by Acad. Nikolai LAVEROV, RAS Vice-President; Leopold LOBKOVSKY, Dr. Sc. (Phys. & Math.), Boris BARANOV, Cand. Sc. (Geol. & Mineral.), Shirshov Institute of Oceanology, RAS; Raisa MAZOVA, Cand. Sc. (Phys. & Math.), Nizhni Novgorod State Technical University; Boris KARP, Cand. Sc. (Geol. & Mineral.), llychev Pacific Oceanological Institute, RAS Far Eastern Branch

In the wake of a disaster that hit Sumatra in December 2004 the world scientific community became conscious of the inadequacy of the present ideas about the most probable sites of great earthquakes and tsunami waves. The long "lull" in the northern stretch of the Sunda island arc came as a warning about the forthcoming natural calamity. That is why the present "silence" in the central part of the Kurile-Kamchatka arc may be a bad omen and calls for non-stop monitoring so as to come up with a reliable intermediate-term forecast for several years.

стр. 4

Three great earthquakes occurred along the boundary of the Indo-Australian and Eurasian plates from December 2004 to July 2005. Their epicenters are marked by asterisks, source areas (200 - 1,300 km) - by yellow ovals, plate boundaries-by red lines; the direction of the Indo-Australian plate's movement relative to the Sunda subduction zone-by arrows.

The northern tip of the Island of Sumatra, which is a part of the Sunda island arc (festoon islands) of the East Indies, or Malay Archipelago, and the adjacent area became a scene of a disastrous earthquake on December 26, 2004*. The initial shock was tremendous indeed, so much so that thousands of kilometers away from the epicenter the surface of Sri Lanka (south of the Indian subcontinent) felt the impact-tremors with an amplitude of more than 9 cm continued for sixty seconds, while 1 cm displacements caused by long-period surface seismic waves were registered on the entire surface of our planet. The amount of energy liberated thereby happened to be equal to the cumulative energy of all earthquakes in these last 15 years. A destructive tsunami hit the shores of Indonesia, Thailand and Bangladesh killing over 280,000. The tsunami tidal wave swept the globe three times to be registered by mareo-graphs on the coasts of all continents. In its magnitude (estimated at 9 to 9.3 points) the Sumatra-Andaman earthquake is assessed second in power ever since the start of instrumental observations in 1887.**

The present system of observations had never registered a great earthquake like that before. It comprises a global network of seismic stations operating in a wide frequency range. This system draws on data supplied by orbital satellites whereby it registers a tsunami wave out in the ocean and measures a rise in the ocean's level. It relies on a network of what we call global positioning (Global Positioning System, GPS) that can spot even insignificant displacements of the terrestrial surface; also, it includes a world network of mareographs registering sea surface variations on shore. Data on the Sumatra earthquake were computer-processed in real time and even ahead of it. Thus, the prognostic map for the propagation of the tsunami drawn by Vassily Titov of the National Oceanic and Atmospheric Administration (Seattle, USA) had appeared on the Internet even before the tsunami wave circumnavigated the globe within twenty-four hours.

At the suggestion of Th. Lay of the University of California (Santa Cruz, USA) it was decided to pool efforts of various groups of seismologists for an in-depth study of what had happened off Sumatra so as to exclude rivalry among them and work out an objective approach. Taking part were 40 seismologists from 23 universities and colleges of seven countries. Already five months after the earthquake the journal Science (May 2005) carried the first articles on the subject. Today dozens of publications have been devoted to the December 2004 disaster. These publications report, among other things, research findings on the sea floor in the rupture area. Research parties dispatched to the area of the Sunda Deep (Trench) immediately after the seismic shock and after (such parties are still there) obtained evidence on the extraordinary nature of the event, the worst over the past forty years.

IN THE WAKE OF A LONG SEISMIC "SILENCE"

The Indo-Australian plate is traveling at a rate of 4 - 5 cm/year submerging under the Sunda island arc of the

* See: "Aftereffects of Seismic Catastrophes", Science in Russia, No. 1, 2006. - Ed.

** An earthquake is classified as great if its magnitude (M) >7.7 points. - Auth.

стр. 5

Source areas of violent earthquakes of the Kurile-Kamchatka island arc and the position of seismic gaps: Possible sites of future earthquakes indicated yellow; the most probable loci of future earthquakes (forecast by Acad. Fedotov) - red; trench axis - by dotted line.

East Indies (Malay Archipelago). This gives rise to a subduction zone along the boundary between the Indo-Australian and the southeastern part of the Eurasian plate (the latter being segmented into the Burmese, Sunda and Andaman microplates). Stresses in the contact zone keep rising to cause and ruptures when the stress value exceeds the fracture strength of the rock. The earthquake of December 2004 occurred in the Sunda subduction zone.

The main shock was localized in the ocean southwest of the northern extremity of the island of Sumatra in the zone of contact between lithospheric plates at a depth of 30 km. Thereupon thousands of tremors (aftershocks) were registered from Sumatra's northern edge up northwest across the Andaman and Nicobar Islands and then toward the southern coast of Myanma. The most remarkable series of shocks occurred in the Andaman Sea, with 150 aftershocks of M>5, in a rather peculiar sequence compared with other great earthquakes. It thus became possible to delineate the source area of the event over nearly 0.25 mn. sq. km stretching from the southeast to the northwest for 1,300 km.

The maximal displacement of tectonic plates relative to one another was found to be about 20 m, and it was centered mainly in the southernmost initial part of the 500 km-long rupture wherefrom it traveled at a very high velocity. Farther north the process slowed down abruptly: it took half an hour for the displacement of the Indian plate with respect to the Andaman Isles to reach 7 - 20 m. Otherwise tsunami waves would have come up along the entire 1,300 km-long zone of ruptures to cause ever heavier destructions on the coasts of India, Thailand and Myanma.

Right after the earthquake seismologists began studying the structure of the sea floor in the region of greatest displacements. Early in 2005 a team of British explorers on board the HMS Scott hydrographic research ship made a bathymetric survey with the use of a multibeam echosounder and obtained a map showing distinct rock-slide structures thought to have triggered the tsunami wave.

However, subsequent investigations carried out by an international expedition aboard the vessel Performance did not confirm that supposition. Using a remote-controlled bathyscaph for surveying a sea floor to a depth of 6 km, seismologists saw that the slides were much older in age and hence could not precipitate the seismic event of December 2004. Yet in the lower part of the island arc slope one detected several scarps about 5 km deep under as if cut off with a knife and rising to 12 meters.

The presence of gouges and fissures on dry land is a typical token of strong underground shocks. But such structures were observed on the oceanic bottom for the first time. In February and March 2005 the research ship Natsushima of Japan likewise detected disjunctive dislocations, though not as deep-about 2 km or so. The ROV Hyper-Dolphin, with a diving depth to 3 km was employed for closer studies.

From October 2005 to August 2006 the research vessel Sonne of Germany took six research parties to the site of the earthquake; they mapped the relief features of the ocean floor and, using benthic seismographs, carried out multipath seismic profiling and studied local seismicity. (One of the authors of the present article, Boris Baranov, took part in an expedition that went thither in the beginning of 2006.) A tsunami early warning system

стр. 6

composed of two watch buoys was set up in the ocean south of the Sunda arc. In the event of an earthquake they were to get information from OBS and pressure sensors on the sea floor and then transmit it to the Tsunami Early Warning System (TEWS) and mobile telephones of private citizens. Unfortunately this warming system did not perform as it should have, for in less than a year after its installation a great earthquake rushed a tsunami wave to Java.

As noted by many seismologists, earthquakes in the subduction zones often group together. This means that some time afterwards (from several months to several years) a great seismic shock in one of the segments of such zones is followed by a shock of compatible force in another segment. Such a pattern of space-and-time dependence is believed to be due to the transfer of stresses along the plates boundary.

The Sumatra-Andaman earthquake was not exceptional at all, for the subsequent eighteen months saw two other violent shocks. The first one of M=8.7 occurred on March 28, 2005-its hypocenter was located at a depth of 30 km under the Banyak Island, roughly about 120 km southeast of the epicenter of the earthquake of December 26, 2004. A 400 km-long rupture came to be ripped open in the same direction. This event caused no tsunami wave because, as it was thought, the region of principal displacements was not confined to the sea floor but took in the chain of islands of the outer (nonvol-canic) arc. None the less houses collapsed on the island Nias, killing more than 1000. The other seismic event occurred on the 17th July of 2006, with the magnitude of the shock equal to 7.7. Its hypocenter was located at a depth of 10 km south of Java. Judging by the distribution of aftershocks, the rupture was about 200 km long. The earthquake and the tsunami wave that broke upon Java claimed over 5000 victims. This outburst of seismic activity with its peak on the northern segments of the arc came in the wake of a long seismic lull.

We should also mention a violent earthquake (M=7.6) that hit Pakistan on October 7, 2005, killing over 80,000 and inflicting immense material damage. Even though it occurred rather far from the focus of the Sumatra-Andaman earthquake, both events were confined to the plates boundary along which stress had traveled.

It's a fact that the frontal regions of subduction zones are "filled" with many sources of violent earthquakes within 50 - 100 years. But some of their sections remain vacant. These are seismic gaps-the most probable sites of the upcoming disastrous earthquakes and tsunami waves. We know that segments of the insular slope opposite the central and southern parts of Sumatra and close to Java were "packed" tight with sources of violent earthquakes. But no data on such seismic events opposite the northern part of Sumatra were available, and the northern sector of the island arc, the Andaman and Nicobar islands including, was relatively quiet. Even though the northern stretch of the Sunda arc was thought to be a seismic gap, no one allowed for a possibility of a seismic event of a magnitude like that.

This standpoint had to be revised after the Sumatra-Andaman earthquake. According to Roger Bilham of the University of Colorado, USA, seismologists came to be aware that their prognoses are conservative in many ways, with only recent history taken into account for plate boundaries, and more distant periods of time skipped (when such extreme events could have occurred). The catastrophe of December 2004 came as a warning that such prognoses could not be placed at the service of society. This lesson is all-important for the Russian Far East, too, where there is a long-living seismic gap in the central part of the Kurile - Kamchatka island arc, and its history still obscure.

CENTRAL KURILE SEISMIC GAP

Proceeding from the dataset on the Kurile-Kamchatka island arc, Acad. Sergei Fedotov of the Institute of Volcanology (Far Eastern Branch of Russian Academy of Sciences) was the first to suggest a mode of identifying the probable sites of forthcoming earthquakes. Since the source areas of most violent seismic events do not overlap, he pointed at seismic gaps situated in the

Extention zone in the central segment of the Kurile-Kamchatka arc limited by fault scarps (red lines and bergstrichs); the seismic gap between the sources of the violent earthquakes of 1952 and 1963 shown in yellow lines; deep-sea trench axis indicated by red.

стр. 7

Seismic cross section in the central part of the insular slope along the seismic gap's trend. From the southwest the extention zone is limited by a fault scarp rising to 4 km. The central part of the zone as well as the northeastern stretch of the Vityaz ridge is composed of several inclined blocks.

southwestern, northeastern and central parts of the arc. The first two gaps are about "full". But the Central Kurile gap, which includes the foci of the earthquakes of 1915 and 1918 as well as a section that experienced the latest powerful shock in 1780, is still "silent".

Acad. Fedotov calculated that the time of a seismic cycle, i.e. recurrence of a violent earthquake in one and the same region is 140+60 years for the Kurile-Kamchatka arc. Consequently, the probability of a seismic event in the Central Kurile gap is high. According to other researchers, tectonic dislocations here take place only on account of the creep (when, encountering no resistance, one tectonic plate glides upon another) and also due to displacements caused by slight and moderate earthquakes. This gap, therefore, is not capable of triggering violent earthquakes. However, as shown by the Sumatra-Andaman disaster, humankind pays a very dear price for erroneous estimates of the seismic potential in some segments of island arcs. A reliable forecast should be based on hard research evidence.

A great earthquake in the central part of the Kurile-Kamchatka arc is all the more dangerous because (as estimated by Yevgeni Kulikov of the Shirshov Institute of Oceanology, RAS) a tsunami wave could devastate both the Kuriles and the coast of the Sea of Okhotsk. It could also menace the neighboring countries-Japan (Hokkaido and Honshu) and the United States (the Hawaiian and Aleutian Islands).

Accordingly the Shirshov Institute of Oceanology (Russian Academy of Sciences) has come forward with an initiative project providing for research in the central part of the Kurile-Kamchatka island arc as a possible site of a catastrophic tsunamigenic earthquake; this project has been upheld by the Presidium of the Russian Academy of Science and the Government of the Russian Federation, while the Ministry of Education and Science of the Russian Federation has backed it financially. A sea expedition sent for exploring the frontal slope of the arc between the Urup (southwest) and Paramushir (northeast) islands was to find out why this particular seismic gap is a long-living one, how great its seismic potential is, and what the consequences of a violent earthquake and tsunami wave could be. Taking part in research works were teams of the Shirshov Institute of Oceanology and the Ilychev Pacific Oceanological Institute (Far Eastern Branch of the Russian Academy of Sciences).

The expedition Kuriles-2005 began its work in August-September 2005 aboard the research vessel AKADEMIK LAVRENTIEV. It explored the insular slope of a trench in the central sector of the Kurile-Kamchatka island arc, where the Pacific plate is underthrusting the Asian continent at an annual rate of 7 cm. Morphologically the arc here is in two parts: the inner, volcanic one (the Kurile Islands) and the outer, nonvolcanic above the submarine ridge Vityaz.

The structure of the frontal segment of the central Kuriles is anomalous for a subduction zone, and is characterized by a large extension structure. The Vityaz ridge is found to be divided into two parts-the southwestern and the northeastern-by fractures which can be interpreted as faults (subvertical or inclined fault with rock blocks displaced in extension). These faults, materialized in 2 - 4 km-high scarps in the bottom relief, confine the edges of a large asymmetric extention zone. The base of its central part, where the surface of the Vityaz is submerged to 3 km, consists of inclined (tilted) blocks typical of extension. The same blocks and grabens (fault troughs of earth's crust) have been detected in the northeastern and southeastern segments of the Vityaz ridge. A sum total of mag-

стр. 8

Generation and propagation of a tsunami wave in the Sea of Okhotsk from a source located in the middle of the Kuriles and composed of 8 blocks: A - uplift of blocks 1 - 8 to different elevation at different rates; B - distribution of wave heights in the Sea of Okhotsk (the highest level of wave run-up along Sakhalin's eastern seaboard attains to 8 m).

netic, seismic, gravimetric and bathymetric data makes it possible to estimate the size of blocks at 50 km or thereabouts, which is much less than the extension of a violent earthquake's source usually above 200 km.

The extension zone is in the shape of a triangle tapering towards the trench, and its width (the distance from the southwestern to the northeastern edge) in the middle of the slope makes up about 275 km. But counting in the entire region of the deformation, the figure will be up to 450 km, or about equal to the linear extension of the seismic gap.

The small dimensions of blocks are thought to be the main cause of the relative seismic quiet in the region, with no violent seismic events taking place so far: the tension built up by these structures is relieved by earthquakes of moderate magnitudes. But this is not to mean that the process will hold on. We know of instances when seismogenic blocks cluster together in subduction zones. In that case an earthquake source takes in several blocks like that, and this goes along with an increase in the magnitude and extension of the rupture. A similar scenario was realized during the Sumatra-Andaman earthquake where, by different estimates, the rupture zone encompassed from 5 to 14 blocks.

The Kamchatka earthquake of November 4, 1952 (M=9.0), was the most violent one within the Kurile-Kamchatka arc over the period of instrumental observations. It is in the list of 11 major seismic events on the globe since 1900. A tsunami wave it triggered wiped out the town of Severo-Kurilsk (on the Paramushir Island) and took a heavy toll of human life. The seismic gap and the source area of that earthquake concur in the rough. Should the focus of a probable seismic event cover the entire gap, including the sites of the 1915 and 1918 earthquakes, in magnitude it could be comparable to the Kamchatka shack of 1952.

The height of a tsunami wave is estimated to be linearly dependent on an earthquake's magnitude. Yet there may be anomalies-for one, due to the high steepness of the fault plane. According to our findings, normal faults are of wide occurrence in the Kurile arc's middle part. Their displacement planes are obviously steeper than with overthrusts characteristic of subduction zones. So this region has optimal conditions for the birth of exceptionally high tsunami waves. Since the blocks of an imminent violent earthquake are small in size, they might be in for significant vertical displacements and act upon the sea water above like pistons. Research findings of the expedition have been used for the identification of seismogenic blocks that make up the seismic gap and for a subsequent modeling of the possible propagation route of a tsunami wave.

A FUTURE EARTHQUAKE WILL TRIGGER A TSUNAMI IN THE SEA OF OKHOTSK

Thus and so, a great earthquake in the frontal part of the Kurile-Kamchatka arc will imperil the Kuriles, the Sea of Okhotsk coast and Sakhalin, an island where an oil and gas mining infrastructure is developing apace.* There were several instances of tsunami waves generated by earthquakes in the Kurile region. But the waves were not high enough to endanger the coastal structures. The point is that the spread of tsunami waves is hindered by a string of the Kurile Islands separating the Sea of Okhotsk from the tsunami source located within the inner slope of the trench. Only two deep seaways, the

* See: "Ecological Studies of the Sakhalin Shelf", Science in Russia, No. 2, 2005; N. Bogdanov. "Russian Shelf, Its Riches", Science in Russia, No. 4, 2003. - Ed.

стр. 9

Straits of Bussol and Krusenstern situated in the central part of the island arc, can act as natural waveguides.

We have made model calculations with the aim of assessing the tsunami threat both for the Kuriles and the Sea of Okhotsk coast; hereby we have taken account of the complex structure of the seismic source area composed of eight seismogenic blocks. And here's what we have found out: minutes after the birth of a tsunami wave part of it would rush into the high seas toward the Hawaiis. Its progress toward the Sea of Okhotsk will be stemmed by the chain of the Kuriles, a factor that will work to increase the wave run-up on the southeastern shores of the isles (over 15 m - on Simushir, 2 m - on Kunashir, 1.9 m - on Shikotan). Although a portion of the wave's energy will be bounced off toward the ocean, the brunt of the attack will go through the deep-sea straits of Krusenstern and Bussol.

After the Kurile Isles there will be two fronts of waves which will form a united front afterwards. The further advance of the tsunami wave depends much on the bathymetry of the water area: upon passing the Bussol strait some of it will travel along the Kurile basin toward Sakhalin and Hokkaido at a rather high velocity (-600 km/h) due to the significant depth of this basin.

The second wave front moving across the Kurile basin will slow down its advancement notably after the Krusenstern Strait because of its journey to the Deryugin basin, much less in depth, east of Sakhalin. The united front of the waves that have traversed the Krusenstern and Bussol Straits will reach the southeastern end of Sakhalin (Cape of Terpenya) in a matter of 50 to 80 min. As shown by our modeling data, the eastern shore of Sakhalin will be overrun primarily by the wave that has swept through the Krusenstern Strait. Traveling toward the northern shore of the Sea of Okhotsk, the left flank of the front will make consecutive attacks on Sakhalin's eastern coast. Their run-up may surge to 2.5 - 8 meters.

Sweeping clear of Sakhalin, the wave front will collapse to become nearly flat; spending its force, it will

стр. 10

keep moving at a moderate speed toward the town of Okhotsk on the northern coast of the Sea of Okhotsk. Yet its flanks will be moving faster westward, to the Shantarskie Islands, and eastward to the Gulf of Shelikov, and reach the coast in about three hours after the earthquake.

David Tappin of the British Geological Service, who has studied the causes of the Sumatra-Andaman earthquake and who has participated in two research parties to the area, said something like this when asked about the time of the next earthquake: It is bound to happen anyway-in a few days or months, we cannot tell... His prediction came true as yet another two seismic events occurred in the Malay Archipelago after the Sumatra-Andaman earthquake.

Yes, a violent earthquake will surely break out within the Kurile-Kamchatka island arc, this is a matter of time. The Central Kurile seismic gap will be one of the most probable regions. The long "lull" in this sector of the arc and the ensuing uncertainty about its potential and background relate to the specific structural characteristics of the insular slope. Yet, as shown by the Sumatra-Andaman earthquake, a long quiet of individual segments of island arcs may erupt into global catastrophes. This scenario cannot be excluded for the middle part of the Kurile arc either.

GEODYNAMIC MONITORING: FIRST TRIAL

Data collected by the Kurile-2005 Expedition and by the modeling of the tsunami wave propagation route make it clear that exploration of this area is a priority task of the federal program for the prognostication and prevention of acts of God. Such studies were carried aboard the Akademik Lavrentiev during its 41st voyage in August-September 2006 by research parties of the Institute of Oceanology and the Pacific Oceanological Institute within the framework of the project launched by the Russian Academy of Sciences providing for close studies into the seismicity and tectonic structure of the Central Kurile seismic gap with respect to a possible great earthquake.

This work is the first experience in a geodynamic monitoring of this segment of the Kurile island arc. Six ocean bottom seismographs were set up on the underwater ridge Vityaz near the Bussol graben for registering seismic activity in the region during 16 days. Hundreds of earthquakes were registered, at M = 1 and above.

Data processing is a laborious and time-consuming job. In the meantime Sergei Kovachev of the Institute of Oceanology has assessed the seismicity distribution pattern proceeding only from the evidence obtained during the first day of the monitoring. Yet an intriguing result is already available, namely concerning the possibal activation of the southwestern boundary of the seismic gap confined to the Bussol graben. The "revival" effect in transcurrent faults prior to a violent seismic shock has been observed in some of the festoon islands, e.g. in the central part of the Hellenic arc* (according to the evidence at the disposal of the Institute of Oceanology). Meanwhile a segment of the seismic gap northeast of the Bussol graben perked up soon after the expedition left the scene. Between September 26 and October 2006 it became an arena of 43 quakes at M>4 (by the data of the National Earthquake Information Center, USA).

All that attests to the need of a persistent monitoring of the seismic gap over many years by means of sea floor observatories equipped with seismographs and inclinometers** and using the data thus obtained for an intermediate-term forecast several years ahead. Such a forecast should be predicated on bits of evidence on processes antecedent to a seismic focus, namely: transformation of stresses, end phases of energy accumulation in seismogenic blocks and preliminary shocks, the fore shocks. A forecast like that is all-important, for even ten years is a brief time given the scope of preventive security measures. We are loath to strike a pessimistic note, of course; but an earthquake by our scenario and an ensuing tsunami wave over 8 meters high can wreck many oil- and gas-producting enterprises on Sakhalin's eastern seaboard and touch off an ecological disaster.

While the present article was being prepared for publication, the forecasts made by S. Fedotov and the authors of this very article started coming true: on November 15, 2006, a violent earthquake broke out next to the Kuriles. Fortunately it was a gentle-not steep/-slope that came to be ripped. A tsunami wave, therefore, traveled into the Pacific Ocean and spared the Sea of Okhotsk shoreline. Since the source area of the seismic event affected only the southwestern half of the central Kurile seismic gap, the seismologists' attention is now riveted to its northwestern half.

Illustrations supplied by the authors

* This island arc is located in the Mediterranean at the junction of the African continent and Eurasia. - Ed.

** Inclinometer (dip compass)-a geophysical device for determining the inclination and direction of a drill-hole. - Ed.