rich layers (VAN WYK DE VRIES and BORGIA, 1996) underlying a volcanic pile provide unstable foundations (see Table 1). Deformation and growth of the volcano is accomplished, at least partly, by means of outward and upward displacements along thrust faults in the incompetent sediments (CoLiNt and BORGIA, 1997; GH.LARD et al., 1996), where such a substrate is not present or much thinner, thrust-related deformation appears minimal or absent (NAKAMURA, 1980, 1982). Examples of volcano collapses that may be attributed to unstable foundations are summarized in Table 1.
Rubble Zones
Using the SeaBEAM mapping system and submersible dives, MALAnoFF (1987) studied Loihi seamount (Hawaii), a site of current submarine hot spot volcanism, and observed widespread mass-wasting on the submarine slopes. GRIGG (1997) made submersible observations on the slopes of Loihi and reports that highly unstable volcanic rubble piles physically restrict faunal development on the summit and flank of Loihi. MOORE et al. (1989) suggest that two thirds of the surface area of Loihi has been modified by landsliding. Similarly, SeaBEAM bathymetry, GLORIA images, and MMR-l sidescan images reveal that the southern slopes of the island of Hawaii are largely covered by the products of mass wasting (SMrrH and SHEPHERD, 1993, 1996) and more than 5,000 cubic km of the Kilauea volcano south flank (Fig. 4) is moving southward and downward (DENLtNGER and OKUBO, 1995; GILLARD et al., 1996). Much of the southern flank of Kilauea volcano is moving, producing an outward displacement estimated to be roughly 0.1 m/yr (BORGIA, 1994). Richard Fiske suggests, "ultimately, there's going to be some sort of failure though we don't know when" (KERR, 1994). Seismic reflection studies
...
1993) and SeaMARC 11 sidescan sonar imagery and bathymetry on the south flank of Savaii Island document extensive failures (KEATING and KAao(,onINA, 1991a,b; KEATING et al., this issue).
Dykes Intrusion
WALKER (1986, 1987) measured the coherent Koolau volcano dyke complex exposed in the eroded shield lavas of east Oahu, Hawaii. At its widest point the complex is 7 km and connains 7,400 dykes. The individual dykes were measured and the median dyke width was found to be less than 0.5 m, increasing downrift (WALKER et al., 1995). When combined, the overall widening of the volcano caused by dyke intrusion totaled over 4 km. The intrusions produce both vertical and horizontal displacements. Because the seaward flanks of volcanoes are unbuttressed, the failure that results from accommodating the vertical and horizontal extension of dyke intrusions occurs as gravitationally induced lateral collapse of the margins.
In the case of oceanic volcanoes, repeated dyke emplacement along a preferential path takes place over long time periods and typically results in the growth of a pronounced topographic ridge. WALKER (1987) points out that in the Koolau volcano (East Oahu, Hawaii) the dykes are clustered parallel to the longitudinal axis of the volcano and display a roughly perpendicular secondary dyke swarm (referred to as the southwest rift zone) on the unbuttressed side of the volcano. This three-part dyke complex is typical of Hawaiian and many other oceanic volcanic structures. Evidence for the collapse of the windward side of the Koolau volcano is seen in GLORIA images of the sea floor (MOORE et al., 1994a). The edifice destabilization process involves physical displacement of the unbuttressed flanks of the volcano. The example cited above for the Koolau dyke complex on Oahu is an important observation of this type of growth, which is believed to be responsible for repeated slope failures. Cable route surveys on the western slopes of the Waianae volcano, Oahu, indicate extensive mass wasting and fragmentation of detrital blocks (CAMPBELL, personal comm., 1997) along the unbuttressed western flank of the Waianae volcano (West Oahu, Fig. 5).
LIPMAN (1995) has suggested that the injection of dykes within the central dyke complex and the rift zones causes lateral extension (or lateral spreading) of Hawaiian volcanoes. While BORGIA and TREVES (1992) argue, "Hawaiian volcanoes are not passive piles of lava but dynamic constructs that deform under their own weight and the pressure of magma intrusion. From the summits of these volcanoes rift zones diverge, extending to the ocean bottom and dissecting the volcanic edifices. The flanks spread away from the rift zones on basal thrusts thought to be located in the sediment layer between the volcanoes and the ocean floor."
Figure 5
NASA satellite imagery of western Oahu, Hawaii. The Koolau volcano to the east is covered by a cloud hank. The summit of the Waianae volcano on the west side of the island is marked by an arcuate formation of fewer and more isolated clouds. The western margin of the Waianae volcano has been removed by landslides, leaving three arcuate cusps in the coastline. The image was provided by Peter Mouginis-Mark courtesy of the Virtually Hawaii Website sponsored by NASA.
Cumulates
Subhorizontal sheet intrusions of magma into the volcanic edifice are also thought to destabilize a volcano. Within Kilauea volcano, RYAN (1988) proposed that a dense mush of olivine crystals and magma liquid pools within a magma chamber 6-10 km below the summit flows outward under the force of gravity. The wedge-shaped cumulate mass, formed beneath the volcano, drives the unbuttressed flank seaward. CLAGUE and DENI..INGI-:R (1994) agreed that a dunite mass at high temperature would flow outward as Ryan proposed. Furthermore, the dense mass would likely provide a surface along which detachment could occur. Under this scenario, the process of displacement of the volcano flanks is similar to that discussed above for dyke injection. BORGIA et al. (1992) suggest a plutonic complex at a depth of roughly 5 km is present seaward of the Mt. Etna summit caldera, lying above a decollement or fault zone. The stability of the Mt. Etna volcano is discussed in several publications including SFE WAR"( et al. (1993), MURRAY and VoIGHT (1996), RASA et al. (1996), Rus) and N .i i (1996), and FIRTH et al. (1996).
ADDSIIKIN el al. (199) and 1)tI.INil N (1995) proposed that the emplacement of a similar magma hods at the unstable Klvuchevskoi volcano in Kamchatka triggered edifice failure resulting in a debris avalanche with a volume of 4-8 cubic km.
termal alteration
Persistent magma injection also weakens the volcanic edifice by thermally altering rocks adjacent to the intrusions. 1=1syOR and Vololrr (1992, 1996) suggest a process in which a slope at the limit of its stability may be induced to slide mechanically. especially when thermally generated increases in pore-fluid pressure accompany the magma emplacements. I)vs (1996) proposes that similar conditions result Irom intrusion degassing. When there is discharge of pressurized fluids from depth via elastic dykes, or faulting associated with deformation and pore collapse.
The unusual mobility of volcanic debris avalanches has often been ascribed to hydrothermal and or magmatic fluids. gases. and or pre-failure fracturing that promotes efficient fragmentation of the collapsing portion of volcanic edifice (Ut. 1983: VonViii el al.. 1983: Stint litT Cl al., 1987). Besides changes in pore pressure. intense hydrothermal alteration chemically alters volcanic rocks oyer time to produce clays that may facilitate edifice destabilization (SmBEh r el al.. 1987: SCrni t Si AN and S t stTDi(irL. 1994: Dt DOICNON et al., 1997). /ones of repeated dyke injection are likely to he weakened by hydrothenmd alteration (STONE and FAN. 1978). Because hydrothermal alteration plays a major part in increasing susceptibility to failure (Sit lit RI if al., 1987). lateral collapse events arc****nonly associated with phrcalic explosions (e.g.. at 13andai-san. Japan. in 1888).
Outcrops on the islands I uPahna and El Ilierro (CARR (l 1994. 1996) in the Canary Islands (Digs. 6 and 7) provide examples of major rift-related flank failures. as well as two examples of lateral collapse events which appear to haie been aborted during the early stages of sliding. The !ault,s of the aborted collapse are well exposed and are characterized by dry fault breccias and ultra-cataclasites. rather than the extremely fluidized gouge muds and mud-rich breccias encountered in exhumed collapses elsewhere in the Canary Islands (DAY. 1996). It is possible that the lack of pressurized fluids on the faults, and a conseqent lack of significant slip weakening (Ranh and IItiRBERT, 1959: Rt( L. 1992) or of brecciation of the collapsing slump blocks by gouge dykes (1)Av, 1996), may' have caused fault movement to cease prior to destabilization. The timing and nature of vertical and lateral collapses on Tenerife (Canary Islands) were discussed by MARTI CI al. (1996.
1997. 1998). ANCOCIII et al. (1998). URGE) FS Cl ul. (1997) and the "It ut (iRotiP (1997). with additional ev idence of collapse identified in studies of offshore deposits by PEARL and JARVIS (1992). Wtl,vVLR ct al. (1992. 1994), MASSON (1994. 1996). WAErs and MASSON (1996). GEiSSLINGER et al. (1996), and Routtss and CR vNIP 11996).
A local earthquake at Yasur volcano (on 'Parma Island in Vanuatu, Fig. 8) produced a flank collapse in hydrotherntally altered (and weakened) rocks that generated a 12-n1-high local tsunami in 1878. Historical records from Port Resolution indicate a large landslide temporarily blocked the mouth of the bay (PATTON, 1894) and uplifted the west side of the hay by 6 m (PA t i oN, 1894; LAWRIE, 1898; MAWSON, 1905). The 1878 Port Resolution tsunami is likely to have resulted from an earthquake triggered submarine landslide. Evidence of failure includes the observed debris flow scar downslope of a fuinarole field on the western side of the
Figure 6
NASA Space Shuttle Photograph SISI 4-OHS-f ~? (taken Sept. 19r shows La Palma Island in the Canary Islands off Weston Africa. The Cumbrc Nueva caldera can be seen at the center of the island.The Ronelli fault zone extends south from the F urnbrc Nueva caldera and may be an impending asymmetric nit zone collapse feature i (55551 i sm. 199t0- Ihis image was provided by Chuck Wood courtesy of the Volcano Website and NASA.
Figure 7
NASA Space Shuttle Photograph STS052-0073-004 shows El Hierro volcano in the Canary Islands off Western Africa. The island last erupted in 1793. CARRACF.OO (1996) suggests that the three peninsulas visible in this image are the volcanic rift zones. The El Golfo avalanche headwall discussed by MASSON (1996) to the northwest side of the volcano is bounded on the east by dark cusp-shaped shadows in this image. This portion of the volcano collapsed seaward about 13 to 17 Ka (MASSON, 1996). The subaerial faults associated with the collapse are visible in the Space Shuttle Photograph. This image. was provided
by Chuck Wood courtesy of the Volcano Website and NASA.
bay, a reported debris flow (PATTON, 1894) and a shallow marine lobate debris apron recorded on the bathymetric map of 1885. CHEN et at. (1995) conclude that landslides slumping seaward of the Yenkahe block pose a serious tsunami hazard. They found rapid uplift of blocks accompanied by hydrothermal alteration of volcanic soils (adding lubrication) and predict sudden block failures are likely to produce future destructive landslides and tsunamis.
Edifice Pore Pressure
Pore pressure plays an important role in gravitational gliding. Increased pore-fluid pressures lower frictional resistance along a fault surface, so that the gravity overcomes the effective normal stress on the fault (Gun) et at., 1982). High pore pressure can occur in various submarine settings, e.g., in association with intrusive rocks, in zones of endo-upwelling (sea-water circulation within a volcanic edifice),
Figure 8
NASA Space Shuttle Photograph of Yasur volcano (Tanna Island, Vanuatu, southwest Pacific Ocean).
This photograph was taken October 4, 1994. An ash plume originating on the northwest coast obscures
a portion of the image. The island is a semicircular fragment of a considerably larger volcanic island,
with much of the western portion of the volcano no longer above sea level. This image was provided by
Chuck Wood courtesy of the Volcano Website and NASA.
in deposits rich in organic material, in areas of rapid sedimentation, or in deposits in which the surface/water boundary has been armored (by Mn crust accumulation, subaerial diagenesis, or other processes).
In 1995, an earthquake of magnitude 6.1 occurred offshore the town of Aegion in the western Gulf of Corinth (Greece). The earthquake triggered small subaerial to submarine sediment failures in at least four sites and created three fan delta
deposits. The failure mechanism was liquefaction caused by elevated pore pressure enhanced by the presence of gas (PAPATHEOooROU and FFRFNTINOS, 1997). The failures occurred on slopes of only 0.2 to 2.1 degrees and produced ground
cracking. rotational slides, elongated slides, sediment gravity flows, and sand boils.
VoIGIIT and EESWORTII (1992) and ELSWORTH and VOIGHT (1992, 1996) have drawn attention to the potential role of dykes in triggering structural failure as a conseqence of raised pore pressure by means of thermal straining that caused liquefaction of overpressured sediments.
The circulation of sea water within a seamount (termed endo-upwelling) has been described by many scientists including ROUGERIE and WAUTHY (1986, 1988, 1993), RouGFRIF et al. (1980, 1991, 1992) in Mataiva, Niau lagoon, Mururoa (French Polynesia) and Tikehau atoll (Tuamotu chain) and on seamounts near New Zealand (PATCHELL, personal comm., 1999). Endo-upwelling results from connection within the edifice that creates positive interstitial pore pressures which can facilitate collapse of the carbonate bank deposits that cap many volcanic edifices.
Reef research in the Pacific and Caribbean shows that maximum cementation and porosity reduction occur along the outer margins of ancient reefs and decrease toward the lagoon. These diagenetically altered reef rims provide a hardened submarine surface, allowing a buildup of interstitial pore pressure. At the interface where the diagenetically altered, hardened surface is intact, the interstitial voids within carbonates can become overpressured and fail. With failure the altered surface is removed and circulation can be re-established. Where the hardened surfaces are no longer intact, discharge can occur.
The accumulation of gases from the decomposition of organic detritus leads to the formation of gas hydrates (MACDONALD et al., 1994). The hydrates can change their physical state, changing from solid to gas with increasing pore pressure during times of reduced hydrostatic pressure and a warmer ocean associated with times of lowered sea level (EsRIG and KIRBY, 1977; KAYEN and LEE, 1991; ROTHWELL et al., 1998; HAQ, 1998). Rapid dissociation of hydrates (occurring after only 1 degree C temperature warming) has been found associated with large-scale slumping of marine sediments (PAULI, et al., 1991). On the shallow flanks of seamounts or island flanks, the frothy degassing sediments indicative of hydrates were observed in pipe dredges from depths of 800 1,200 m in the western Pacific Ocean (in the Enderbury Island area). Thus, hydrates may be a potential contributor to sub-marine landslides on seamounts as well.
Creep Associated with Unbuttressed Structures
Failed margins and aprons of destabilized material around oceanic volcanoes shows that the seaward-facing flanks of island volcanoes are inevitably the least buttressed. This observation also applies to coastal volcanoes such as Etna (MCGUIRE, 1996), where the topography becomes increasingly elevated inland, and to island volcanoes such as Stromboli (KOKELAAR and ROMAGNOLI, 1995; TIBALDI et al., 1994; TIBALDI, 1996) or Hawaii where younger centers (such as Kilauea) can be buttressed by older edifices (e.g., Mauna Loa). MooRE (1987) points out that portions of the slopes of Hawaii are characterized by arcs of anomalously steep terrain and bulging zones that represent seaward-creeping gravitational failures. The process of gravitational failure can involve the slow displacement or creep of sectors of the edifice in the form of giant slumps, coseismic downfaulting, rapid-moving catastrophic debris avalanches, or a combination of any of the three. The morphological asymmetry resulting from creep eventually leads to the preferential release of accumulated exogenetic and endogenetic stresses, such as surface overloading or to repeated dyke emplacement in a seaward direction.
Where individual islands in a volcanic archipelago are sufficiently close, for example in the Hawaii-Emperor Chain preferential edifice buttressing ensures that the majority of collapse events occur perpendicular to the length of the chain. In the case of more widely-spaced or less linear archipelago, such as the Canary Islands, this relationship is less well defined.
Submarine mapping shows that landsliding has occurred in the extinct north-west Hawaiian chain (Fig. 9) where dyke injection terminated millions of years ago. While the active part of the chain has debris avalanches extending 150 km to fill the Hawaiian Deep (a moat around the active hot spot), the northwestward chain displays debris avalanches extending twice as far from the chain, 250-350 km from the islands. The evidence for continued mass wasting long after the active construction phase of volcanism ceases, indicates that creep (coupled with other contributing factors) is an important failure process. ROBERTS and CRAMP (1996) point out that "creep is a process that can be considered as both a trigger mechanism and a failure type." HILI_ et al. (1982) use seismic and geotechnical data to suggest that while strain rates are low, rock deformation by creep continues over long time periods and is a significant destabilization process.
Buried Faults
Buried faults may be zones of weakness for subsequent deformation when the surface load from accumulation of lava becomes sufficiently large or when horizon-tal displacement from dykes becomes sufficiently great. LIPMAN (1995) reports that, "rapidly emplaced giant submarine landslides have catastrophically removed entire flanks of Hawaiian volcanoes..." Much of the early slide material from the southern flank of Mauna Loa, along the Wai'ohinu-Kao'iki fault zone, is buried beneath the present-day Kilauea volcano. Mauna Loa is an older volcano and its western flank has features thought to be the faulted head walls of the South Kona landslides. The headwalls are associated with a landslide estimated to be roughly 100 ky old but buried by post-landslide lavas.
NORMARK et al. (1993a) suggest large-scale landsliding takes place intermittently throughout the formation of the volcano so that slope failure deposits are incorporated into the edifice, with little outcrop evidence of the prior displacements remaining. Buried faults represent past structural failures. These faults are potential zones of weakness within the volcano that could again fail under stress. Reactiva-tion of these faults is likely to result in greater edifice destabilization and degradation. Future collapses are likely to occur along these pre-existing seaward dipping zones of low friction.
During 1975, Kilauea volcano experienced a failure event at Kalapana, involving an earthquake and associated tsunami, deformation from the summit to the coast, fault offsets on the Hilina system, and underwater block slumps (LIPMAN, 1995). During the Kalapana event, pre-existing faults were locally reactivated.
Outcrops on the islands of LaPalma and El Hierro (CARRACEDO, 1994, 1996) in the Canary Islands provide excellent evidence of prior rift-related flank failures. Faults with limited displacement are well exposed and are characterized by dry fault breccias and ultra-cataclasites, whereas extremely fluidized gouge muds and mud-rich breccias are encountered in exhumed collapses of large displacement (DAY, 1996), like those suggested by LIPMAN (1995). Drilling completed as part of nuclear test work in the Pacific also provides evidence of the internal faulted nature of Pacific oceanic islands. Published reports by CHAVEAU et al. (1967), GoGuEE (1982), and BoucliEz and LECOMTE (1995a,b) attest to faulting of the volcanic pedestal underlying the atolls of Mururoa and Fangataufa.
Exogenetic Factors Associated with Failure
Subaerial Deposits
The subaerial portion of mid-ocean island edifices is typically shield volcanoes consisting of superposed stacks of alternating lava flows and pyroclastic deposits (SWANSON and CHRISTIANSEN, 1973). Portions of the flanks near coastlines consist of weak hyaloclastites (PETERSON, 1976); rubble, pyroclastic layers (DZURISIN et al., 1984); and littoral****s (PETERSON, 1976; DZURISIN et al., 1995; DECKER and CIIR1sT1ANSEN, 1984; REHM and HALBECH, 1982). Such deposits typically have low cohesive strength and are more susceptible to erosion, deformation and ultimately collapse.
In the coastal areas, unconsolidated deposits or poorly consolidated ash can be complexly interspersed with beach or reef deposits, resulting from subsidence of the volcano and sea-level changes (HAWAII ScIENTH,IC DRILLING PROJECT TEAM, 1996; BEESON et al., 1996). Layers of thermally altered soil and ash lying between lava flows as well as a 3-m thick soil layer have been encountered in a drill hole near Puna, Hawaii (THOMAS, personal comm., 1997). Loosely consolidated sediments and a'a lava flows are structurally weak and provide low friction surfaces (and ball-hearing-like materials) which facilitate gravity-driven slope failures.
MARK and MOORE (1987) suggest a marked change in slope is evident at the subaerial-submarine zone because the rapid chilling effect of water increases the flow's effective viscosity, shortening and thickening flow lobes and causing flows to spread over broad fronts. They suggest that as a flow goes into the sea, its effective density drops by 1g/cubic cm, which results in lava piling up at the land-sea contact, creating an oversteepened flow front that soon fails. Lava is currently entering the sea on the southeast coast of Kilauea volcano where it has repeatedly built lava deltas that have collapsed into the sea. These collapses arc estimated to involve on the order of 5,000 cubic km of material (MOORE et al., 1989).
Examples of collapse of subaerial deposits include the subaerial landslides at Ritter volcano in Papua New Guinea (PNG) reported by JOHNSON (1987). KOKI:-EAAR and ROMAGNOLI (1995) review of the sedimentation and Blast population evolution observed in the collapsing slope deposits of Stromboli volcano. MooRe (1985) describes subaerial and submarine activity at Surtsey volcano, Iceland (from drill core material). MooRE and FISKE (1969), FORNARI et al. (1979), MOORE et al. (1990), and MOORE and CHADWICK (1995) describe subaerial and submarine gravity driven slope collapse features around the Hawaiian islands.
Submarine Deposits
The deposits found at shallow submarine depths off islands are prone to be poorly consolidated, highly porous, weakly cohesive, and highly susceptible to failure (JoNEs, 1995; MOORE and FISKE, 1969). The lack of mechanical cohesion and strength occurs for several reasons. Volcanism at depths of 500 1,000 m (KOKELAAR, 1986), often produces violent water/rock interactions that reduces the lava flow material to small fragments by thermal shock. Violent eruptions of Rumble IV and V Seamounts off New Zealand (WRIGHT, 1996) have produced submarine fire fountaining. At the shallow depths on the summit and upper flanks, the low****ining pressures of the overlying water allow lava vesicularity as high as 40-50% (MooRE, 1985). The eruptive activity in shallow marine waters (less than 500-1000 m) produces abundant fragmental debris, glassy hyaloclastites and fragmented pillow lavas, mixed with marine sediments (beach and reef), ash and cinder debris, all of low mechanical strength with the consistency of poorly sorted rubble (KOKELAAR, 1986, WRIGHT, 1996).
VINCENT (1995. 1997) pointed out the problems of structural stability of the "transitional zone" (i.e., the zone of subaerial and shallow submarine volcanism) at Mururoa Atoll. Subsurface drill cores indicate that the subaerial basalts have been altered to clays by hydrothermal alteration and deep weathering associated with a tropical climate. Drilling in the upper 500-800 m of the summit area of the atoll encountered carbonate bank deposits, pillow lava breccias overlain by the products of phreatomagmatic volcanism, and****lomeratic deposits of pebbles of varying igneous compositions in a matrix of clay. This****lomeratic material is overlain by a sequence of palcosols (GUILLE et al., 1993). These stratigraphic relations indicate that the deposits accumulated on a highly eroded and weathered volcanic substrate, which constitutes a potential failure surface.
Drill-core and seismic profile data from Mururoa shows that beds near the center of the volcano have dip slopes of 10 degrees or less for volcanic deposits and 2-3 degrees in the sedimentary formations. Near the top of the volcanic formations and near the atoll rim, beds dip more steeply (15 to 25 degrees). Vitric tuffs situated deeper in the volcano also dip seaward. These multiple outward dipping horizons combined with the weakly consolidated materials (breccias,****lomerates, ashes, tuffs, and clays) provide the ideal setting for low friction slip surfaces to act as decoupling surfaces for gravitational collapses of the summit region (VINCENT, 1997).
Guii,LE et al. (1993) found****ave amphitheaters mapped on the submarine hanks of Mururoa and interpreted them as "former valleys draining the volcano or slump structures, which are commonly observed on islands formed in a similar geodynamic setting (Hawaii and Reunion Islands). These structures have provided channels for considerable transport of sediments that were accumulated on the bathyal slope at depths between - 2,000 and - 3,000 m... as well as on the deep ocean floor (below 3,000 m), especially between Mururoa and Fangataufa." In addition, they report that shallow submarine volcanic rocks, characterized by the
presence of highly brecciated lavas, between 550 and 600 ni, except at Viviane
(south side of atoll) where the same unit abruptly occurs at -- 700 to 900 m. "In this sector, the large volume of subaerial volcanic rocks, their very deep position, buried beneath more than 300 m of submarine volcanic rocks, as well as the pronounced****avity of the coral rim, may be due to caldera-type collapse or even an ancient landslide on the slopes of the volcano." Guit.LE et al. (1993) postulate that landsliding occurred during the period of emergence of the subaerial volcano about 11.8 Ma. BOUCHEZ and LECOMTE (1995a,ó) report these submarine valley deposits are subject to debris avalanches and have experienced natural slides over geologic time.
Karst Features
ASHMORE (1973) found karst structures on the carbonate platform of Johnston Atoll (in the northern Line Islands, Central Pacific Basin). Subsequently KEATING (1987) using a submersible, explored the flanks of the Johnston carbonate bank and found caves with intact stalactites, stalagmites, drip pool rims, and tables (Fig. 10). These karst features are clearly the products of subaerial dissolution during times of emergence, related to ancient changes of sea level. Only a small fraction of the Johnston carbonate cap currently remains emergent.
Karst features documented by swath mapping of the Mid-Pacific Seamounts are widespread (VAN WAASBERGEN and WINTERER, 1993; WINTERER et al., 1995; MANO (personal comm., 1995). Studies by BUIGUES (1982) on Mururoa Atoll, using rocks collected from drill cores, revealed several zones of mega-porosity around 330 m depth, 280 and 290 m, 230 and 260 m, 120 and 150 m, and around
Figure 10
Sketch of one of two large caves discovered at depths of 400 m below sea level on the margins of Johnston Atoll (in the central Pacific, Northern Line Islands chain). A series of stalactites and stalagmites were observed as well as pedestals and pool-rim features. This drawing was made by W. W.Sager.
90 m. The mega-porosity was interpreted as resulting from subaerial weathering of exposed limestones, with likely development of cave systems. The presence of karst features such as caves and river channels in the limestone cap overlying the volcanic basement at numerous atolls indicates that solution plays an important role in removing carbonate material, leaving large void spaces (mega-porosity) and reducing the mechanical strength of the carbonate cap portion of volcanic islands and facilitating gravity failures of the summit cap.
Fracturing
The carbonate banks associated with large volcanic islands as well as the smaller atolls display abundant fracturing. These fractures appear to mark zones of weakness where failures have taken place and are likely to fail again. In the central Pacific, KEATING (1997b, 1985) reported many rectangular cavities within the flanks of Johnston Island where fractures intersected and house-size blocks of detached carbonate rocks left voids in the flanks of the carbonate atoll. KEATING also described the occurrence of dissolution caves, water discharge, intersecting fracture patterns, fracture cavities, and slump blocks. Swath mapping of the southern margin of Johnson Atoll found a long, steep vertical scarp truncating the southern margin of the atoll, which is interpreted as a landslide headwall. At the toe of the scarp was a debris pile over 700 m thick (KEATING, 1994a), including limestone blocks up to 1 km across.
Chapman Seamount (in the Line Islands Chain, Kiribati, Central Pacific Ocean) shows evidence of a large submarine landslide (Fig. 11) that took place after the development of a carbonate reef cap (KEATING et al. 1991; KEATING, 1998). Since the Chapman landslide involved both the upper portion of the volcanic edifice and the overlying reef hank, the failure took place long after the volcanic edifice sank below sea level (KEATING el al., 1991). A similar submarine landslide was observed on Horizon Guyot within the Marshall Islands (Bh:RGERSEN, 1995).
On Niue Island, near the Tongan Islands in the SW Pacific Ocean, a remarkable example of surface fracturing is evident. The atoll has a raised peripheral reef about 60 m above modern sea level (FoRBes, personal comm., 1996). Magnetic studies indicate that a large section of the SW flank has been removed, either blown away as a result of explosive volcanism or detached and moved as a giant landslide (HILL, 1996). NUNN (personal comm., 1998) suggests Alofi Bay is probably the site of an old landslide. SCHOFIELD (1959) reported that large areas of the western
Figure 11
West Chapman guyot in the southern Line Islands chain (Kiribati) has been surveyed using the SeaMARC 11 side-scan sonar system. A large landslide was found on the south side of the seamount. Seismic profiling establishes that the landslide took with it a large piece of the carbonate cap. This guyot has been volcanically reactivated after the landslide.
slopes of the Niue appear to have been removed, leaving the unusual coastal****iguration, the eccentric position of the center of the caldera, and the large chasm within the limestone platform. Schofield's report includes a photograph taken from a boat between the limestone walls of the chasm. NUNN (personal comm., 1999) suggests the chasm may represent an old line of weakness, and any impending slide may occur along the same zone.
KAYEN el al. (1989) report slumping and sedimentary gravity flows on Horizon Guyot (Central Pacific Ocean) of the sediment cap at dip slope angles of only 1.6 to 2.0 degrees. The slump blocks disrupt the surface, and extend to sub-bottom depths of - 10 to - 46 m. They concluded that current-induced beveling combined with infrequent earthquake loading was responsible for the initiation of slumps. Slope failure was also observed between 1,600 and 1,850 m depth on the perimeter of the same atoll (I IHN et al., I985a and b) where gravity flows traveled (probably at high speed) long distances down the guyot flank. DADE and HUPPERT (1998) discuss the nature of long-runout debris avalanches. Ocean drilling of the rim of Resolution Guyot in the Mid-Pacific Mountain found that the periphery of the guyot consisted of lagoonal facies sediments (WINTERER and SAGER, 1995). SAGER (personal comm., 1998) suggested that landslides around the periphery of the guyot may have played an important role in removing the atoll rim, leaving only the central drowned lagoon. Seismic reflection studies of the carbonate bank off Oahu, Hawaii, show that the submarine bank is extensively fractured (GREGORY and KROFNKF, 1982).
Steep Slopes
SIPBERu (1984) found a correlation between slope angle and the frequency of major slope failures on Quaternary volcanoes. A similar relationship is noted by MURRAY and VOIGHT (1996). Higher cliffs retreat more rapidly because they generate higher shear stresses and suffer larger landslides (RICHARDS and LORRIMAN, 1987). Dramatic slope failures occur most often where the slope angle exceeds twenty degrees.
Within the Hawaiian volcanic chain the mean slope angles were calculated for the submarine flanks, using available seismic profiles (KEATING, 1994h), the slope angles ranged from 6 to 11 degrees. Seamounts formed by non-hotspot processes have consistently greater slope angles (8 to 12 degrees). MARK and MOORE (1987) have also reported on slope angles around the Hawaiian Islands, using gridded sounding data. The MARK and MOORE (1987) and K1AIING (1994b) measurements show that surface roughness increases with increasing slope angle. In examining areas associated with the heads of submarine landslides, MARK and MOORE (1987) report that the slopes adjacent to the landslide areas exceed 19 degrees. LENAT el al. (1989) report the submarine slopes of Reunion Island reach peak values of 24 degrees between depths of 500 m and - 100 m. On average the sslopes reach 15 degrees at these depths, depending on the nature of the shore (valley, caldera rims, etc.). ROWLAND (1996) reports slopes of 6-12 degree for portions of the subaerial portions of Volcan Fernandina, Galapagos Islands, with generalized slopes less than 20 degrees (Fig. 12).
LABAZUY (1996) has suggested that rapid growth of volcanoes like Kilauea, and asymmetric growth such as Reunion, can cause oversteepening of flanks and induce slope failure. DUFFIELD et al. (1982) describe landslide blocks and the growth of Piton de la Fournaise on Reunion Island (Figs. 12 and 13) and CIILVALLILR and BACHFLFRY (1981) discuss the structural evolution of Reunion Island. Major edifice collapses in the Canary Islands appear to be an order of magnitude smaller than those of Hawaii (hundreds rather than thousands of cubic kilometers, MCGUIRF, 1996).
Sea Level Change
Early studies of the processes that modify island morphology, particularly on the role of sea level change, are numerous and include DALY (1924), DAVIS (1926. 1928), HOII-MEISTER and LADD (1944), COTTON (1969), MENARD (1969, 1983. 1984) among others. NUNN (1994) summarizes many of these studies involving oceanic islands. Studies of sea level change have continued throughout the century, and have been a major emphasis in earth sciences in the last decade. Recent observational studies and modeling of the geoid, suggest the major redistribution of planetary water during the glacial-interglacial alternations do not produce uniform records of global sea level rise and fall, since the crust does not deform uniformly. Eustatic sea level variation tends to produce records of a more regional significance.
Studies of continental shelf regions suggest that there is increased seismicity, due to crustal loading or unloading, providing a trigger for the lateral collapse of already destabilized edifices, e.g., McGUIRF. et al. (1997). Sea level highs lead to shedding of skeletal sands from the inner shelf, producing coarsening and thickening upward successions of turbidites (VFCSEI and SANDERS, 1997). Rising sea level will offset the dissipative effects of a widening shore platform with rapid cliff retreat rates (TRENIIAILE and BRYNF, 1986. and TRENHAILE. 1989). As sea level falls, excess pore pressures can be created in sediments under impermeable or semi-permeable layers, when the pressure gradient in the sediments fail to equilibrate with hydrostatic pressure. Streams will incise older deposits and rivers will deposit abundant sediment on the incised surface (BARNIIARDI el al., 1997). Additionally, erosion of the newly exposed island summits (and flanks) will take place.
VORREN el al. (1991) points out that erosion during glaciations led to rapid sedimentation and progradation. This rapid sedimentation has long been recognized as a contributor to landslides on continental margins (MORTON, 1993). These same patterns of oscillating sea level and sediment delivery to the sea occur on the flanks of oceanic volcanoes. A fall in glacio-eustatic sea level lowers bottom water temperature and, when coupled with high sedimentation rate, produces excess pore pressures in sediments on carbonate platforms and seamount slopes. Flanks loaded with low-cohesion detritus, fail when earthquakes cause dynamic loading (LABIA-2G and VORREN, 1993). NISBLT and PIPER (1998) and WEAVER and KuuPERs (1983) suggest sediment landslides are most common close to glacial sea level minimum.
Large changes in global sea levels reaching 130 m have occurred over the last 18,000 years, with sudden rises of 11.5 m in less than 160 years ( i 50 years according to BI.ANCHON and SHAW, 1995). These sea level changes modify stress
Figure 12
Relief map of Hawaii, the Galapagos Islands, and Reunion Island illustrating the elevation in 1,000 in concours. Boundaries between volcanoes are drawn with white dashed lines. Volcano rift zones are drawn with solid black lines. Figure 12A The relief map of the Island of Hawaii is based upon the lISGS DEM topographic database which has pseudo-illumination from the north. The rift zones on the south side of Hawaii mark the zones of active dyke intrusions for the Kilauea volcano. The flank of Kilauea south of the rift zones moves downward and southward (the crater is shown as a small white circle, adjacent to the 1,000 in concour).
Figure 12B (continued)
The relief map of the West Galapagos Islands is based upon the TOPSAR single-pass interferometric radar database with pseudo-illumination from the east. It shows that the volcanoes in the Galapagos chain arc largely intact. This island chain is situated on relatively young sea floor with little sediment underlying the volcanoes, which appears to be an important factor effecting volcanic stability. Figure I2CShaded relief map of Reunion Island (Indian Ocean) based upon Shuttle Imaging Radar-C
(SIR-C') database and digitized concours (from October 1994 mission). This margin of the Piton de Fournaise volcano (between the eastern rift zones marked by black lines) at the southeast corner of the island is moving downward and away from the caldera. Sonar imaging of the sea floor adjacent to this collapse feature documents the giant volcanic landslide and debris field (e.g., LABAZUY, 1996; Dr irrtet.o et al., 1982; I.uNAT Cl al., 1990). These images were provided by Scott Rowland.
regimes and water pore pressures, resulting in edifice destabilization, particularly on an island with bare platforms and on those composed of soft volcanic materials such as ash (SUNAMURA, 1988). continued oscillations of sea level will fill and bury old fault scars and perpetuate more sliding on the rubble piles derived from prior slides (MORTON, 1993).
Loading Effects
ScHMINCKE (1995) suggests that small magnitude variations in sea level may affect shallow magma chambers, whereas large magnitude falls of sea level fall can affect the stresses at greater depths, perhaps in the lithosphere. Relative changes in sea level, such as the 100 m suggested by WALLMANN et al. (1988) would have the
Figure 13
NASA Space Shuttle Phototgraph of Reunion volcano (Indian Ocean). The volcanic crater and associated collapsed margin are seen on the southeastern corner of the island. This image was provided by Chuck Wood courtesy of the Volcano Website and NASA.
potential to initiate stress changes both in oceanic and continental volcanoes. WALLMANN et al. (1988) suggests that eruptions at Patelleria Volcano in the Strait of Sicily at about 67 Ka occurred during a major low stand of sea level and the most recent episode of volcanism took place about 20 Ka during the last glacial maximum.
GUDMUNDSSON (1983) and others have suggested glacial retreat triggered basaltic eruptions in Iceland. WAmEMANN et al. (1988) relate the volcanism to stress changes associated with bending of the underlying plate associated with a 100 m fall of sea level. They found that for some magma chambers, the magnitude of the tensile stress maximum is of the correct order of magnitude for dyke propagation to occur along fractures inboard (inside) the chamber margin. Others have suggested that loading of the periphery of the volcanic edifice would affect pore pressures and prompt additional structural failures.
McGuIRF et al. (1997) find evidence for increased explosive volcanism coincident with more rapid changes in sea level. The rate of change, rather than absolute sea level appears to be most important, with rapid sea level draw down promoting eruption at island volcanoes, through a reduction in the radial compressive stresses.
Weathering
Another environmental factor of importance in edifice failure is precipitation. With increased precipitation, tropical weathering will increase the alteration of rocks to clays, increase pore pressures in rocks, increase the amount of detritus derived from the volcano and moved offshore and increase the loading the Hawaiian Island slopes with low cohesion materials. GAVENDA (1992) suggested that greater rainfall during glacial periods produces more highly weathered soils and increased sedimentation. Gavenda suggests that interglacials would have produced drier conditions at low elevations, however they would probably produce greater precipitation at higher elevations based on the modeling that assumes predominant tradewind rainfall and estimates of rainfall based on glaciocustatic-related elevation changes and distances from mountain summits (GAVENDA, personal comm., 1999). (in some parts of the Pacific this pattern may not apply due to differences in wind patterns, ocean circulation, etc. according to NUNN, personal comm., 1999).
Because many mass movements associated with high precipitation begin high on the volcanic slope, the energy for post-failure transport of the detached material is great enough to exceed the obstruction of the bathymetric arch surrounding the Hawaiian Islands (MucK and UNDERWOOD, 1990). These large debris avalanches move detritus great distances from the slopes and often erode the abyssal sea floor (GARCIA and HuLt., 1994: GARCIA, 1996).
Overloaded Slopes
The subaerial amphitheaters often occur in portions of the islands characterized by the highest precipitation and most rugged topography in which erosion is most intense. Since increased precipitation strongly changes pore pressures as a result of increased groundwater recharge, decreased slope stability and gravitational sliding results. These factors act together to load debris on the volcanic slopes. Storm surges can then help to move the material far offshore, loading the submarine slopes where landslides and slumps can subsequently transport the debris (characterized by water-saturated materials with very low cohesion, TRIBBLE, 1991) even further.
MOORE et al. (1994a) and NORMARK et al. (1993a) observed that submarine avalanches on oceanic volcanoes generally have a well-defined amphitheater at their head. (These amphitheaters arc described as u-shaped valleys and horseshoe shaped valleys, by other authors.) The amphitheaters are often adjacent to areas of highest concentration of precipitation on the island. Marine terraces within the amphitheaters indicate the heads of these amphitheaters were originally formed subaerially and have subsequently drowned. The horseshoe-shaped valleys in the subaerial portion of the avalanches are better developed than those below sea level.
During 1953, submarine landslides reportedly cut undersea cables in the Samoa island chain, and a 2 m tsunami wave was observed (Sol,oviev and Go, 1984). Most likely this landslide resulted from the loading of the slopes rather than volcanism. Submarine volcanism was last recorded in 1866 (THOMsoN, 1921; KEATING, 1992a) and subaerial volcanism occurred between 1905 and 1911.
Eruptions and Seismic Shaking
Among the most obvious triggering mechanisms for landslides are eruptions and strong volcanogenic earthquakes and associated displacements. SEED (1968) reports historic landslides associated with seismic activity at volcanoes documented as dated as 373 B.C. KEEFER (1984) reviews 40 strong earthquakes which generated landslides during the years 1811 to 1980 and reports that most of these occur in materials which had not previously failed. DUDLFV and LEI- (1988) summarize the earthquake related landslides and tsunamis for Hawaii, and EISSLER and KANAMORI (1986) review large earthquakes on Hawaii since 1940. The Hawaiian Islands experience frequent tsunamigenic earthquakes (WALKER, 1994). During this century, loss of life has resulted from both distant earthquake tsunamis (with runups of up to 16 m) and from locally generated tsunamis. Twelve tsunamis locally generated by earthquakes have produced runups to 14 m (Cox and MORGAN, 1977). Reports of Hawaiian tsunamis of "uncertain origin" (unwitnessed and/or undetected trigger events) are common (Cox and MORGAN, 1977) and many may be related to submarine landslides (WALKER, personal comm., 1995). Tsunamis on the island of Kosrae (Federated States of Micronesia) were recorded in 1890 (with many people killed) and in 1985. The later tsunami was not associated with an earthquake and may have been locally generated by a submarine landslide (WALKER and MCCREERY, 1988).
The observations related to locally generated tsunamis are of particular importance since even a small land slump such as that on Ilawaii in 1975 can produce one of the largest earthquakes felt on the island, resulting in deaths and very extensive property damage (DUDLEY and LEE, 1988). In 1868, movements on slumps on the Island of Hawaii caused a 7.8 earthquake, 16 m tsunami, one million dollars of damage and destroyed several Hawaiian settlements, taking the lives of 46 people. In 1919, a submarine landslide off the Kona coast (Island of Hawaii) was associated with a '7.5 magnitude earthquake and a 5 m tsunami that carried one woman out to sea. In 1951, two separate tsunamis occurred in one day at Napoopoo (on the Island of Hawaii). The first was associated with an earthquake, two hours later when a large piece of the steep cliff at Kcalakakua Bay fell into the sea a second local tsunami was generated. In 1975, 60 kin of Kilauea's south coast subsided 3.5 m and moved seaward 8 m, producing a 10 m tsunami that caught 34 people in a campground, killing two and injuring others. The earthquake and tsunami caused four million dollars of damage. DuD1.EY and Eel, (1988) report "since 1948, there has been no period longer than 35 years without a local tsunami in Hawaii."
During the 1907 eruption of Matavanu Volcano on Savait, in the Samoan island group, a local tsunami was generated with waves traveling 90 110 m inland (LANDER and LOCKRIDGE, 1989). ANDERSON (1910) reports that the largest tsunami generated by the Matavanu eruptive activity was a 3--3.6 m high tsunami.
Eruptions of Mt. St. Augustine volcano (Alaskan Peninsula) have repeatedly coincided with volcanic edifice collapses (KIENLE et al., 1987). BEGET and KIENLE (1992) recognize at least 11 major debris avalanches at Augustine that have followed reconstructive dome-growth oversteepening, separated by an average time interval of only 150 200 years. Such a short repeat time for edifice failure is unusual and may pose a hazard to the local settlements around Cook Inlet from volcanogenic tsunamis (WAYTHOMAS, this issue). BEGET and KIENLE (1992) have correlated an 1883 debris avalanche at Augustine volcano with a tsunami reported at English Bay, approximately 80 kin to the south. The 1883 avalanche (Fig. 14) extended 9 km from the summit and traveled 3- 8 km beyond the coastline, extending the shoreline by 2 km. Wave heights of 7 to 9 m were reported at English Bay which carried off fishing boats and deluged houses, although the tsunami occurred at low tide.
The historic records indicate that the waves traveled at a very slow rate and reached the English Bay community with the initial wave described as a 6-9 m high wall of water. Numerous debris avalanches that reach the sea are also reported at other Aleutian Arc volcanoes (MILLER et al., 1998). SIEBERT et al. (1989) suggest a similar volcanogenic tsunami event will occur sometime during the next century. WAYTHOMAS (this issue) describes the distribution of tsunami deposits around Cook Inlet. The observations of SIEBE:RT et al. (1989) indicate that in the ocean island framework even small debris avalanches (Fig. 15) of low volume can generate very hazardous local tsunamis. Waythomas, however, finds very little evidence in the geologic record that a hazardous tsunami occurred.
In 1888, the Ritter Island volcano (PNG) experienced a tsunamigenic**** collapse accompanied by eruptive activity, and now only a small portion of the island remains above sea level (JOHNSON, 1987). Johnson suggests that tilting of the sea floor beneath Ritter volcano probably contributed to the slope failure. He compares the slope collapse to that at ML Mavuvama****, Unten volcano (Kyushu. Japan) where a small slope collapse of only 0.48 cubic km sent an avalanche to the sea. generating a tsunami that caused 15,001) deaths (Stint It , 1984). Johnson also notes that numerous small avalanches may contribute to massive structural failure but not produce a single amphitheater which dominates the morphology. Johnson points out that Bamus and IJlawun volcanoes in the Bismarck volcanic arc have prehistoric gravity slides that may have been tsunami-genic and suggests that Ulawun volcano in the Bismarck volcanic arc is a prime candidate for further catastrophic slope failure. Local legend indicates that about ten generations ago the side of the Donw Peaks volcano (Fly-h ighland.s Province) erupted water and rock from a large crater lake. This is consistent with hydrothermal undermining. avalanches and lahars. AI_t_Frv and Wool) (1980) suggest that an event took place 100 120 years ago, in which the Doma Peak volcano experienced numerous Collapses.
Figure 14
Aerial photograph of the central portion of the 1883 Burr Point debris avalanche at Ml. St. Augustine volcano_ Alaska. The photograph was taken from an northern vantage point looking to the south.
Figure 15
Aerial photograph of the distal portion of the West Augustine Island debris avalanche at Mt, St. Augustine solcano, Alaska. The photograph was taken fout a small aircraft, viewing the volcano in a southct a said direction
Hagen volcano (also in the FIs-llighlands Province) is strongly asymmetrical
 |
|
|
|
| |
|
|
