Chapter Two

Geology

Geology teaches us that the surface of the land and the distribution of land and water is everywhere slowly changing. It further teaches us that the forms of life which inhabit that surface have, during every period of which we possess any record, been also slowly changing.—WALLACE 1869

INTRODUCTION

The world changes, but to most of us, our Earth is the symbol of stability. Viewed from the standpoint of a human life span, and indeed generations of humans, the foundations of Earth remain unchanged. However, on a geological time scale, the earth has undergone radical transformations from its birth to present. These slow but ongoing changes are modifying the surface of the earth, reshaping the continents and ocean basins. More rapid changes are occurring on the earth's surface as a result of human activity. Examples of anthropogenic actions which have caused major modifications to the earth's topography include the construction of dams, open-pit mining, and rice cultivation. Flannery (1994) suggests that the 60,000 years of human occupation in Australia have impacted the Australian continent to such an extent that virtually all of the continent's ecosystems are in some sense man-made. In Indonesia, Javanese and Balinese rice cultivation have had a major impact on the ecology of those islands (see Whitten et al. 1996), as had open-pit mining on the islands of Sulawesi and Irian Jaya. But human impacts on the Indonesian Archipelago are the topic of a later chapter. Of interest to us now is the land beneath our feet, its birth and evolution. Several theories evolved during the 1900s leading to the now accepted theory of plate tectonics. However, prior to discussing plate tectonics and its role in shaping the Indonesian Archipelago, we will briefly review the structure of the earth.

STRUCTURE OF THE EARTH

The earth has a layered structure comprising the core, mantle, and crust (fig. 2.1). The core is divided into a solid inner core and a nickel-iron outer core. The outer core is liquid, and it is the circulation within the outer core that generates the earth's magnetic field. This point will be significant when we later discuss magnetic anomalies.

Figure 2.1. The earth's layered structure. Cross-section through one hemisphere of the earth showing the inner core (1), the outer core (2), the lower mantle (3), the upper mantle (4), and the crust (5). The inner core is approximately 2414 km in diameter. The outer core is 2253 km thick, as is the lower mantle, the upper mantle is roughly 644 km in thickness, and the crust varies from 3.2 km under parts of the oceans to 120 km thick beneath mountains. Figure not to scale.

The mantle is mostly solid, consisting of dark, heavy rocks which are rich in iron-magnesium silicates such as olivine and pyroxene (Rhodes 1991). The asthenosphere is part of the upper mantle. It is known as the zone of mobility, since it is "nearly molten" (Gross 1990) with convection movements and isostatic adjustments occurring within this layer. Magmas (molten rock) may also be generated (Bates and Jackson 1980). The lithosphere includes the crust and part of the upper mantle. It is 50-100 km thick. The lithosphere floats on the asthenosphere.

The crust, or what we know as the earth's "surface", forms die upper portion of the lithosphere. There are two types of crust: continental and oceanic. Continental crust is generally 30-40 km thick, but may be 120 km thick beneath mountains and is rich in aluminum and silica. Oceanic crust is 3-7 km thick and rich in magnesium and iron. Continental crust is thicker but less dense than oceanic crust, so it floats higher in the asthenosphere than oceanic crust. This is an example of isostasy, an equilibrium condition comparable to floating. However, there are two different concepts of the mechanism of isostasy: the Airy hypothesis and the Pratt hypothesis (fig. 2.2).

The Airy hypothesis postulates an equilibrium of crustal blocks of the same density but of different thickness (i.e., topographically higher mountains are of the same density as other crustal blocks, but have greater mass and deeper roots) (Bates and Jackson 1980). This is an example of floatational equilibrium (Spencer 1972) and can be compared to blocks of wood floating in water.

Figure 2.2. Isostasy: The Airy hypothesis and the Pratt hypothesis. The Airy hypothesis (A) is based on the premise that the crustal blocks are of equal density but the roots are at different levels. The Pratt hypothesis (B) assumes that the blocks are of different densities but the differences are compensated at a certain depth. The Heiskanen hypothesis (C) is a combination of the Airy and Pratt hypotheses.

Spencer 1972, Kingston 1988.

The Pratt hypothesis postulates an equilibrium of crustal blocks of different densities (i.e., topographically higher mountains are less dense than topographically lower units such as the ocean floor, while the depth of crustal material is everywhere the same) (Bates and Jackson 1980). The depth at which the effects of the different densities are balanced (i.e., the level to which all the blocks sink) is the level of compensation.

The lesser-known Heiskanen hypothesis combines assumptions from Airy and Pratt to account for mountain roots, and the variations in crustal densities. It is currently thought that neither the Airy nor the Pratt hypothesis can fully explain the isostatic equilibrium which exists.

The lithosphere is broken into several pieces or "plates" and it is the movement of these plates which is known as plate tectonics.

Mantle Convection

Mantle convection is responsible for the movement of the lithospheric plates. Since the earth's crust is a part of the mantle, it is carried "piggyback", with the mantle's movements. The mantle is warmed at the core-mantle boundary, and then slowly rises to the surface where it cools and sinks again.

Cooling of the mantle takes place through several processes:

· volcanic eruptions at mid-ocean ridges and hot spots;

· seawater circulation through new crust;

· heat conduction through the ocean floor (Gross 1990).

Once cooled, the thickened and denser lithospheric plates are eventually drawn down through subduction trenches, coming to rest on the core-mantle boundary. Over a period of millions of years, the plates will warm and decrease in density until they are sufficiently buoyant to rise through the mantle and begin the cycle again. Images of rising and sinking mantle can be seen using seismic tomography which is similar to computerized axial tomography or the "CAT scans" used in the field of medicine to image human organs (Gross 1990).

Hot Spots. Some spots on the core-mantle boundary are anomalously hot. At these hot spots, plumes of molten rock rise through the mantle and form volcanoes on the surface. Hot spots do not move with the overlying mantle and lithospheric plates. They remain in the same location for tens of millions of years, recording the movement of the overlying plate with a chain of volcanoes (Gross 1990). The Hawaiian islands and the Emperor Seamounts were formed over a hot spot which is currently found to the southeast of the Hawaiian islands. Loihi is still submerged, but will eventually join the chain of volcanic islands. There are no hot spots in the Indonesian Archipelago.

Hot spots are also believed to be responsible for the formation of several aseismic ridges such as the Ninety-east Ridge in the Indian Ocean, and the Walvis and Rio Grande Rises in the South Atlantic (Brown et al. 1989).

Plate Tectonics

The Beginnings of the Theory of Plate Tectonics. The theory of continental drift (or more appropriately, continental displacement) was first proposed in 1915 by Alfred Wegner after he observed the jigsaw puzzle fit of the eastern coast of South America with the west coast of Africa. However, he was not the first to observe this intriguing fact, for Alexander von Humbolt, the naturalist-explorer, had made a similar observation in 1801 (Seibold and Berger 1982). Wegner's book, The Origin of Continents and Oceans, was originally published in 1915 in German, followed by the English translation in 1924. Wegner contended that the similarity of the rocks and fossils of the two coasts supported his theory. He theorized that all the continents were once part of a single continent known as Pangaea. The continents moved apart by the displacement of large plates of continental (sialic) crust, moving freely across a substratum of oceanic (simatic) crust (Bates and Jackson 1980). However his hypothesis was not widely accepted by the scientific community. The major objections were due to his proposed mechanisms. Wegner suggested that the continents were rigid plates moving through the ocean basins. The motions were driven by the variations in the gravitational attraction of the earth's equatorial bulge and the westward drift due to the attractions of the Sun and the Moon (Gross 1990). In addition, Wegner's fossil evidence was not definitive.

Prior to the first mapping of the ocean floor in the 1950s, it was believed that the ocean basins and continents were stable features of the earth. Robert Dietz, in 1961, published a paper entitled "Continent and Ocean Basin Evolution by Spreading of the Sea Floor" introducing the term "sea-floor spreading". Based on the ocean mapping data, Harry H. Hess, in 1962, hypothesized in "History of Ocean Basins," that the earth's outer surface was in motion, causing continents to fragment, move and create new ocean basins. Hess also proposed that new oceanic crust was being formed at mid-ocean ridges by volcanic activity, and destroyed in trenches (Gross 1990). J. Tuzo Wilson in his 1965 paper, "A New Class of Faults and Their Bearing on Continental Drift," elaborated on the ideas, strengthening the concept with the addition of a new type of plate boundary, transform faults.

Figure 2.3. Earth's magnetic poles. Current magnetic orientation is "normal" with the north and south magnetic poles close to the north and south geographic poles respectively. During periods of reversed polarity, the north and south poles are interchanged.

Magnetic Anomalies. Magnetic anomalies on the ocean floor were first noted during submarine detection activities. However, it was not until 1963 that the striped pattern parallel to mid-ocean ridges could be. explained. Drummond Matthews and Fred Vine postulated that the patterns represented reversals in the earth's magnetic field. As new oceanic crust is extruded from mid-ocean ridges, the minerals in the cooling rock are aligned in accordance with the existing magnetic orientation. Each stripe represents a section of the ocean floor formed during a particular magnetic orientation, with the adjacent stripe indicative of a different magnetic orientation or magnetic reversal. At present, the magnetic orientation is "normal" (i.e., the north and south magnetic poles are relatively close to the north and south geographic poles) (fig. 2.3). The.alternating bands of rock with different magnetic orientations create the distinctive "striped" pattern we call magnetic anomalies (fig. 2.4).

Sea floor produced now and during other periods of normal magnetic orientation show strong or positive magnetic values. During periods of reversed magnetic fields, the north and south magnetic poles are reversed, and the rocks record a weak or negative pattern. Reversals of the magnetic field occur approximately every hundred thousand to a few million years. It has been estimated that in the past 76 million years, 171 magnetic field reversals have occurred (Beiser and Krauskoff 1975). Why these reversals take place is not yet known.

Figure 2.4. Magnetic anomalies. A schematic portrayal of how the distinctive striped pattern of ocean floor magnetic anomalies originates. The sea floor presently being formed along the mid-ocean ridge is of "normal" magnetic orientation. Earlier episodes of normal magnetic fields are marked by the other shaded bands. The white bands indicate sea floor formed during periods of reverse polarity. Note the symmetry of the magnetic anomalies on either side of the mid-ocean ridge.

Modified after Ross 1977, and Gross 1990.

Not all oceanic crust displays magnetic anomalies. During a period of the earth's history, 80-120 million years ago, magnetic reversals did not occur. As a result, no magnetic anomalies are present in rocks formed during that interval. Deep burial and intense heat may also erase the pattern of magnetic anomalies (Gross 1990).

Magnetic anomalies are also a means to map the age and rate of spreading of the ocean floor. By correlating the pattern of oceanic magnetic anomalies with the pattern observed in rocks of known ages on land, we can determine the ages of the various sections of ocean crust, and estimate the rate of sea-floor spreading.

Spreading rates range from less than one centimetre per year on the Mid-Atlantic Ridge, to 16 centimetres per year on the East Pacific Rise. In all cases, the youngest crust is found in a band straddling the spreading ridge, with increasingly older crust on each side of the ridge as the distance from the ridge crest increases. The faster the spreading rate, the thicker the central band of young crust. The oldest oceanic crust, estimated to be 190 million years old, is found in the North Pacific near Asia, and along the margins of the North and South Atlantic. It is believed that approximately half the ocean floor is less than 80 million years old (Gross 1990).

Further support for Wegner's Pangaea theory was provided by Sir Edward Bullard in 1965. Using a common depth contour as the edges of all the continents, Bullard was able to piece together the continents to form a supercontinent resembling Pangaea. There were some areas of overlap where relatively new features such as coral reefs and river deltas had developed, but overall the fit was good enough to support Wegner and the Pangaea concept (Gross 1990).

Plate Tectonics -The Parts and the Processes. Plate tectonics is the movement of the earth's lithospheric plates (composed of the upper mantle and crust) on the asthenosphere. There are seven major plates and many small plates. The major plates are: Pacific, Indo-Australian, North American, South American, Eurasian, African, and Antarctic. Smaller plates include the Philippine Plate, China Plate, Gorda Plate, Cocos Plate, Nazca Plate, Caribbean Plate, Scotia Plate, Arabian Plate, Iranian Plate, Hellenic Plate, and Juan de Fuca Plate (fig. 2.5).

Plate Boundaries. All plates are in contact with several other plates. There are three types of plate boundaries:

a) divergent or constructive margins

b) convergent or destructive margins

c) conservative boundaries or transform faults.

Divergent or Constructive Margins. Divergent margins are also known as Atlantic, passive, aseismic or constructive margins. Divergent margins develop when continents rift apart and form new ocean basins. As a result, continental crust and the adjacent oceanic crust are part of the same plate. As the rift widens, the continental margin grows further from the spreading centre and closer to the stable interior. Micro-continents may form if these pieces of continental crust are isolated as a result of rifting or other plate movements (Brown et al. 1989). The actual cause of a divergent margin is not fully understood, but it is thought to begin with the development of crustal stretching, extensional faults, rift basins, and possibly regional uplift and volcanism initiating sea-floor spreading (Hutchinson 1992).

A divergent margin, as shown in figure 2.6, generally features the following physical characteristics:

a) a continental shelf, gently sloping (average gradient of 0.1°) from the shore and extending to a depth not exceeding 130 m; may be 1500 m wide

b) a steep (average gradient of 4°) continental slope, 20 to 100 km wide, continues from the continental shelf to a depth of 4000 to 5000 m, and may feature submarine canyons

c) the continental rise, (average gradient of 1 °), may be up to 600 km wide and leads into the abyssal plains (Hutchinson 1992; Brown et al. 1989).

Spreading ridges or spreading centres are the site of new crust formation and occur at divergent or constructive margins. As new crust cools and moves away from the spreading ridge, it increases in density and thickness. New crust is only a few kilometres thick, while old crust reaches thicknesses of 150 km. With increasing density, the crust sinks deeper into the asthenosphere from a depth of 2500 m to 6000 m as it ages (Gross 1990). Two extensively studied examples of early stage continental rifting are the northern Red Sea and the Salton Trough in Baja, California (Hutchinson 1992). Active spreading centres in this region are found in the Andaman Sea, Aru Trough, Bismarck Sea, and the Woodlark Basin (Petroconsultants Australasia 1991).

Figure 2.5. Tectonic plates of the world. Simplified world map showing the major plates and some minor plates.

After Brown et al. 1989, Gross 1990, and Ganeri 1994.

Figure 2.6. Divergent margins. A generalized cross-section showing the main features of divergent margins: (1) continental shelf, (2) continental slope, (3) continental rise, (4) continental margin, (5) abyssal plain, (6) oceanic ridge. Not to scale.

Modified after Brown et al. 1989, and Hutchinson 1992.

Convergent or Destructive Margins. Convergent margins are also known as active, seismic, Pacific-type, or destructive margins. On land, visible signs of convergent margins are active volcanoes, frequent earthquakes, island arcs, and young mountains (Gross 1990). Below the sea surface, subduction trenches mark locations of convergent or destructive margins.

Subduction Zones. Most subduction zones involve the subduction of oceanic crust beneath lower density continental crust. Oceanic crust is destroyed as it is drawn into the mantle. Earthquakes occur as the plates drag past each other. Deep- and intermediate-focus earthquakes (100-700 km below the surface) are indicative of subduction zones (Gross 1990). General features of a subduction zone are: trench, accretionary wedge and fore-arc ridge, fore-arc basin, volcanic arc (island arc), and back-arc basin (fig. 2.7). The trench is where the descending plate first contacts the material from the overriding plate as it heads towards the mantle. The accretionary wedge forms when sediments and other materials are scraped from the subducting plate and accumulate (like the action of a snowplow pushing the snow ahead to clear a path), at the front of the overriding plate (Hamilton 1989). It is composed of deformed rocks. The fore-arc ridge is the summit of the accretionary wedge. The fore-arc basin is found landward of the accretionary wedge and contains relatively less-deformed sediments (Hutchinson 1992). The frontal arc located between the accretionary wedge and the volcanic arc is a zone of uplift and deformation. The volcanic arc is a zone of active igneous activity (Hutchinson 1992). The back-arc region is found behind the volcanic arc and may contain marginal (back-arc) basins or ancient, inactive arcs.

Figure 2.7. General features of a convergent margin are: (1) active trench, (2) accretionary wedge, (3) fore-arc ridge, (4) fore-arc basin, (5) frontal arc, (6) volcanic arc, (7) back-arc region.

Modified after Curray 1989, and Hutchinson 1992.

The Benioff zone, also known as the Benioff seismic zone, is a seismic zone where earthquake foci cluster. An earthquake focus is the point within the earth which is the centre of an earthquake (Bates and Jackson 1980). The Benioff zone stretches downward from the oceanic trench, dipping toward the continents usually at an angle of 45°.

The Scenario Involving Old Subducting Crust. When the subducting or descending oceanic plate is old and therefore dense, it descends as a steeply dipping slab. This is a common scenario in the western Pacific basin where the oldest oceanic crust occurs (fig. 2.8).

Subduction of the oceanic crust may also be accompanied by back-arc spreading and basin formation as is the case in the area of the Mariana Trench (Gross 1990). According to Hutchinson (1992), subduction involving old oceanic crust is "low-stress", and commonly features a small accretionary wedge, few large earthquakes, igneous rocks with a narrow basaltic compositional range, a steeply dipping descending slab, and a well-developed back-arc basin.

The Scenario Involving Young Subducting Crust. When the subducting plate is composed of relatively young, still-buoyant crust, it descends as a shallow-dipping slab (fig. 2.9). The east coast of the Pacific (i.e., the west coast of North and South America) is an example of where this type of shallow subduction is taking place. Active volcanoes and young mountains line the coast. The mountains are formed from material scraped off the subducting plate (Gross 1990). Subduction involving young oceanic crust is classified by Hutchinson (1992) as "high-stress", and features a large accretionary prism, large shallow earthquakes, igneous rocks of varied composition, and a shallow-dipping plate.

Figure 2.8. Subduction involving old oceanic crust. The main features to note are: (1) steeply dipping oceanic plate, (2) deep trough with little accumulation of sediments or a small accretionary wedge on the upper plate, (3) active island arc, (4) well-developed back-arc basin and, (5) extinct arc. (Few large earthquakes.)

Modified after Gross 1990, Curray 1989, and Hutchinson 1992.

When material is scraped off an oceanic plate and accretes to a continental plate, it is known as an ophiolite. Ophiolites are a specific type of exotic terrane which is any fragment of continental or oceanic plate welded onto a continent. Ophiolites are often rich sources of sulphide minerals formed at spreading centres, and are mined for copper, lead, zinc, and silver (Gross 1990).

Subduction is presently occurring off the coasts of Sumatra, Java, along the East Sunda Arc, Banda, Seram, Sangihe, North Sulawesi, Cotabato/West Sangihe, and Halmahera (Petroconsultants Australasia 1991). The trenches where subduction is taking place are known as active trenches, as opposed to inactive trenches where subduction has ceased.

Collision of Two Continental Plates. Collision of two continental plates, does not produce a subduction scenario because both plates are composed of fairly buoyant crust. Instead, overriding and uplift can occur with one plate being folded and thrust upon the other (Gross 1990). This is what happened when the Indian Plate collided with the Eurasian Plate resulting in the formation of the Himalaya Mountains. Collisions are also taking place in Sulawesi, and Timor-Tanimbar (Petroconsultants Australasia 1991).

Subduction Involving Two Oceanic Plates. When one oceanic plate is subducted beneath another oceanic plate, a trench is present, and a volcanic island arc forms on the overriding plate. On the opposite side of the island arc, away from the trench, commonly a marginal or back-arc basin forms as a result of sea-floor spreading.

Figure 2.9. Subduction involving young oceanic crust. The main features to note are: (1) shallowly-dipping oceanic plate, (2) relatively shallow trench, (3) large accretionary wedge, (4) fore-arc basin, (5) active volcanoes. Earthquakes are generally large and shallow.

Modified after Gross 1990, and Hutchinson 1992.

Conservative Boundaries or Transform Faults. The third type of plate boundary involves lateral movement with transform faults (fig. 2.10). Transform faults are a type of strike-slip fault in which the plates slide horizontally past each other with oceanic crust neither created nor destroyed (Roberts 1989). Many transform faults are associated with mid-ocean ridges and are seismically active. They run perpendicular to the line of the ridge, offsetting the ridge, and terminate where they meet the ends of the two offset ridge segments which they connect (Brown et al. 1989). The extension of a transform fault beyond those points is known as a. fracture zone, and is characterized by no relative sideways motion, and low seismicity (i.e., small earthquakes) as it is located within a single plate. Examples of transform faults in Indonesia are in Sumatra and Sorong (Petroconsultants Australasia 1991). One of the world's best-known examples of a transform fault is the San Andreas Fault in California, U.S A.

Figure 2.10. Transform faults. The ridges of the earth are marked by a series of transform faults which offset the ridges resulting in the irregular outlines. Movement along the faults during seafloor spreading produces shallow earthquakes marking the active section of the fracture zone. Beyond the active area the fracture zone features steep ridges and valleys. (1) Mid-Atlantic Ridge, (2) Southwest Indian Ridge, (3) Indian Ridge, (4) Southeast Indian Ridge, (5) Ninety-east Ridge, (6) Java Trench.

Modified after Ganeri 1994, Earth 1992, and Gross 1990.

Earthquakes

Rocks, like people, can be affected by stress. Stress builds up in rocks as lithospheric plates move past each other. When the stress is greater than the strength of the rocks, the rocks are strained, then fail (i.e., fracture along zones of weakness), faults are formed, and the energy is released in the form of an earthquake (Keller 1979). Although plates are often described as "sliding" past another plate, the movement is not smooth and there is much friction where the plates meet. Movement along faults produces seismic waves to depths of 700 km (Ross 1977), and in the near-surface of the earth.

An earthquake may be described as a sudden motion of the earth caused by faulting or volcanic activity. As many as a million earthquakes each year are recorded on seismographs, with most escaping the notice of human populations.

The centre of an earthquake is called the focus or hypocentre. The focus is the point deep within the earth of the initial rupture and where the strain energy is first converted to elastic wave energy (Bates and Jackson 1980). Shallow-focus earthquakes are those with a focus at 70 km or shallower. Deep-focus earthquakes begin below 300 km (Kingston 1988). The point on the earth's surface above the focus is called the epicentre. The epicentre is generally located directly over the focus, except when the earthquake originates deep within a subduction zone.

The Mercalli Intensity Scale. The Mercalli intensity scale, devised in 1902, is an arbitrary scale of earthquake intensity, ranging from I (detected only by instruments) to XII (almost total destruction). Its adaptation to urban conditions is known as the modified Mercalli scale (Bates and Jackson 1980).

Modified Mercalli scale (from Whitten 1972, and Spencer 1972):

I. Instrumental. Detected only by seismographs.

II. Feeble. Noticed by only a few persons at rest. Delicately suspended objects may swing.

III. Slight. Resembles vibrations caused by heavy vehicle traffic.

IV. Moderate. Felt by most people. May waken some. Rocking of free-standing objects. Walls crack. Sensation like heavy truck striking building.

V. Rather strong. Sleepers awakened. Bells ring. Widely felt.

VI. Strong. Felt by all. Trees sway. Some damage from overturning and falling of objects.

VII. Very strong. General alarm. Everyone runs outdoors. Damage is negligible in buildings of good design and construction. Slight to moderate damage in well-built ordinary structures.

VIII. Destructive. Fall of factory smoke stacks, columns, monuments, and walls. Heavy furniture overturned.

IX. Ruinous. Ground begins to crack. Houses collapse. Underground pipes break.

X. Disastrous. Ground badly cracked. Many buildings destroyed. Some landslides. Rails bent. Water splashed over banks.

XI. Very disastrous. Few buildings remain standing. Bridges and railways destroyed. Broad fissures in the ground.

XII. Catastrophic. Total destruction. Waves seen on ground surface. Objects thrown into the air.

The modified Mercalli intensity scale is based on effects which can be observed, and refers to the "violence of the earthquake motion" (Spencer 1972). However, this scale is difficult to apply for accurate, quantitative, worldwide comparisons of earthquakes. In 1935, the Richter magnitude scale was devised.

The Richter Magnitude Scale. The Richter magnitude scale, more commonly used today, and generally referred to as the Richter scale, is a logarithmic scale, based on the amount of energy released at the focus, as recorded by a seismograph. A seismograph records earthquake waves. The amplitude of the largest wave determines the magnitude (Keller 1979). An earthquake of magnitude 6 produces a displacement on a seismograph 10 times larger than does a magnitude of 5. A magnitude 5 earthquake releases approximately 10²¹ ergs of energy which is equivalent to the first atomic bomb detonated in 1945, or 20,000 tons of TNT (Spencer 1972). However, although the total energy released may be comparable, its mode of dispersal produces very different effects depending on whether it was released in a highly concentrated form as in the atomic bomb, or widely dispersed as in an earthquake. Table 2.1 provides a qualitative description of the Richter scale.

Effects of Earthquakes. In addition to the well-known disastrous effects of earthquakes on population centres, earthquakes affect marine environments (see section on coral reefs and natural disturbances). The primary effects of earthquakes are violent ground/substrate motion, surface rupture, and permanent substrate displacement of a metre or more (Keller 1979). Secondary effects maybe divided into short-term events such as landslides, tsunamis and floods. Long-range effects are regional subsidence or emergence of landmasses.

Distribution of Earthquakes. Active seismicity marks plate boundaries. The zones of seismicity tend to be narrow when associated with mid-oceanic spreading centres, and strike-slip faults. Wider zones occur above subducting plates, and in those parts of continents undergoing distributed extensional, strike-slip, and compressional deformation (Hamilton 1979). Earthquakes of a magnitude greater than 8.0 occur primarily along subducting plate boundaries, and less frequently in continental strike-slip and compressional deformation situations.

Distribution of Earthquakes Worldwide. The earth can be divided into 10 regions based on seismic activity:

1. The circum-Pacific belt contains most of the shallow- and intermediate-depth earthquakes, and almost all deep-focus earthquakes;

2. The Alpine belt of Europe and Asia, containing the Alps and Himalayas, is the other significant zone of shallow and intermediate-depth earthquakes;

3. The Pamir-Baikal zone of central Asia;

4. The Atlantic-Arctic belt;

5. The belt of the central Indian Ocean;

6. Rift zones, notably those of east Africa;

7. A wide triangular active area in eastern Asia, between the Alpine belt and the Pamir-Baikal zone;

8. Minor seismic areas, usually in regions of older mountain building;

9. The central basin of the northern Pacific Ocean; almost nonseismic except for the Hawaiian islands;

10.The stable central shields of the continents, also nearly nonseismic.

Distribution of Earthquakes in Indonesia. Indonesia, located within the famed "Ring of Fire," is frequently hit by earthquakes covering a range of magnitudes. Approximately 10% of the world's seismicity occurs in the Indonesian Archipelago. Katili (1985) gives a brief summary of the distribution of Indonesia's shallow, intermediate, and deep earthquakes (fig. 2.11).

Table 2.1. Scales of magnitude on the Richter scale.

Figure 2.11. Distribution of shallow and deep earthquake epicentres in Indonesia.

After Katili 1985, and Hamilton 1974.

Indonesia's shallow-focus earthquakes most commonly occur above the subducting plate boundaries of the Java-Sumatra Trench and the Banda Trench, within the active collision zone between Sulawesi and Halmahera, and associated with transcurrent faults (i.e., the Great Sumatran fault zone, the Palu-Koro-Matano fault zone, the Gorontalo fault zone, and the Sorong fault zone).

The intermediate-depth (focus depth of 100-300 km) earthquakes are generally spread along "the whole volcanic arc divided in the middle by the axis of active volcanoes in Sumatra, Java, the Lesser Sunda Islands, Sulawesi and Halmahera" (Katili 1985).

The deep earthquakes (focus depths of 500-800 km) are clustered along an east-west trending belt from the Java Sea to the Banda Sea, and a north-south trending belt from Sulawesi to Mindanao.

VOLCANOES

To many, the mere mention of "Indonesia" brings to mind the archipelagic nation's most famous volcanoes, Krakatau and Tambora. Krakatau, immortalized by Hollywood's movie industry, although geographically misplaced in the film Krakatau, East of Java, ranked fourth in the world's greatest historic eruptions (Hutchison 1982). The 1815 eruption of Tambora on the island of Sumbawa was the most violent explosion of recorded history, yet is relatively unknown outside geological circles. Indonesia's two most famous eruptions will be discussed in greater detail after an introduction to volcanoes and volcanic activity.

The Shapes of Volcanoes

Volcanoes come in several shapes and sizes. "Volcanoes show a wide variety of forms, depending largely upon the composition of the erupted material and hence the style of eruption" (Thorpe and Brown 1985). The three main shapes are cinder cones, shield volcanoes and composite or stratovolcanoes.

Several factors determine the shape of a volcano:

· Land surface;

· Type and nature of material ejected (i.e., viscous lava and cinder build up steep cones, while more fluid lava flows further away from the vent and results in wide-based mountains);

· Forcefulness of the eruption;

· Duration of the activity;

· Eruptive history of the volcano.

Cinder Cones. Cinder cones are built when pyroclastics, such as cinders, and ash, pile up around the vent. A crater normally surrounds the vent. The slopes are generally steep (greater than 10°) and symmetrical. These volcanoes are usually basaltic or andesitic. The magma contains a relatively high gas content, resulting in higher explosivity (van Bemmelen 1949) than cumulo-volcanoes (volcanic domes) or lava shields (shield volcanoes of the basaltic type). Examples: Lamongan, East Java and Vesuvius, Italy.

Shield Volcanoes. Shield volcanoes feature the largest cones. This type of volcano has a broad, shield, or low-slope profile. Its diameter may be between 100 and 200 km (Thorpe and Brown 1985). Shield volcanoes represent the largest discrete volcanic form, but they are rarely preserved in the geological record, since they occur mainly as oceanic islands (Thorpe and Brown 1985). They are produced by eruption of low-viscosity lava from either a central vent, fissures or parasitic vents. The flow generally consists of very fluid basaltic lava or rhyolitic ash flows (Bates and Jackson 1980). The gas content of the magma is low. They are also sometimes referred to as lava domes. Examples: Sukadana, South Sumatra and Mauna Loa, Hawaii.

Composite Cones (Stratovolcanoes). Stratovolcanoes, or composite cones, have a central vent from which lavas as well as pyroclastics are expelled. This type of volcano is composed of alternating layers of lava and pyroclastics, resulting from the prolonged activity of a central vent causing the formation of bedded-volcanoes (van Bemmelen 1949).

Stratovolcanoes are the most common shape for andesitic volcanoes. They have steep, irregular conical forms with diameters of 10-40 km. The volcano often has a shape combining cinder and shield volcanoes. There are often many dikes and sills. The lava, viscous and acidic, is generally restricted to the area of the volcano, but the pyroclastic material may be transported by wind more than 1000 km (Thorpe and Brown 1985). These deposits are often used by stratigraphers as markers or reference points. Examples: Merapi, Central Java and Fujiyama, Japan.

Composite volcanoes clearly show a zonation of volcanic products which may be divided into the central, proximal arid distal zones with increasing distance from the central vent (Thorpe and Brown 1985).

The central zone is located within 2 km of the central vent. It is characterized by lava conduits. The volcanic products associated with this zone are coarse, poorly-sorted pyroclastic materials which have been deposited close to the vent.

The proximal zone, situated 5-15 km from the central vent, contains a higher proportion of lava flows and a variety of pyroclastic flow deposits.

The distal zone is found beyond the proximal zone, and consists of pyroclastic flow deposits associated with fine air-fall deposits dispersed by the wind away from the volcano.

Other Volcanic Landforms

Caldera. A caldera is a large, basin-shaped volcanic depression, more or less circular, the diameter of which is many times greater than that of the included vent or vents (Bates and Jackson 1980). There are two main types of volcanic calderas depending on the mode of formation (i.e., collapse caldera, and explosion caldera). A collapse caldera is formed when the top of the magma chamber collapses into a void or cavity created by the removal of magma either by large volume eruptions of lava or pyroclastics, by subterranean withdrawal of magma, or contraction of magma as it cools and crystallizes. The top of the volcano collapses into the void (subsidence phenomena) forming an enclosed or partially enclosed depression called a caldera which maybe tens of kilometres across. Most calderas are of this type (Bates and Jackson 1980). An explosion caldera is formed by the explosive removal of the upper part of a volcanic cone. This type of caldera is extremely rare, and is small in size according to Bates and Jackson (1980).

Plateau Basalts. Plateau basalts are the most extensive volcanic landform, with the basaltic lavas covering areas up to 10⁵ km² (Thorpe and Brown 1985). The lava originates from cracks or fissures (i.e., fissure eruptions) rather than from a central vent (Rhodes 1991). Plateau basalts erupted in rapid succession over vast areas and have, at times, flooded sectors of the earth's surface on a regional scale (Bates and Jackson 1980). Examples are known from India, Iceland and the United States of America where the Columbia River Plateau has an estimated volume of 417,000 km covering an area greater than 250,000 km (Spencer 1972). These lavas range in age from Eocene (40-60 million years old) to Pleistocene and Recent (i.e., less than 2 million years old).

Volcanic Domes. Volcanic domes are dome-shaped, bulbous masses of hardened lava. They form above and around volcanic vents. When a volcanic dome forms on the side of, or close to a larger volcanic cone, it is known as a parasitic volcano. Parasitic volcanoes are generally 1-2 km diameter, created by a single, short-lived eruption (Thorpe and Brown 1985).

Lahar. Lahar is an Indonesian word which has been adopted by geologists worldwide to describe mudflows which contain debris and angular blocks mostly of volcanic origin (van Bemmelen 1949). The transporting medium is a mixture of cool (<100°C) liquid water and gas. The transported materials include poorly-sorted large volcanic fragments in a finer ash-size matrix (Thorpe and Brown 1985). Some of the materials may be derived directly from magma reaching the surface or juvenile materials (Bates and Jackson 1980). According to Thorpe and Brown (1985), eruptions of volcanic material through lakes, below ice or during heavy tropical rain may generate lahars composed largely or entirely of juvenile volcanic material.

Van Bemmelen (1949) described normal lahars or cold mudflows as not exclusively volcanic. They originate by heavy rainfall on slopes covered with loose material or by earthquakes (e.g., Benkulen 1933) and are typical for tropical volcanoes (e.g., Merapi). They were also known as rain lahars but may more correctly be termed debris flows or slumps. These debris flows are especially common where deposits form in shallow water conditions subject to volcanic instability and seismic activity (Thorpe and Brown 1985).

Hot mudflows are caused by the emptying of a crater lake through crater wall collapse or explosion. Under these circumstances, the lake water is mixed with hot volcanic material (e.g., Kelud, East Java 1811, 1826, 1835, 1848, 1864, 1901, and 1919) (van Bemmelen 1949).

Lahars tend to be deposited in low-lying land close to a volcano, although lahars have been found up to 300 km from their source (Thorpe and Brown 1985). Rain lahars are typical of tropical volcanoes and influence the outline of the cones (van Bemmelen 1949).

The shape of a volcano is obviously a reflection of the volcanic activity which is in turn determined by several physical factors.

Factors Determining the Types

of Volcanic Activity

Several factors determine the type of volcanic activity:

· temperature, composition, fluidity/viscosity, and pressure build-up of the magma

· types of release conduits or openings through which the magma may be released (i.e., summit craters, parasitic craters, radial fissures, regional fissures)

· where the eruption occurs (i.e., submarine, sublacustrine or under a lake, subglacial, or subaerial).

Eruptions are generally classified as either explosive, lava or mixed eruptions.

An explosive eruption is an eruption that is characterized by the energetic ejection of pyroclastic material (Bates and Jackson 1980). There are two main types of explosive eruptions. Incandescent eruptions describe episodes involving glowing pyroclastic material such as ash flows, and Nuée ardentes. Phreatic explosions involve the explosion of steam, mud and other incandescent (non-glowing) materials.

A lava eruption is characterized by the emission of lava, generally free of explosions.

A mixed eruption includes both the emission of lava and the explosive ejection of pyroclasts.

Specific types of volcanic eruptions have been named after type localities (e.g., Hawaiian, Strombolian, Plinian, Pelean, Vulcanian, or Katmaian type eruptions).

Types of Volcanic Eruptions.

HAWAIIAN TYPE. The Hawaiian type eruption is characteristic of shield volcanoes. The height of the eruptive column is very low (i.e., materials are not sent high into the atmosphere). The area affected by ash fall is less than 0.1 km² and explosive episodes are rare. Fissures on the sides of the volcano, are the conduits from which the highly fluid basaltic lava is extruded (Bates and Jackson 1980). Example: Mauna Loa, Hawaii.

STROMBOLIAN TYPE. The Strombolian type eruption has a moderate eruptive column height of up to 1 km. Ash falls within a 5 km² radius. The fluid, basaltic lava spurts, fountain-like, from a central crater. The Stromboli Volcano on the Lipari Islands, Italy, typifies this type of eruption pattern.

PLINIAN TYPE. The Plinian type of eruption features a towering eruptive column reaching up to 20 km. Nuées ardentes are common. The area affected by ash fall may be as much as 1000 km². Unlike the previous two eruptive types which were cone-building, Plinian eruptions are sheet-building. Plinian eruptions are characterized by "explosive eruptions in which a steady, turbulent stream of fragmented magma and magmatic gas is released at a high velocity from a vent" (Bates and Jackson 1980). Examples: Tambora, and Krakatau.

PELEAN TYPE. The Pelean type eruption is characterized by gaseous clouds (nuées ardentes) and/or the development of volcanic domes. Example: Mount Pelee, Martinique.

VULCANIAN/VESUVIAN/VULCANO TYPE. The Vulcanian type eruption, also known as the Vesuvian or Vulcano type, is characterized by the explosive expulsion of new lava fragments. Although the fragments are glowing when they leave the vent, they are too solid or viscous to be altered in shape while airborne. Material expelled includes blocks and ash. Example: Mount Vesuvius, Italy.

KATMAIAN TYPE. The Katmaian type eruption describes the violent, explosive release of vast amounts of pumice and ash, followed by an ash flow and extensive fumarole activity. Fumaroles are holes or vents through which gases and vapors escape. They are characterized by quiet emission and sulphur deposition. Fumaroles are much more common than geysers, but less of a tourist attraction. They may be situated along fissures or in clusters known as a fumarole field. An Indonesian example is Wurlali Volcano crater, Damar Island, Banda Sea. Another example is Mount Katmai, Alaska, which includes the Valley of Ten Thousand Smokes, an area which, as its name suggests, is rich in fumaroles.

PAROXYSMAL ERUPTION. A paroxysmal eruption is an eruption of the Katmaian, Pelean, Plinian or Vulcanian type.

PHREATIC ERUPTION. A phreatic eruption involves the explosion of steam, mud, or other material that is not incandescent. It is caused by the heating and consequent expansion of groundwater due to an underlying igneous heat source (Bates and Jackson 1980). The greatest phreatic eruption in recorded history was in 1933 at Suoh, South Sumatra (Hutchison 1989b).

Products of Volcanic Activity

The materials which emanate during volcanic activity fall into one of three groups; lava, gases, and pyroclastics (solid fragmental materials).

Pyroclastics. "Pyroclastic" describes clastic rock material which has formed by volcanic explosion or aerial expulsion from a volcanic vent (Bates and Jackson 1980). Pyroclastic rocks are composed of materials fragmented by explosive activity (Thorpe and Brown 1985). Pyroclastics are generally classified according to size and shape. Pyroclastics also include partially fluidized mixtures of particles and gases that travel up to 150 m/s and have internal temperatures of up to 600°C. Pyroclastics are denser than the atmosphere but may be of comparable density to seawater (Sigurdsson et al. 1991).

Pyroclastic Classification,

GENETIC CLASSIFICATION. Pyroclastics may be separated into two categories on the basis of how they are deposited, or their genetic classification.

Air-fall deposits result when pyroclastic material is erupted into the atmosphere and settles out downwind from the vent. Fine-grained pyroclastic material may travel more than 1000 km (Thorpe and Brown 1985).

Pyroclastic flows result from transport of solid fragments of volcanic rock in a fluid (gas or liquid) matrix away from the volcano. Unlike the air-fall deposits which move up and away, pyroclastic flows travel close to the ground, moving away from the vent laterally. Pyroclastic flows may travel up to 100 km. Since the flows may have volumes of up to 3000 km³, the shields or plateaux created by such flow deposits may cover areas of 10⁵ - 10⁶ km² (Thorpe and Brown 1985). These plateaux are more likely to be preserved in the geologic record than basaltic plateaux, since the pyroclastic plateaux tend to be found in inland continental areas distant from destructive margins.

Nuée ardente is a French term meaning "glowing cloud", which aptly describes the swiftly flowing, turbulent, often incandescent, gaseous cloud which contains ash and other pyroclastics in its lower section (Bates and Jackson 1980). Up to several hundred thousand cubic metres of red-hot blocks may be associated with the cloud of hot air and ashes. This glowing cloud is extremely destructive, as was the case in 1902 when the eruption of Mount Pelee released a nuée ardente that wiped out the town of St. Pierre, Martinique.

LITHOLOGICAL CLASSIFICATION. Pyroclastics may be classified according to type of material and grain size, or a lithological classification.

Blocks are the largest in size, with diameters greater than 64 mm, and may weigh several tons. Blocks are ejected in a solid state.

Bombs are also 64 mm in diameter or larger, but unlike blocks, bombs are ejected while still viscous, so they are shaped as they travel through the air. Bombs, therefore, tend to be somewhat rounded, although they exist in a variety of shapes and have similarly diverse internal structures ranging from vesicular to hollow.

Lapilli fragments range in size from 2 to 64 mm. A lapillus does not have a characteristic shape, and may be viscous or solid when it hits the ground.

Ash consists of fine, unconsolidated pyroclastic material, with a diameter size range of 1/16- 2 mm.

Dust is fine volcanic ash, less than 1/16 mm in size.

Pumice is a light-coloured, porous, glassy, highly vesicular rock which is formed when gases bubble through a highly viscous rhyolitic lava. This rock is of such low density that it floats on water.

Scoria is similar to pumice (i.e., a vesicular volcanic rock), but contains magnesium and iron and is heavier and darker. Scoria is generally formed from andesitic or basaltic lava.

Tuffs are rocks composed of compacted, consolidated volcanic ash or dust.

Agglomerate is a pyroclastic rock consisting of medium- to large-sized pyroclastic fragments embedded in a finer-grained matrix of volcanic ash.

Ignimbrite is a pyroclastic rock formed by the eruption of an ash flow. It is composed of crystals and rock fragments lying in a matrix of glass shards (Roberts 1989). Ignimbrites are essentially consolidated ash flows and nuées ardentes. They may be massive and continuous over long distances resembling lava flows.

From the discussion above, it is apparent that volcanoes produce more than lava. Thorpe and Brown (1985) point out that all volcanoes erupt both lava and solid pyroclastics, although the proportions vary depending on the type of volcano. For example, basaltic volcanoes such as those in Hawaii generally expel 80% lava and only 20% pyroclastics. The volcanoes of island arcs and active continental margins tend to erupt less than 10% andesitic lava, and more than 90% pyroclastics. The quantities of pyroclastic rocks, especially on island arcs and continental margins, may be underestimated because they are more easily eroded and wind-dispersed than solid lava (Thorpe and Brown 1985).

Lava. Lava includes both molten extrusive rock and its solidifed product (Thorpe and Brown 1985). It is erupted at temperatures ranging between 900° and 1200°C (Rhodes 1991). Three main forms have been described; aa or blocky lava, pahoehoe or ropy lava, and pillow lava.

Aa, or blocky lava, is produced by sudden gas escape. It has a rough, jagged surface. Aa is a Hawaiian word pronounced "ah-ah".

Pahoehoe, or ropy lava, is formed at higher temperatures than blocky lava. This basaltic lava is typified by a smooth, billowy, or ropy surface. Pahoéhoé is pronounced "pa-ho-e-ho-e," and is also a Hawaiian term.

Pillow lava is erupted subaqueously, or below the water surface. Pillow-shaped masses are formed. This type of lava formation is typical of deep-ocean eruptions, and submarine volcanoes. The lava may be either basaltic or andesitic (Bates and Jackson 1980). Pillow lavas have glassy, chilled margins which resulted when the lava encountered the relatively cold water. They may also contain marine sediments and possibly chert (quartz) from radiolarian tests or skeletons (Spencer 1972).

Gases. The most abundant gas released during volcanic activity is H₂O or water vapour. Other commonly released gases are: CO₂, N₂, SO₂, SO₃, S₂, H₂, CO, Cl₂, and A (argon).

Distribution of Volcanoes

Of 800 active volcanoes, 75% are in the "Ring of Fire", the belt of volcanoes and frequent seismic activity nearly encircling the Pacific.

Volcanoes and volcanic islands are common in the ocean. It is estimated that there are many tens of thousands of extinct volcanoes with an elevation of at least 1 km above the sea floor.

Most volcanic eruptions occur at mid-ocean ridges. Recent work by scientists aboard the JOIDES (Joint Oceanographic Institutions for Deep Earth Sampling) Resolution, drilling off the east coast of Greenland in October 1995, revealed 60 million-year-old lava several hundred metres beneath the sea floor. Cores were obtained for depths as great as 8000 metres.

We have never seen the scale or the kind of eruptions like the ones in the rock we were studying. It's like taking the volcanoes in Hawaii and multiplying [them] a hundredfold. The eruption-rate was huge.—

JAMES ALLAN, STAFF SCIENTIST, JOIDES RESOLUTION

THE GLOBE AND MAIL, NOVEMBER 11,1995

The discovery of the massive amounts of 60 million-year-old lava, suggests that Greenland and Europe did not drift apart, but "were forced apart by colossal eruptions over the course of several million years. The dark volcanic rock found under the ocean floor matches the age and characteristics of rocks found in northern Europe and Greenland" (The Globe and Mail, November 11, 1995 p.D8).

Seamounts. Some volcanoes are situated in the middle of ocean basins. These shield volcanoes may be several kilometres high breaking the sea surface to form islands. Examples of active shield volcanoes are the island of Hawaii, and Reunion Island (west Indian Ocean).

When the volcano becomes dormant, it will slowly be eroded until it disappears beneath the ocean surface, and becomes a seamount. A guyot is a type of seamount with a flat top. The Emperor Seamounts are a subsurface chain of extinct volcanoes estimated to be 40-60 million years old.

Distribution of Active Volcanism in Indonesia

Indonesia has more than 500 young volcanoes. During recorded history, 80 of these volcanoes have erupted, and 50 are in the solfataric and fumarole stage (Neumann van Padang 1951). The types of eruptions exhibited by the Indonesian volcanoes are:

· Quiet emission and sulphur deposition at fumaroles;

· Phreatic explosions;

· "Normal volcanic activity" characterized by paroxysmal gas outbursts, followed by viscous lava seeping from the vent;

· Plinian eruptions;

· Nuées ardentes.

Van Bemmelen (1949) described the following rows and groups of active volcanism in Indonesia and adjacent regions:

a. The volcanic inner arc of the Andaman-Nicobar group (with Narcondam and Barren-Island);

b. The volcanic inner arc of the Sunda Mountain system (Sumatra, Java, Lesser- Sunda and Banda inner arc);

c. Batu-Tara – Yersey "Reef – Niewerkerk – Emperor of China – Api (north of Wetar);

d. Una-Una in the Gulf of Tomini;

e. Minahasa – Sangihe – Ragay zone (Philippines);

f. Manila zone and its SW branches (Philippines);

g. The Halmahera-Ternate row, which can be traced via Batjan and Kofiau to the volcanic northern range of the "Vogelkop" Peninsula of New Guinea, with the Umsini;

h. The Ruk arc along the northeast coast of New Guinea;

i. Volcanoes in the eastern part of the central Mountains of New Guinea.

a-f: Belong to the young mountain systems around the more stable Asiatic Sunda Area;

g-i: Related to the young orogenesis in the geosynclinal belt between Melanesia and Australia.

Volcanic Inner Arc. Van Bemmelen (1949) also describes the volcanic inner arc of Indonesia consisting of the islands:

· Narcondam-Barren;

· Sumatra – Java – Lesser Sunda Islands – Banda inner arc;

· Minahasa – Sangihe;

· North Halmahera – Ternate – Batjan;

· Northern range of the Vogelkop of New Guinea;

· Ruk (or Rook) row of volcanic islets with extinct volcanism along the Alor-Romang section.

Non-Volcanic Outer Arc. Van Bemmelen's non-volcanic outer arc included these islands (see fig. 2.12):

· Andamans and Nicobars;

· Islands west of Sumatra-submarine ridge south of Java – Savu – Roti – Timor – Banda outer arc (Tanimbar – Kai – Seram);

· Maju Ridge;

· Southern range of the Vogelkop of New Guinea.

Indonesia's Most Famous Eruptions

Tambora. The eruption of Tambora in 1815 was the greatest eruption in recorded history. It is estimated that the energy released was 8.4 x 10²⁶ ergs. The volume of material erupted was approximated at 100-318 km³. Tambora's cone with an estimated volume of 30 km³ was blown away. Andesitic to basaltic ash layers from Tambora were found in the Indian Ocean up to 600 km towards the south, but not towards the north (Hutchison 1989b). Ash layers 60 cm thick, were found at a distance of 70 km. This Plinian eruption killed 92,000 people (van Bemmelen 1949). The tremendous amounts of ash thrown into the atmosphere resulted in an abnormally cold summer in the Northern Hemisphere.

Krakatau. The eruption of Krakatau in 1883 was the fourth-greatest eruption in recorded history, killing 36,417 people. It has been estimated that 18 km³ of pumice and ash erupted (Hutchison 1989b) with an energy expenditure of 1 x 1025 ergs. The coasts of Java and Sumatra were hit by at least 19 tsunamis, with the largest estimated at 30-40 m high. The clouds of dust, gases, and debris rose 27.2 km into the atmosphere, colouring sunsets for 2 years.

Initial phreatic activity began in May 1883 and continued through until early August. The paroxysmal eruption began on August 26, with a Plinian eruption column approximately 20 km high. Several loud explosions were heard on the morning of August 27. Pyroclastic surges and flows, rich in pumice, were the principle products. The discharge of pyroclastic flows into the sea are thought to be responsible for 20% of all volcanogenic tsunamis (Latter 1981).

The eruptions radically altered the size and layout of the Krakatau islands. Prior to the 1883 eruption, Krakatau was a lush island known as Rakata Besar with three volcanic peaks; Rakata, Danan, and Perbuatan. Danan, Perbuatan, and the northwestern section of Rakata were destroyed by the 1883 eruption. The formation of a 300-m-deep, 6-km-wide caldera destroyed much of Krakatau Island, leaving only a portion of Rakata, and the peripheral islands of Sertung and Panjang. To the north, two islands were created on a broad, shallow marine bank. Steers (elevation 3 m) and Calymeyer (elevation 6.5 m) were eroded to below sea level in approximately one year (Sigurdsson et al. 1991).

Evidence suggests that the Krakatau caldera is a collapse caldera due to the similarity in volumes of the caldera and the deposits. The volume of the caldera was estimated at 8.9 km³. The dense-rock equivalent volume of the total erupted products comprising subaerial falls, subaerial flows, and submarine flows, is 9.0 km (Sigurdsson et al. 1991).

Although Krakatau continues to receive significant attention from scientists from Indonesia and worldwide (a symposium entitled "The Symposium on 100th Year Development of Krakatau and its Surroundings" was held in Jakarta from 2327 August, 1983 commemorating the 100th anniversary of the Krakatau eruption), most of the focus has been on the terrestrial facets of Krakatau.

Figure 2,12. Distribution of active volcanoes in Indonesia.

Modified from Hutchison 1989b, and Katili 1985.

KRAKATAU BELOW SEA LEVEL. Umbgrove (1947) briefly discussed the effect of the 1883 eruptions on the coral reefs. Umbgrove's findings agreed with Davis (1928) that corals do not grow well on loose volcanic substrate. The subject of coral reef recolonization on lava flows is covered in greater depth in this book; in the chapter on natural disturbances.

The nature of the submarine deposits around Krakatau had not been studied until recently. According to Sigurdsson et al. (1991), "the details of the eruptive processes that generated the deposits and the relationship of these processes to the origin of the deadly tsunamis have been a source of controversy and much speculation." The study carried out in 1990 focused on the Krakatau submarine deposits.

The results of Sigurdsson et al. (1991) suggested the following:

· The submarine deposits from the 1883 Krakatau eruption are of pyroclastic origin, essentially identical to the subaerial deposits found on the Krakatau islands.

· The pyroclastic flows entered the sea in all directions, with the greatest amount of pyroclastics deposited west of Sertung.

· Steers and Calymeyer were ephemeral islands which only remained above water for a year following the eruptions. They were built of submarine flow deposits on a shallow section of sea floor.

· In the shallow seas surrounding Krakatau, a total of 13.6 km of pyroclastics was deposited, with 0.4 km³ redeposited in subsequent erosion.

· The closely matching volumes of erupted magma (within the mapped area of 1,030 km²) and the caldera, support the collapse model of caldera formation.

Historical observations of hot, subaerial flows that traveled over 50 km across the ocean surface suggest that gravitational segregation of proximal pyroclastic flows led to the development of a dense basal flow that traveled underwater, and a more dilute and buoyant upper part that was able to remain above sea level for great distances. — SIGURDSSON ET AL. 1991

PRE-ERUPTION BATHYMETRY AROUND KRAKATAU. Prior to the 1883 eruption, a broad platform at a depth of 40 m was located to the north of Krakatau. The platform was separated from the volcano by a concentric moat with a maximum depth of 60 m. The sea floor to the south and west of Krakatau had gentle slopes to depths of more than 100 m.

POST-ERUPTION BATHYMETRY (1990) AROUND KRAKATAU. Verbeek (1885) estimated that the 1883 eruption deposited 12.5 km of volcanic material in the Sunda Straits. The submarine deposits surrounding Krakatau are largely of pyroclastic flow origin as confirmed by Sigurdsson et al. (1991).

Pyroclastic material from the 1883 eruption appears to have been deposited concentrically around Krakatau with a mean thickness of 20 m, and a maximum thickness of 90 m to the west due to prevailing wind conditions during the eruption. The survey by Sigurdsson et al. (1991) indicated that Steers and Calymeyer are now wave-cut platforms 10-15 m below sea level. Hummocks, in this case pyroclastic mounds, are found in the region between the platforms and the islands of Sertung and Panjang, where they are 6-10 m high, spaced 150-500 m apart. The moat surrounding much of the Krakatau island group was not filled in during the eruption and continues to exist today. The moat is generally 50-60 m deep but in some areas, large pyroclastic mounds 18 m high, reach within 10-15 m of the surface.

THE PYROCLASTIC DEPOSITS. The volume of the deposits within the mapped area of 1030 km² was estimated in 1990 at 13.2 km³ by Sigurdsson et al. (1991). Erosion is primarily responsible for the differences of up to 21 m in the contours of the 1886 and 1990 bathymetric charts. The amount eroded varies inversely with the original depth to the top of the deposit, with most erosion occurring in less than 15 m. Data derived from 50 cores, indicates that the mean erosion of the primary pyroclastic deposit was 29%.

Sigurdsson et al. (1991) have estimated that 0.4 km³ of the pyroclastic deposit has been eroded and redeposited, giving a total volume of 13.6 km which is only slightly higher than Verbeek's (1885) estimate of 12.5 km.

The 1883 submarine deposits consist of pumice-rich dacitic pyroclastic flows with a thin veneer of reworked pyroclastic material. Although there is some variation, generally the deposits are poorly sorted, containing angular pumice fragments, and lithic clasts in a light grey silty ash matrix. The size of the fragments varies from 10 cm to more than 1 m.

ANAK KRAKATAU. The present-day Anak Krakatau is the fourth to bear this name. Its predecessors were each destroyed by wave erosion soon after formation. According to van Bemmelen (1949), Anak Krakatau IV has been active since 29 December 1927. It first appeared as an island in 26 January 1928, building up in the centre of the caldera, and has been more or less active since then.

Krakatau displays a periodicity in its eruptive activity involving three cycles of magmatic differentiation (increasing silica content of the eruption products), two cataclysmic outbursts at the end of the cycle, followed by the collapse of the volcano. According to van Bemmelen (1949) it takes several centuries for the "chemical composition of the eruption products" to change from basaltic to dacitic-rhyolitic so "no repetition of the calamity of 1883 has to be feared in the near future."

TECTONICS OF THE INDONESIAN ARCHIPELAGO

Starting from a number of continental nuclei the sialic crust has grown in thickness and extent during the geological evolution of this region. This process of growth of the sialic crust was accompanied by orogenic revolutions, which have welded new rigid belts to the continental shields. The Pacific Mountain System had its main development in the Lower Mesozoic and consolidated the continental area of SE Asia and the older circum-Australian girdle. In the intervening region between Asia and Australia the process of mountain building continued and is still in full swing. It is to be expected that in the geologic future this region will also consolidate, forming a continental bridge between Asia and Australia.—VAN BEMMELEN 1949

Introduction

The Southeast Asian region comprises Indonesia, Thailand, the Philippines, Malaysia, Singapore or all the association of South-East Asian (ASEAN) countries, plus Myanmar (formerly Burma), and Indochina (Kampuchea, Laos and Vietnam). Three of the seven major tectonic plates of the world, the Eurasian, Pacific and Indo-Australian Plates, are shaping this region, in addition to several smaller plates such as the Philippines, Burma, Moluccas, Bismarck, Solomon and Caroline Plates. Some countries, Indonesia for example, are situated on more than one tectonic plate.

The tectonic activity in west Indonesia differs significantly from the events in east Indonesia. Subduction tectonics characterize west Indonesia with frontal subduction south of Java, and oblique subduction west of Sumatra. Collision tectonics between the Eurasian and Indo-Australian Plates dominate southeast Indonesia. Subduction tectonics are important in northeast Indonesia (Zen 1993).

The Cratons. A craton is a part of the earth's crust that has attained stability and has been little deformed for a prolonged period (Bates and Jackson 1980). The Sunda craton, or Sundaland, is the Southeast Asian continental part of the Eurasian Plate (Hutchison 1989b). The continental Australian craton is found on the Indo-Australian Plate, and extends from Australia to New Guinea. The area between the Sunda Platform and the Sahul Platform is known as Wallacea, first named by R. E. Dickerson in 1928, after the naturalist Alfred Russel Wallace.

Plate Motions

The Indo-Australian Plate. The Indo-Australian Plate shares boundaries with several plates: the Eurasian (Southeast Asian), Burma, Caroline, Pacific, Bismarck and Solomon Plates.

The Indo-Australian Plate is subducting beneath the Burma Plate at a rate of 5 cm a^-1 (Addicott and Richards 1981) off the coast of Sumatra. This area of subduction is noted for its absence of deep earthquakes. The Benioff zone does not appear to extend beyond 200 km (Hatherton and Dickenson 1969, Newcomb and McCann 1987).

From the Java Trench to the Timor and Tanimbar Troughs, the Indo-Australian Plate is converging with the Southeast Asian section of the Eurasian Plate. The Benioff zone is well defined, steep and extends to depths of 600 km. Most of the active volcanoes lie above a zone defined between the 100 and 200 km depth contours (Hutchison 1989b).

East of the Aru Basin, the Indo-Australian Plate is in contact with the Caroline/Pacific Plate, and in Papua New Guinea contact is with the Bismarck and Solomon Plates (Petroconsultants Australasia 1991).

The Eurasian Plate. The Eurasian (Southeast Asian) Plate is being subducted into four trenches:

· Cotabato/West Sangihe Trench;

· North Sulawesi Trench;

· Manila Trench, Philippines;

· Negros Trench, Philippines.

However, it is, at present, the overriding plate in its convergence with the Indo-Australian Plate.

The Pacific Plate. The Pacific Plate proper does not appear to have a major influence on the tectonics of Indonesia. However, the microplates sandwiched between the converging Eurasian, Indo-Australian, and Pacific Plates are moved by boundary interaction with the megaplates (Hamilton 1979). This interaction by the Indo-Australian and Pacific Plates is accompanied by subduction, major rotations, transform faults, and minor spreading. According to Hamilton (1979), "the eastern Indonesian zone of interaction between all three megaplates is still more complex." An interesting and highly complex area of convergence is the Banda Sea region where the Indo-Australian, Caroline and Eurasian Plates meet to form a triple junction.

Historical Setting

Nishimura and Suparka (1990) interpret the structural features of Indonesia in terms of collisions between island arcs and continental margins. But the complete tectonic history of Indonesia is still far from being solved. Further field work and research are needed to unlock some of the mysteries of Indonesia's tectonic history. Before delving into a detailed study of the complex plate tectonics which are shaping Indonesia, we will familiarize ourselves with the historical geological setting.

Hutchison (1989b) states that "there is compelling evidence that all the terrains of Southeast Asia were an integral part of Gondwanaland and most probably were attached to the northern Australian margin" during the early Paleozoic. The evidence referred to by Hutchison, are stratigraphical and paleontological records found in Cambrian and Ordovician rocks. It is believed that the blocks containing Eurasian fossils of Permian age rifted away from Gondwana during the Ordovician or Silurian. Those rocks displaying Permian Gondwana fossils remained attached or close to Gondwana until the Carboniferous-Permian (Hutchison 1989b). Audley-Charles (1983) and Pandolfi (1993), however, are of the opinion that all terrains remained attached to Gondwana until the Jurassic rifting event (160 Ma [million years ago]). Audley-Charles explains the differences in fossil assemblages as representing either coastal (Eurasian) or inland (Gondwana) zones of the supercontinent.

The two halves of the Malay Archipelago (Sumatra, Java, Borneo, Sulawesi, Moluccas and Timor; usually also including the Philippines, and sometimes New Guinea) remained far apart from 200 Ma (early Triassic) to (50 Ma) middle Eocene (Audley-Charles 1981). From the mid-Miocene (15 Ma) to the Pliocene (5 Ma) a progressive collision, the mid-Miocene collision, started between the Australian/New Guinea component of Gondwana and the Asian component of Laurasia (Audley-Charles 1981). East Indonesia was largely created by this great mid-Miocene collision (Pandolfi 1993).

The Dispersal of Gondwana and Laurasia. Gondwana separated from Laurasia 180 Ma with the opening of the Atlantic Ocean between Africa and North America (Pitman and Talwani 1972). The breakup of Laurasia and Gondwana resulted in the closing of the Tethys Ocean (Audley-Charles et al. 1981). The present-day Mediterranean Sea is the only major remnant of the Tethys Ocean (Brown et al.1989). Laurasia was then split by the separation of North America from Greenland and of both from Eurasia (Smith and Briden 1977).

Gondwana broke up 140 Ma with the northward drifting of India. India separated from Africa and South America, as well as Antarctica-Australia/New Guinea. At this time Australia/New Guinea also included part of eastern Indonesia, and some of the southwest Pacific islands (Audley-Charles et al. 1981). India's northward drift was relatively rapid, with estimated rates at 35-175 mm per year (Johnson et al. 1976). India collided with Eurasia 55 Ma in the mid-Eocene (Klootwijk and Peirce 1979). The present Indonesian Archipelago is composed of Gondwana and Laurasian landmasses (fig. 2.13).

The Early History of the Indonesian Archipelago

The first major step in the formation of the Indonesian Archipelago, as it exists today, began 53 Ma with the separation of Australia/New Guinea from Antarctica. We know that Australia/New Guinea began its northward trek 53 Ma from microfossils which date the growth and spreading of the Indian-Antarctic Ridge (Johnson et al. 1976), and from paleomagnetic data (Smith and Briden 1977). The northward movement of the Australian Plate continues today as a result of spreading activity on the Southeast Indian Ridge between Australia and Antarctica.

Sundaland. Sundaland refers to the landmass of Southeast Asia, which was above sea level during the Pleistocene lows in sea levels. During that time, land was continuous from peninsular Southeast Asia across to Sumatra, Java and Borneo, and the western shallow part of the South China Sea as far east as the Natuna islands (Hutchison 1989b). According to Katili (1984), west Sulawesi was also a part of Sundaland, but the early Tertiary opening of the Makassar Basin caused western Sulawesi to rift away from Borneo.

Continental Sundaland is composed of an assortment of continental blocks which rifted from Gondwana in pre-Mesozoic time. The collisions and suturing of the blocks to form Sundaland was completed by late Triassic-early Jurassic (Hutchison 1989b). The Sunda craton is now aseismic, except for activity on its active margins, and some faults within Sundaland.

Java. Late Cenozoic volcanic rocks rise above Neogene marine strata on the island of Java. The basement rocks are late Cretaceous or early Tertiary melange with some ophiolites. There is no evidence of old continental crust (Hamilton 1979).

Sumatra. Sumatra forms an integral part of continental Sundaland. The island is composed of a Cenozoic volcanic arc built upon a Lower Paleozoic continental crust (Hutchison 1989b).

Borneo. Southwest Kalimantan may be an extension of the Malay Peninsula (Hartono and Tjokrosapoetro 1986), or a continental fragment derived from Indo-China either during the mid-Cretaceous (Hutchison 1989b) or late Cretaceous (Gatinsky and Hutchison 1984; Metcalfe 1988). Southeast Kalimantan is a micro-continental fragment (Hutchison 1989b) which was joined with southwest Sulawesi prior to the opening of the Makassar Strait in the Oligocene (McCabe and Cole 1989).

Figure 2.13. Map of the Indonesian Archipelago indicating the Gondwana (shaded areas) and Laurasian affinities.

Modified after Audley-Charles 1981.

The Sunda Arc

The Sunda Arc extends from the eastern Himalaya Mountains, through western Myanmar, the western Andaman Sea, Sumatra and Java, and continues into the Banda Arc of eastern Indonesia. It represents a classic plate-edge convergent or subduction margin, with active andesitic volcanism along most of the arc (Curray 1989). The orientation of the plate margin varies from approximately easterly in the Java sector, to northerly in the Andaman-Nicobar sector, hence the oblique subduction off the coast of Sumatra (Hutchison 1989b). South of Java, the Indo Australia Plate is moving northward at a rate of 7-7.5 cm/yr., therefore "Australia could be one kilometre closer to Indonesia in as little as 20,000 years" (The Globe and Mail, 12 December 1994).

The classic features of a subduction zone are well developed off the coasts of Java and Sumatra. Present are trench, fore-arc ridge, fore-arc basin, volcanic arc (island arc), and back-arc basin.

The Sunda Trench is typical of subduction trenches of continental margins or mature island arcs. The inner and outer walls of the trench only slope by approximately 7° (Hamilton 1989). Young or mid-ocean island arc systems where there is little sediment-feeding an accretionary wedge, may have steeper walls. The Ganges and Brahmaputra Rivers supplied most of the Sunda Trench sediments as far southeast as Java (Moore et al. 1982). Those sediments were transported more than 3000 km but the source has now been cut off as a result of the collision of the Ninety-east Ridge and the trench in the Andaman sector (Hamilton 1989). Off the coast of Java, the Java Trench is 6-7 km deep. The trench located in offshore South Sumatra is approximately 5 km in depth, shallowing up to 3.5 km, west of the Nicobar islands (Hutchison 1989b).

Hamilton (1989) estimates that 3000 km of Indian Ocean lithosphere has been subducted into the Sunda subduction system during the 30 million years of its operation. The crust now being subducted into the trench south of Java has been dated as late Cretaceous (Johnson et al. 1976).

The Sunda accretionary wedge is approximately 15 km thick, and located 75150 km from the trench (Hamilton 1989). The fore-arc ridge is the crest of the accretionary wedge debris accumulated in front of the leading edge of the overriding plate (Hamilton 1989). The fore-arc ridge is 1-3 km below the surface off Java and South Sumatra, but along most of Sumatra the fore-arc ridge is at a depth of less than 1 km. Exposures of the accretionary wedge have been studied on the island of Nias where the ridge is several hundred metres above sea level (Moore and Karig et al. 1980; Moore et al. 1980).

Hamilton (1989) suggests that as the leading edge is eroded, the fore-arc ridge migrates landward relative to that plate and the fore-arc basin is narrowed. Karig (1982) offers the contrary view, that the fore-arc ridge migrates seaward with time, due to growth of the stabilized accretionary wedge.

The Sunda fore-arc basin, located between the fore-arc ridge and the shoreline, is 150-200 km wide (Beaudry and Moore 1981; Hamilton 1979; Karig et al. 1980). Off the coast of Java the fore-arc ridge is 2-3 km deep, and a well-developed fore-arc basin is present.

The Sunda magmatic arc is characterized by active volcanoes above the 100 km depth contour of the Benioff zone. The magmatic arc changes from continental (silica-rich rocks) in Sumatra, to transitional (near-continental thicknesses, mélanges, and mafic/basic to intermediate rocks, e.g., pyroxene andesite, high-alumina basalt) in Java, to mature oceanic island arc (mafic/basic to intermediate rocks) in Bali and Lombok (Hamilton 1989). The composition of the young volcanic rocks reflect the composition of the crust through which the magmas have erupted. Much of Sumatra has been continental since the late Paleozoic, but Java has been constructed of post-Jurassic subduction related processes of magmatism and tectonic accretion (Hamilton 1989).

The Sunda Arc is separated from the Banda Arc by the Pan tar fracture located between the islands of Pan tar and Alor, and the Sumba fracture separating Sumba and Flores from Sumbawa (Nishimura and Suparka 1986).

The Banda Arc

The Banda Arc is generally divided into an inner and an outer arc, with the Inner Banda Arc considered an extension of the Sunda Arc that trends eastward from Java, curving northward around Banda, then westward toward Ambon (Hutchison 1989b).

The Inner Banda Arc. The Inner Banda Arc is basically a volcanic extension of western Sumatra and Java (Michaux 1991). Convergence between the Eurasian Plate and the northern margin of Australia/New Guinea resulted in the complete subduction of the oceanic crust from the northern margin (Bowin et al. 1980). The volcanic island arc that was produced during the period of subduction is known as the Inner Banda Arc. The Inner Banda Arc is composed of the islands from Bali eastward through Sumbawa and Flores to Wetar, the Damar islands, Manuk, Banda Api, and Ambon. Also included in the Inner Banda Arc are submarine volcanoes (Emperor of China, Nieuwerkerk, and Gunung Api) which are located north of Alor and Wetar. The inner-arc islands are composed of upper Cenozoic calc-alkalic volcanic rocks (Hamilton 1979).

The Outer Banda Arc. Once all the oceanic crust had been subducted from the northern margin of the Australia/New Guinea Plate, what followed was a convergence of continental crust of Australia/New Guinea and the Inner Banda volcanic arc. This collision produced rock deformation in the collision zone, leading to uplift, and the creation of the non-volcanic Outer Banda Arc (Audley-Charles and Milsom 1974; Carter et al. 1976; Bowin et al. 1980). This outer arc is primarily formed of Tertiary subduction melange and imbricated complexes (Hamilton 1979). Fossiliferous Australian continental margin rocks have been found underlying the Outer Banda Arc islands of Sawu, Roti, Timor, Leti, Babar, Tanimbar, Kai, Seram, Buru, and the Watubela islands located between Kai and Seram.

The Banda Arc Trench overlies continental crust, in contrast to the Sunda Arc Trench which overlies oceanic crust (Hamilton 1989). The trench is more than 6000 m deep south of Bali and Sumbawa, 5000 m deep north of Buru, and elsewhere averages 2000 m deep (Katili 1991).

The accretionary wedge is well below sea level from Java to Flores but eastward it rises high on the continental crust, visible on the islands of Flores, Tanimbar, Kai, Seram and Buru. Uplifted Quaternary reefs with elevations of 1000 m above sea level are a result of thickening of the accretionary wedge, and "ramping farther onto continental crust" (Hamilton 1989).

The fore-arc ridge comprises the islands of Sawu, Roti, Timor, Tanimbar, Kai, Seram and Buru. These islands are primarily composed of subduction melange, and imbricated complexes of Tertiary age (Katili 1991). The fore-arc basin is continuous except at Sumba. The basin deepens symmetrically along the two limbs of the Banda Arc to reach its maximum depth (7.5 km) at the axis of the curve, the Weber Deep. Other basins include the Lombok Basin and the Sawu Basin. Sumba may be a minicontinent rifted from the Java Sea Shelf, a fragment from northwest Australia (Hamilton 1989; Chamalaun et al. 1982), or a section of uplifted fore-arc basin (Silver et al. 1983).

The width and volume of the magmatic arc decreases eastward along the south limb of the Banda Arc and corresponds to a decreasing age of the commencement of magmatism, from early Miocene in the west to Pliocene in the east (Hamilton 1989). The volcanic arc consists of the islands of Bali, Sumbawa, Flores, Wetar, Damar, and Banda Api.

The Banda Arc and Plate Motions

The Indo-Australian Plate is being subducted beneath the southern Banda Arc subduction zone. Convergence began with the separation of Australia/New Guinea from Antarctica 85 Ma, but it was only in the Pliocene that continental crust collided with the Banda Arc (Hutchison 1989b). As a result of this Pliocene collision, Australian continental crust now underlies the outer slope of the Timor-Tanimbar Trough (Charlton 1991).

The stratigraphy of Tanimbar suggests that subduction ended during the late Pliocene. The coastline of Tanimbar features extensive raised reefs which indicate that uplift has occurred and the island is rising. Charlton (1991) offers the explanation that the partially subducted continental lithosphere has detached from the subducting oceanic lithosphere and is rebounding isostatically towards the earth's surface. The collision of arc island and continent is at a later stage on the island of Timor. It is believed that the collision began in late Miocene (Berry and Grady 1981; Berry and McDougall 1986). Timor is also showing signs of uplift with Quaternary reefs that can be found with elevations of 1300 m (Rosidi et al. 1981). These uplifted reefs are discussed in more detail in chapter 16, Atolls and Raised Islands.

Subduction of oceanic lithosphere in the Timor Trough continued until the arrival of the Australian continental lithosphere. The relatively buoyant continental crust was not easily subducted, resulting in an end to subduction and extinction of the volcanic arc to the north of Timor (i.e., Alor, Wetar, and Romang). The collisions continue, therefore most of the Banda Arc is now in the process of extinguishing (Hutchison 1989b).

The Lesser Sunda Islands (Bali west to Timor). The size of the islands decreases as you travel from west to east through the islands of Bali, Lombok, Sumbawa, Flores, and Wetar. This size decrease is especially marked east of Wetar, and may be a reflection of the amount of oceanic crust subducted, or an indication that the present volcanic arc east of Wetar is younger or that the original volcanic arc east of Wetar has been overridden by the Australian continental margin (Bowin et al. 1980).

The subsequent collision of Australian continental crust with the islands of Alor and Wetar to create Timor occurred in the early Pliocene after all the oceanic lithosphere had been consumed by subduction. Post-collision land has also arisen in the collision zone. In the islands east of Java the oldest rocks are Miocene and great uplift occurred during the Plio-Pleistocene.

The Moluccas/Maluku. The sinuous distribution of the Moluccas is due to a continuous 10 million-year period of collision, from 15 to 5 Ma, between the irregular northern edge of the Australia/New Guinea continent and the island arcs to the northwest (Audley-Charles 1981).

The northern group of Moluccan islands in the vicinity of Halmahera is geologically related to the northern Australia/New Guinea margin. The southern group, centred on Seram, is part of the rift system connected to northwestern Australia (Michaux 1991). Halmahera was included in Parker and Gealey's (1985) outer Pacific Arc, because of its similarities with Irian Jaya north of the Sorong Fault (Dow and Sukamto 1984). Bakan, Obi and Buru are derived from the edge of the Australia/New Guinea craton (Michaux 1991).

New Guinea (Irian Jaya). Central and southern New Guinea sit upon Australian continental crust (Sander and Humphrey 1975). Northern New Guinea is underlain by continental margin "transitional" crust locally overthrust by ophiolites (Bain 1973), resulting from the collision 15 Ma of Australia/New Guinea with a Tertiary island arc that now forms the north coast ranges (Johnson and Jaques 1980). Pigram and Davies (1987) identified 32 tectono-stratigraphic terranes of both continental and oceanic affinities within the New Guinea Orogen. Some terranes were formed in distant ocean basins during the Eocene, and did not attach to the New Guinea landmass until the late Oligocene. Pigram and Davies (1987) suggested that, although no terrane accretion has occurred to the east of Cenderawasih Bay (formerly Sarera Bay) in northwest Irian Jaya since the Pliocene (2 Ma), the opening of the Woodlark Basin is currently breaking up the eastern end of the East Papua composite terrace (Pandolfi 1993).

Approximately 45% of the New Guinea terranes have continental affinities with most of the rocks displaced from the relative proximity of the northern edge of the Australian craton. Other terranes, however, were formerly parts of Gondwana that were detached early in the Mesozoic and evolved away from the Australian craton before accreting in the mid- to late Oligocene. The oceanic terranes were derived from a wide diversity of environments, including plateaus, seamounts, island arc complexes, and abyssal depths as exemplified by deepwater cherts and carbonates (Pandolfi 1993).

Now that we have discussed the tectonic history of western and eastern Indonesia, there remains one island, which does not completely fit into either division…the centrally located island of Sulawesi.

Sulawesi

The history of the oddly shaped island of Sulawesi has long baffled geologists. The currently favoured hypothesis suggests that Sulawesi was the site of a Miocene collision between an east-facing island arc and a passive Paleozoic microcontinent (Sula microcontinent) situated on the downgoing plate (McCabe and Cole 1989). This hypothesis is based on the work of Hamilton (1973, 1978, 1979), Katili (1978), and Silver et al. (1983). Western Sulawesi was originally a part of Sundaland, while eastern Sulawesi is composed of subduction melange (Hamilton 1979; Katili 1978), and ophiolite (Hamilton 1989). Uplift occurred when a continental fragment of Australia/New Guinea known as the Banggai-Sula Spur, or alternatively the Sula Platform, Sula Peninsula, or Sula microcontinent, collided with the rest of Sulawesi (Audley-Charles 1981). The collision has been dated as mid-Miocene (15 Ma) based on the age of overthrusting and imbrication (or shingling, from the Latin word imbrex, meaning a roof-tile) in the east arm of Sulawesi by Kundig (1956), and as late Miocene by McCabe and Cole (1989). It is believed that the Banggai-Sula Spur/Sula Platform rifted from New Guinea sometime between the Cretaceous and late Miocene (Pigram et al. 1985), while Hamilton (1979) dates the movement by strike-slip fault along the Sorong Fault of the Banggai-Sula microcontinent as mid-Tertiary. It is not known with much accuracy from where the Banggai-Sula Spur originated, although it is known that the microcontinent rifted from the western two-thirds of New Guinea. It has been speculated, based on stratigraphic comparisons, that the Banggai-Sula Spur came from north of Misool (Audley-Charles et al. 1972), the Bird's Head of Irian Jaya (Norvick 1979), the Bird's Neck of Irian Jaya (Hamilton 1979), or from central New Guinea (Pigram et al. 1985).

Located off Southeast Sulawesi, the islands of Buton, Muna and Wowoni are believed to have a similar history as Banggai-Sula, representing a continental fragment that originated from New Guinea (Hamilton 1989). Buton shares a similar stratigraphy with the islands of Buru and Seram (Wiryosujono and Hainim 1978).

Audley-Charles et al. (1972), Hamilton (1979), and Hutchison (1989a), have suggested that western Sulawesi was during times of low sea level connected to Borneo. The most likely areas for an inter-island connection are across the southern and central parts of the Makassar Strait where presently there is only a narrow channel of deep water. Sea level changes such as those in the late Pliocene and Pleistocene would have at times exposed a land bridge between Borneo and Sulawesi (Audley-Charles 1981). Dating of marine sediments in the Makassar Strait suggest that the strait has been a marine basin since the Eocene and possibly as far back as the Cretaceous. Hamilton (1979) postulated that Sulawesi rifted eastward in response to back-arc spreading under the southeast section of Sundaland during the mid-Tertiary. This spreading resulted in the opening of the Makassar Strait. According to Hutchison (1989a) the western arm of Sulawesi, as well as the Paternoster Block (now located off the east coast of Kalimantan), rifted from Sundaland during the opening of the Makassar Strait.

The northern arm is a young volcanic arc built on oceanic basement (Hutchison 1989b). The North Arm rotated with respect to the South Arm during late Tertiary. The work of Otofuji et al. (1981) suggested that the North Arm has rotated more than 90° between Eocene and early Miocene. Plio-Pleistocene paleomagnetic directions from the North Arm are indistinguishable from the present dipole direction, suggesting that the North Arm has not undergone any significant shift in position since that time. The rotation is believed to be a result of the collision between the Sula microcontinent and Sulawesi during the Miocene (Silver et al. 1983; McCabe 1984).

The similarity of the paleopole positions for the South Arm of Sulawesi and southeastern Borneo suggests that the South Arm of Sulawesi was in close proximity to Borneo and a part of the same tectonic block since the Cretaceous (Haile et al. 1977 and Sasajima et al. 1980). Both areas have undergone an early to middle Tertiary counterclockwise rotation which was completed by Miocene time.

Paleomagnetic data (Haile 1978) from Mesozoic cherts in the Southeast Arm indicate that the rocks were formed at much higher latitudes. These findings support the theory that some of the Southeast Arm rocks were formed in a southern Mesozoic ocean basin and traveled northward by plate motions until they were eventually accreted onto the equatorial Sulawesi island arc (McCabe and Cole 1989).

Katili expresses a different view of the geological history of Sulawesi and the Makassar Strait. Katili (1978) believes that the Makassar Strait was open during the pre-Pliocene, closed in the late Pliocene, and then reopened by post-Pliocene rifting. Katili (1991) suggests that Sulawesi emerged as a double island arc east of Kalimantan during the Miocene. This view is in contrast to Hamilton (1979) who viewed Sulawesi as being close to Borneo until mid-Tertiary rifting opened the North and South Makassar Basins. Late Miocene to early Pliocene, the east and west arms of southern Sulawesi spread apart, along a rift zone which is now the 2 km deep Gulf of Bone (Hamilton 1978; Norvick 1979). Audley-Charles et al. (1972) separated the eastern and western arms of Sulawesi by placing the east and southeast arms at the margin of continental Australia during its drift northwards.

Charlton (1991) has proposed that the Gulf of Bone is an example of extensional crustal collapse following arc-continent collision. The Southwest Arm is a Miocene or earlier volcanic arc, and the Southeast Arm is a lower-middle Miocene arc-continent collision complex. Uplift in the continental metamorphic rocks in the Southeast Arm has been dated as 16.5-20 Ma (lower-middle Miocene) and it has been suggested that this uplift is related to extension in the Gulf of Bone (Helmers et al. 1990, oral comm. in Charlton 1991).

The Wallace Line

Wallace's Line is studied in virtually all biology classes, but the position of the line is geologically-dependent, a result of plate tectonic movements. The zoogeographical boundary known as Wallace's Line, separates the Oriental (Asian) and the Australian fauna and flora. Alfred Russel Wallace, in 1863, proposed that the division ran between Bali and Lombok, extending between Borneo and Sulawesi (formerly known as Celebes), and between the Philippines and Indonesia (Hutchison 1989b). In 1910, he revised the line (fig. 2.14) so that it lies east of Sulawesi (George 1981). This revised position of Wallace's Line fits more closely with the current thinking on the history of plate tectonics in the Indonesian Archipelago.

How did geology determine the position of the Wallace Line?

The mid-Miocene to Pliocene collision of Australia/New Guinea with Asia 15 5 Ma, occurred in the region of Wallace's Line. "The collision brought two originally separate faunas and floras into direct contact, ultimately giving rise to the present-day distribution of plants and animals" (Audley-Charles et al. 1981).

Geological History and Biography

Three types of geological events may have significant biogeographical consequences: (1) longitudinal displacement, (2) land connections and sea barriers, and (3) sea level history and speciation.

1. Longitudinal Displacement. Displacement of faunas may have occurred in relation to the microcontinents rifted from eastern Gondwana that now form Southeast Asia, and through motion of the numerous terranes that now compose the northern portion of New Guinea (Pandolfi 1993). The Southeast Asia micro-continents are significant because they may have transported tropical marine faunas as they moved north and westward (i.e., New Guinea marine faunas to Southeast Asia). McKenna (1973) coined the term "Noah's Arks" for faunas which have been transported passively on geotectonic plates.

Figure 2.14. Wallace's Line. The original (1863-80) and revised (1910) positions of the Wallace Line.

From George 1981.

2. Land Connections and Sea Barriers. Land barriers affect biogeography in two ways. First, land barriers impede dispersal of marine taxa with planktonic larvae, and second, they enhance speciation by providing a mechanism for isolating populations of marine taxa by fragmenting species ranges (Pandolfi 1993).

The following two pre-Quaternary events may have isolated populations between the Indian and Pacific Oceans in the vicinity of Wallace's Line (Whitmore 1987):

a) A substantial land barrier from Laurasia to Australia in Cretaceous time (Audley-Charles 1987) may have been created by the movement, northward drift, and rotation of Australia/New Guinea and the Southeast Asia micro-continents;

b) The second direct link according to Pandolfi (1993), or the first direct connection according to Audley-Charles (1981), between Australia and Southeast Asia was a result of the great mid-Miocene collision between Sulawesi and the Sula Peninsula/Banggai-Sula Spur. As a result of this collision there was either continuous land or only narrow sea gaps between Laurasian Borneo and Gondwanic New Guinea from late mid-Miocene (12 Ma) to late Pliocene.

In several areas, land emerged as a result of the mid-Miocene collision (Audley-Charles 1981):

• New Guinea - tracts of land up to 2000 km long emerged above the sea by mid-Miocene (Thompson 1967; Froidevaux 1974);

• West part of Sulawesi, intermittently linked to Borneo by exposure of parts of the Makassar Strait;

• Buru, Seram, and Misool appear to have been submerged but Banggai, Sula and Seram possibly began to emerge by late Miocene or early Pliocene and certainly by late Pliocene. Buru may have been exposed by late Pliocene (Audley-Charles 1981).

Audley-Charles (1981) suggests that a migration route for the terrestrial biota and a barrier for the marine biota could have been established between Sulawesi and eastern Australia/New Guinea by the latest Miocene or early Pliocene and by the late Pliocene it was probably as well-established as at present.

Shallow-water early to middle Miocene carbonates occur on most of the eastern Indonesian microcontinents and northern island arcs. Late Miocene to Quaternary reefs occur on the western part of the Sula Peninsula (Sukamto 1975), suggesting that the collision zone between Australia/New Guinea and Sulawesi involved shallow marine waters if not land above sea level. The presence of Pleistocene terraces of raised fringing reefs in East Sulawesi and Timor suggests that these areas also remained above sea level after their post-collision emergence in the mid-Miocene (Audley-Charles 1981).

Transient land may have provided a stepping stone between benthic marine, and coastal fish populations living on both sides of the Indian Ocean. The Ninety-east Ridge of the Indian Ocean was exposed during the Eocene and Oligocene (Kemp and Harris 1975). When the Ninety-east Ridge became submerged, it may have created a broad oceanic barrier (i.e., open ocean with too great a distance for the dispersal and survival of larvae), which disrupted previously mixed marine populations.

3. Sea Level History and Speciation. Most parts of the Southeast Asia microcontinents (South Tibet, Burma, Thai-Malay Peninsula and Sumatra) were above sea level during the Jurassic rifting event, and the late Cretaceous rise in sea level seemed to have had minimal effects on their emergence.

Quaternary Tectonics and Sea Level Fluctuations. Morley and Flenley (1987) reported that during the past 140,000 years, major sea level lows have ranged from about 50-150 m lower than today. Sea temperatures dropped by 2°C, 18,000 years ago. Morley and Flenley have proposed that the Quaternary minimum sea level was as low as 200 m below present sea level. Such estimates may be possible in limited areas during glacial maxima, but estimates of 120-130 m maximum are more likely (Chappell and Shackleton 1986; Shackelton 1987).

During the Pleistocene low sea level stands, there was almost continuous land between Australia and Asia. Both the Sunda Platform and the Sahul Shelf between Australia and New Guinea would have been exposed. Drops in sea level would also have connected Sumatra, Malaysia, Java and Borneo, all of which lie in the shallow seas of the Sunda continental shelf (Audley-Charles 1981). These changes in sea level could also have linked Borneo with Sulawesi in areas where the sea depth is presently less than 180 m.

BASINS

There are several definitions for the word basin, but two definitions are of particular importance to us: basin - 1) a great depression in the surface of the lithosphere occupied by an ocean (Merriam-Webster's Collegiate Dictionary Tenth Edition, 1993), and 2) a low area in the earth's crust, of tectonic origin, in which sediments have accumulated (Bates and Jackson 1980). We will first discuss the ocean/sea basins of the Indonesian region, followed by the sedimentary basins of Indonesia which are of major economic importance because of their hydrocarbon potential.

Ocean Basins – Formation to Demise

Continents must break apart so that ocean basins may form. The cycle of heating and cooling of the mantle is disrupted by a stationary continental plate. Heat, unable to escape from the upwelling magma, causes the mantle to expand, forcing upward the overlying continental crust. This uplift continues until the crust breaks apart and collapses, resulting in the formation of a rift valley. The valley fills with fresh water, forming a deep lake which eventually connects with the ocean. The freshwater lake then becomes a long, narrow ocean. A mid-ocean ridge supplies new oceanic crust, resulting in the expansion of the ocean basin. As the crust ages and cools, it begins to sink into the mantle, setting up a process of subduction. As the subduction continues, the ocean basin narrows. With the subduction of the 50 mid-ocean ridge, the ocean basin soon closes. Continued collision of the two continental plates-results in the formation of mountains (see table 2.2). This cycle of ocean-basin formation and destruction occurs over a period of 400 million years (Gross 1990).

The Formation of the Current Ocean Basins

The current spreading cycle began 225 million years ago with a single continent known as Pangaea, and a single ocean called Panthalassa. Pangaea began to split 200-180 million years ago with the Tethys Sea forming in the rift. The continents of Gondwana in the south and Laurasia in the north were formed from the breakup of Pangaea (Gross 1990). The Atlantic and Indian Oceans had appeared by about 65 million years ago, but North America, Europe, and Asia were still one landmass. The Tethys Sea became the Mediterranean Sea (Brown et al. 1989).

The Ocean Provinces. There are three main oceans at present (fig. 2.15). The oceans are divided into six provinces (Menard and Smith 1966) of the ocean floor: (1) shelf and slope, (2) continental rise, (3) deep ocean floor, (4) volcano and volcanic ridge, (5) rise and ridge, and (6) trench. The relative percentages of these provinces for the Pacific, Atlantic and Indian Oceans are illustrated in figures 2.16, 2.17 and 2.18, respectively.

The Pacific Ocean. Panthalassa shrank to half its original size and became the Pacific Ocean (fig. 2.16). The Pacific Ocean is therefore the oldest ocean existing today. It is mostly bordered by subduction zones, and is gradually decreasing in size.

Table 2.2. Life cycle of an ocean basin.

Figure 2.15. The present oceans. Percentage of ocean area occupied by the Pacific, Atlantic*, and Indian Oceans.

* includes the Arctic Ocean

Data from Menard and Smith 1966.

The joining of North and South America, and the northward drift of Australia, separated the Pacific Ocean from the Atlantic and Indian Oceans.

The Atlantic Ocean. The North and South Atlantic Basins formed as the Mid-Atlantic Ridge developed (Gross 1990). This mid-ocean ridge is causing the expansion of the Atlantic Ocean basin in an east-west direction. In the oldest parts of the Atlantic (i.e., the South Sandwich Trench in Antarctica, and small island arc systems in the Caribbean), subduction has begun. The Atlantic (fig. 2.17) will begin to close as subduction becomes more prevalent (Gross 1990).

The Indian Ocean. The youngest and shallowest ocean basin, the Indian Ocean, began with the breakup of Gondwana 125 million years ago (Gross 1990). The separation of Africa, Antarctica, India and Australia, followed by the collision of India with Asia, led to the opening of the Indian Ocean (fig. 2.18).

The evolution of the Indian Ocean was determined by three episodes of seafloor spreading as evidenced by magnetic anomalies and supported by data from the Deep Sea Drilling Project (Johnson et al. 1976). The first phase of spreading began with the separation of India from Australia, suggested by the magnetic anomalies which increase in age away from the Java Trench in a south-southeast direction towards the continental margin of Australia. Spreading rates are estimated at 3.5 to 4.8 Cm.a. The age of the sea floor in this area, known as the Argo Abyssal Plain, ranges from 131 Ma (early Cretaceous) to 161 Ma (late Jurassic), as oceanic crust continues to be subducted into the Java Trench (Hutchison 1989a).

Figure 2.16. The Pacific Ocean floor. Features are: (1) shelf and slope, (2) continental rise, (3) deep ocean floor, (4) volcano and volcanic ridge, (5) rise and ridge, and (6) trench.

Data from Menard and Smith 1966.

Figure 2.17. The Atlantic Ocean floor. Features as described in figure 2.16.

Data from Menard and Smith 1966.

Figure 2.18. The Indian Ocean floor. Features as described in figure 2.16.

Data from Menard and Smith 1966.

Magnetic anomalies in the Wallaby Plateau east of the Wharton Basin indicate a different direction of spreading, trending 30° instead of 60° as found in the Argo Abyssal Plain. Hutchison (1989b) suggests that this difference may indicate that while India was moving towards the northwest, other continental fragments may have moved more northerly. The presence of Jurassic and Cretaceous oceanic lithosphere in Borneo and the Philippines are possible evidence, although only north Palawan and the Banggai-Sula Spur are generally recognized as continental fragments from Australia (Hutchison 1989a). According to Hutchison (1989b), spreading continued with the axis aligned northeast in the Argo Plain, and east-southeast farther south until 90 Ma when Australia began rifting from Antarctica. However, other researchers (Weissel and Hayes 1972; Johnson et al. 1976; Weissel et al. 1977; Hamilton 1979; Audley-Charles et al. 1981; Lee and McCabe 1986) believe that Australia did not separate from Antarctica until approximately 50 Ma (53 Ma) when the still-active spreading ridge between Australia and Antarctic formed. The change in spreading direction at approximately 90 Ma occurred as India concluded its clockwise rotation, and north-south spreading dominated in the northern Wharton Basin and Mid-Indian Basin with the two spreading axes joined by the Ninety-east Ridge transform fault (Hamilton 1979).

From 90 to 53 Ma, spreading rates were 15-17 cm.a^-1 with the axis aligned in an east-west direction, offset by major north-south transform faults. It was during this period of north-south spreading that India moved rapidly northward, and the northern Wharton Basin and the Mid-Indian Basin opened. This spreading phase ended when activity along the Wharton axis ceased, and Australia and Antarctica separated (Hamilton 1979). Following the end of this phase, and possibly for the next 20 Ma, Antarctica, Australia, and India were on separate plates with spreading confined to east of the Ninety-east Ridge (Sclater and Fisher 1974; Johnson et al. 1976).

The third major phase of spreading in the eastern Indian Ocean began after 53 Ma, along the Southeast Indian Ocean Ridge located between Australia and Antarctica. Direction of spreading was northeast-southwest, causing the older oceanic crust in the north to be subducted into the Sumatra-Java Trench (Hutchison 1989b). The Indo-Australian Plate was created when the Indian and Australian Plates joined approximately 32 Ma (Sclater and Fisher 1974). The northward movement of Australia to its present position 15 million years ago, resulted in the present morphology of the ocean basin.

As summarized by McGowran (1978), the three major spreading events which shaped the evolution of the Indian Ocean in the area between India, Australia, and Antarctica were:

1. The development and activity of a NE/SW trending spreading ridge from the mid-late Jurassic to the late Cretaceous;

2. An active east-west trending spreading system from late Cretaceous to the Eocene;

3. Development of the present NW/SE spreading ridge beginning in the Eocene.

Marginal Seas

According to Hutchison (1989b), the presence of marginal seas "are a most characteristic feature of south-east Asia, distinguishing it from all other regions of the world." The marginal seas in this region are primarily associated with island arcs and result from the opening of back-arc or marginal basins. Two well-known examples of marginal seas forming in areas of continental rifting are the Red Sea and the Gulf of California (Ross 1977).

Banda Sea Basin. The Banda Sea is located where the Indo-Australian, Eurasian and Pacific Plates meet. It is predominantly of oceanic crust as indicated by seismic refraction, gravity, heat flow, and magnetic data analysis (Purdy and Detrick 1978; Bowin et al. 1980; Pigram and Panggabean 1983; Lapouille et al. 1985). The Weber Deep is the deepest point with depths greater than 7 km, with most of the Banda Sea 4-5 km deep, surrounded by shallower water of 1 km or less (Hutchison 1989b). The Banda Sea is divided into two basins, the north and south Banda Sea; although some researchers favour a common origin for the two basins, others, such as Hutchison (1989b), suggest that the Banda Sea Basin should be divided into two or more smaller basins of different character and age, and Silver et al. (1985) have proposed a Molucca-Pacific Plate origin for the North Banda Sea. Two main theories exist regarding the origin of the Banda Sea: 1) old oceanic crust trapped by the bending of the Banda Arc (Katili 1975) or specifically, trapped Mesozoic lithosphere of Indian Ocean affinities (Bowin et al. 1980; Karta 1985; Lapouille et al. 1985; Lee and McCabe 1986), or 2) an active marginal basin of Cenozoic age (Hamilton 1979, 1989). Réhault et al. (1991) proposed that the basement of the North Banda Sea was a remnant of a more recent oceanic crust squeezed between large strike-slip faults which are a continuation of those found onshore in Sulawesi. During the January 1992 French-Indonesian "Geobandut"/Banda Sea cruise, aboard the Indonesian research vessel Barunajaya III, rock samples (pillow lavas) from the basement of the North Banda Sea were dredged from two sites. The pillow lavas were classified as transitional basalts, close to the boundary between the alkaline and sub-alkaline series, similar to those found in the east African rift zone (Réhault et al. 1993), while the other sample was an intermediate lava oversaturated in silica, trachyandesite type, also near to the boundary between the alkaline and sub-alkaline series. The geochemistry of both samples indicates similarities to basalts from back-arc basins (Saunders and Tarney 1984). K⁴⁰ -Ar⁴⁰ isotopic dating of the samples (7 Ma) correlated with those obtained by Silver et al. (1985) for samples collected from the Lucipara Ridge. From their results, Réhault et al. (1993) concluded that the North Banda Sea Basin is composed of young oceanic crust created in a back-arc setting less than 10 Ma, which may have then been "reengaged in a westward subduction with a reactivation of the previous transform directions into sinistral strike-slip faults."

Celebes Sea Basin. The Celebes Sea Basin is an extinct marginal basin (Hutchison 1989b) with a flat sea floor, and depths of 4-5 km. The oceanic crust is covered by 2-3 km of sediments. Heat-flow measurements correspond to an age of 51 Ma (Parsons and Sclater 1977). Magnetic anomalies found in this basin have been interpreted as episodes of sea-floor spreading 65-72 Ma (Lee and McCabe 1986), or 42-47 Ma (Weissel 1980). Moving in a northwest direction, younger crust is encountered in a similar pattern as that encountered in the Banda Sea, which, according to Hutchison (1989b), suggests that the Celebes Sea also represents a piece of Indian Ocean lithosphere trapped behind younger arc-trench systems.

Sulu Sea Basins. The Sulu Sea is divided into two basins: the Outer Sulu Sea Basin has a thick sediment layer and is located in the northwest, while the Inner Sulu Sea Basin is of oceanic crust floor and is the southeast basin (Mascle and Biscarrat 1979). The two basins are divided by an extinct submerged ridge, the Cagayan Ridge (Hutchison 1989). Much of the past history of the Sulu Sea has been subducted into the Sulu Archipelago arc-trench system; however, the remaining crust, found at depths of 3.7-4.4 km, has been dated by magnetic anomalies as Oligocene by Weissel (1980) and Eocene age by Lee and McCabe (1986). Hutchison (1989b) proposes that the Sulu, Celebes, and Banda Sea Basins were a continuous ocean basin from the Cretaceous to early Tertiary, fragmented by Neogene (Miocene and/or Pliocene) arc-trench systems (Lee and McCabe 1986), and increasing in age from early Cretaceous in the Banda Sea to late Cretaceous in the Celebes Sea, to Eocene in the Sulu Sea located in the north (Lee and McCabe 1986).

South China Sea Basin. As described by Hutchison (1989b), there continues to be disagreement regarding the age and direction of spreading which opened the South China Sea. Pautot et al. (1986) suggested that the whole South China Sea Basin resulted from northwest-southeast spreading, in contrast to Taylor and Hayes (1983) who favoured north-south spreading. Ru and Pigott (1986) divided the South China Sea basin into the southwest and the east basin. Magnetic anomalies trend northeast-southwest in the southwest basin. Heat-flow and bathymetry data suggest that the age of the basin is approximately 55 Ma (Ru and Pigott 1986). The eastern half of the basin bears east-west trending magnetic anomalies, active from mid-Oligocene to early Miocene (32-17 Ma) (Taylor and Hayes 1983). Thus the southwest basin with its northwest-southeast spreading is thought to be late Cretaceous in age, while the east basin with its north-south spreading is late Eocene to early Miocene (Ru and Pigott 1986). It is thought that the South China Sea is a result of continental margin rifting, rather than back-arc spreading (Hutchison 1989b).

The Andaman Sea Basin. The Andaman Sea is an extensional basin that began opening at the rate of 3.7 cm.a^-1, 13 Ma during the mid-Miocene (Hutchison 1989b). In the central basin, spreading is north-northwest, to the north, spreading is northwest-southeast, while in the southern section, short spreading ridges are offset by transform faults. The north Sumatran volcanic arc is a result of subduction of Andaman Sea crust beneath the north coast of Sumatra (Hutchison 1989b).

Other ocean basins in our area of interest include the Flores, New Guinea and Manus Basins.

SEDIMENTARY BASINS

Classification

Classification of basins has been done by Murphy (1975), Soeparjadi et al. (1975), Bally and Snelson (1980), Horn (1980), and Kingston et al. (1983). Hutchison (1989b) summarizes the different classifications and includes a map of the main Tertiary basins of Southeast Asia. Hutchison's book, Geological Evolution of South-East Asia, may be consulted for additional information; Petroconsultants Australasia (1991) list 10 types of basins. Although some of the terms may differ, the type of basin being described is the same. In table 2.3, equivalent basin classification terms have been paired for easier reference.

The following summary of basin classification is from Hutchison (1989b) with equivalent terms from Petroconsultants Australasia (1991) in parentheses:

A) Basins on rifted lithosphere (ocean basins)

· Occur during the early stages of continental rifting;

· Include major grabens, aulacogens (failed arms of rift systems), and small ocean basins;

· Examples in Indonesia include the Tarakan, West Natuna, Kutei, North Makassar, South Makassar, and Bone Basins.

B) Atlantic-type miogeoclinal margins (rifted continental margins)

· Represent one side of a major rift system which developed into an ocean;

· Formerly known as miogeosynclines;

· Examples found in Indonesia include the Salawati and Bintuni Basins;

C) Cratonic basins (intra-cratonic)

· down-faulted or sag basins in continental interiors on pre-Mesozoic crust

· no known examples in Indonesia

D) Deep sea trenches (convergent accretionary wedge)

· Present where oceanic crust is being subducted;

· Generally relatively low sediment supply, and low geothermal gradients,therefore poor hydrocarbon potential;

E) Fore-deep or retro-arc basins (suture)

· Occur where continental crust subducts beneath continental or island arc crust at a low angle (i.e., a collision zone);

· No examples in Indonesia;

F) Fore-arc basins (fore-arc)

· Occur between an accretionary prism and a volcanic arc;

· Examples in Indonesia include the Sibolga, Bengkulu, South Java, Melawi, Ketungau, Savu, and Bula Basins;

G) Intra-arc basins (intra-arc)

· Small basins which are found within the volcanic arc and are controlled by faults and associated with trenches (Kingston et al. 1983);

· Not recorded in Indonesia

H) Marginal sea basins (wrench and rift)

· Back-arc basins where continental crust has been stretched and thinned until it splits and new crust has spread from the crack;

· Another method of formation is the trapping of older oceanic crust behind newer arc-trench systems;

· Generally of low hydrocarbon potential except along the shallow rifted margins;

I) Back-arc basins (back-arc)

· The difference between back-arc basins and marginal sea basins is that the back-arc basins do not contain new oceanic crust: the continental crust has been stretched but not to the point where spreading occurs;

· In this region, most of these basins have resulted from major wrench faulting combined with extension, behind the volcanic arc;

· Examples are the North Sumatra, Central Sumatra, South Sumatra, Sunda, Billiton sub-basin, northwest Java, East Java, and Barito Basins;

Table. 2.3. Basin classification. Comparison of terms used by Hutchison (1989b) and Petroconsultants Australasia (1991).

J) Basins on and peripheral to microcontinents

· "Basins may develop between and marginally on continental fragments which have rifted from Atlantic-type margins" (Hutchison 1984, 1986a);

· Indonesian examples include the northeast Natuna Basin, and the Paternoster Platform;

Table 2.4. Sedimentary basins of Indonesia. A list of the sedimentary basins in Indonesia including basin classification according to Hutchison (1989b) and hydrocarbon-producing basins (*) from Patmosukismo et al. (1989). Basin types: A = basins on rifted crust; B = Atlantic-type miogeoclinal margins; F = fore-arc basins; I = back-arc basins; and J = basins on or adjacent to microcontinents. Basins without designating letters have not been classified. For location of basins see map in figure 2.19.

Figure 2.19. Sedimentary basins of Indonesia. Numbers correspond to basins in table 2.4.

From Patmosukismo et al. 1989.

K) Pannonian-type (foreland)

· These back-arc basins are found on continental or transitional crust and may be associated with major strike-slip fault zones;

· No examples in Indonesia;

In general, the major tectonically-influenced processes responsible for the creation of basins are: extension, deformation and thermal response, and thermal relaxation subsequent to tectonic activity (Petroconsultants Australasia 1991).

Hutchison (1989b) gives a good overview of basin classification as well as some Indonesian examples. Patmosukismo et al. (1989) provide a more complete list of 60 sedimentary basins in Indonesia, but do not classify the basins (see table 2.4. Sedimentary basins of Indonesia, and fig. 2.19 Sedimentary basins of Indonesia). Patmosukismo et al. (1989) have placed 23 basins into western Indonesia and 37 basins in eastern Indonesia for a total of 60 basins. The western basins are characterized as being relatively large, and located onshore or in shallow seas. The eastern basins are by comparison small, and generally located in deep seas. Fifteen of the 60 basins were noted as hydrocarbon-producing basins. The Paternoster Platform (#61) was only listed by Hutchison (1989b).

Box 2.1. The significance of seismic profiles.

M. J. Evans, Jakarta, Indonesia.

Explanation of what a seismic section represents

Seismic surveying provides geoscientists with a tool to interpret the subsurface structure of the earth, both on land and at sea. In the marine environment seismic surveys are acquired using a specially-built boat, which tows an energy source and recording instruments in a cable which may be up to 5 km in length. The energy source generates sound waves, which pass through the sea and penetrate the rock strata of the earth. The sound waves are reflected from the interfaces between rock layers, and the reflected signals are recorded by sensitive instruments called hydrophones in the survey cable. This is analogous to the way in which a ship's echo sounder measures the depth of the sea from the time taken for a pulse of sound to return to the ship after reflection at the seabed.

The recorded seismic data are processed using sophisticated and powerful computer technologies, involving the manipulation of very large amounts of numerical data. This is usually done at processing centres on land, although state-of-the-art processing equipment can now be taken onboard the survey vessel and the data processed as it is recorded. The results of processing the data provide an image of the subsurface in time, since the vertical axis of the image represents the time taken for the sound pulse to travel from the ship to the rock interface at which it is reflected, and back to the ship again (this is known as two-way time).

Interpretation of seismic line

An example of a processed seismic line from the offshore environment in northeast Kalimantan is shown in figure 2.20. This seismic line is oriented southwest-northeast, and shows a buried limestone reef complex.

Figure 2.20 shows a series of strong seismic reflectors which define a buried "hill", the steep crest of which occurs at approximately 1.5 seconds two-way time (TWT). The more gently inclined flanks of the feature intersect the left-hand vertical time axis at approximately 2.5 seconds TWT. The seismic reflection configuration is interpreted to represent the external morphology of an ancient reefal build-up.

Reef initiation and growth occurred during the Tertiary era, and on the basis of stratigraphic knowledge from boreholes in the region the reef is interpreted to be of Oligocene to Miocene age. The generally flat-lying seismic reflectors that surround the reefal body and overlie it are interpreted to be fine-grained marine sediments, deposited during a rise of sea level. Reef growth was unable to keep pace with the rate of sea level rise, and ultimately it was drowned, probably during the late Miocene.

The reefal complex shown in figure 2.20 is a large feature, being up to 8 km in width, and because of its size and overall geometry it is interpreted as an isolated carbonate platform or pinnacle reef complex. Organic reef deposits are indicated by the two pinnacle-like features which occur on the southwest margin, the largest of which continued growing until at least 1.3 seconds TWT. A smaller reef pinnacle is present on the northwest flank of the platform, with its crest at

Figure 2.20. Seismic section of a carbonate build-up. Water depth is approximately 450 m.

Figure is courtesy of Perusahaan Pertambangan Minyak dan Gas Bumi Negara (Pertamina).

approximately 1.8 seconds TWT. A package of generally flat-lying seismic reflectors (figure 2.20), occupying the central area between the two described reef pinnacles, is interpreted to represent interior platform sediments, deposited in a back-reef environment. Eroded and redeposited carbonate debris, derived from the reef complex, is banked-up against the lower flanks of the platform margin.

Much of our knowledge of Indonesia's geology, and indeed the world's geology, has been derived from the activities of the petroleum and mining industries. Their exploration activities have made significant contributions to geological research in Indonesia. Another benefit of the search for economically viable resources is the technology which has been developed to support the exploration. One tool which is widely used to determine what lies beneath the surface is applied seismology or seismic surveying which, produces images known as seismic profiles.

PALEOGEOGRAPHY OF THE INDONESIAN ARCHIPELAGO

Late Jurassic-Cretaceous (160-65 Ma)

During the late Jurassic-Cretaceous, Australia was in its pre-rift Gondwana position to the south of its present location. From paleomagnetic data and stratigraphic studies, it has been determined that Sula, Timor, Sumba, Seram, and Buru were part of the northern Australia - New Guinea margin (Visser and Hermes 1962; Audley-Charles et al. 1972; Hamilton 1973,1978,1979;-Katili 1975,1978; Carter et al. 1976; Barber et al. 1977; Chamalaun and Grady 1978; Norvick 1979).

Mesozoic paleomagnetic data from the Malay Peninsula, Borneo, and western Sulawesi (McElhinny et al. 1974; Haile 1978; Haile et al. 1977) suggest that these regions were a single tectonic block located far to the north of Australia, and not a portion of Gondwana.

Magnetic anomalies indicate that an east-northeast-trending spreading complex existed between Australia and Sundaland (Falvey 1972; Sclater and Fisher 1974; Larson 1975; Heirtzler et al. 1978). A north-dipping subduction zone was present beneath Sunda.

Paleocene - Eocene (65 - 36 Ma)

Australia separated from Antarctica in the early Tertiary (Weissel and Hayes 1972; Weissel et al. 1977). Spreading in the ridge between Australia and Sundaland ceased in the Eocene. The Eurasian and Indo-Australian Plates appear to have been relatively stable with respect to each other. No evidence exists of a major subduction zone along the southern margin of the Sunda craton. Rifting of Sulawesi from southeastern Sundaland began. South-dipping subduction began beneath northern Borneo and southern Palawan, causing Borneo to begin a counterclockwise rotation (Haile et al. 1977), overriding Paleocene oceanic crust in the proto-South China Sea Basin. Subduction continued in southern Borneo. The Banda, Celebes, Sulu Basin spreading centre migrated northward from its Cretaceous position to equatorial latitudes. The Philippines was located southeast of its present position as an active island arc, with collisions along its eastern boundary, and subduction along its western margin (McCabe and Cole 1989).

At the end of the Eocene, the first major ice sheets appeared in Antarctica. The advance and retreat of the ice during the Oligocene and Miocene led to unstable sea levels.

Oligocene (36 - 25 Ma)

Australia continued to drift northward. The Makassar Strait, between Borneo and Sulawesi, began opening as a back-arc basin behind the eastward moving South Arm of Sulawesi (Hamilton 1979). Paleomagnetic and stratigraphic data (Haile et al. 1977; Haile 1979) indicate that Borneo continued its counterclockwise rotation until late Tertiary. The results of Haile et al. (1977) suggest 50° of counterclockwise rotation but no latitudinal change. The opening of the South China Sea began with the southward drift of north Palawan. Subduction continued along the western Philippines margin. The positions of the Philippine Arc, the various pieces of Papua New Guinea, and eastern Indonesia were very different from present positions. Major deformation was taking place in northwest Borneo, Sumatra, Sulawesi and Papua New Guinea (Petroconsultants Australasia 1991).

The mid-Oligocene was marked by relatively low sea level, with broad shelves exposed to form low-relief landmasses.

Early to Middle Miocene (25-12 Ma)

In the early-mid-Miocene, Australia maintained its northward drift. The Banda Sea was segmented by the development of a large east-west shear zone along the northern edge of the Indo-Australia Plate (Silver and Smith 1983). Buton, the Banda Ridges (Silver et al. 1985), and the Sula microcontinent (Hamilton 1979; Silver and Smith 1983) translated westward toward eastern Sulawesi. The movement of these blocks isolated the South Banda Basin from the North Banda Basin and began to separate the Celebes and Banda Seas. The Palawan Trench became inactive, possibly due to a Miocene collision event between the trench and submerged continental blocks in the South China Sea (Hamilton 1979; Ludwig et al. 1979; Taylor and Hayes 1980, 1983). The cessation of subduction in the Palawan Trench resulted in an end to the counterclockwise rotation of Borneo (Schmidtke et al. 1985). In the early Miocene, the direction of sea-floor spreading shifted from north-south to southwest-northeast in the South China Sea, then ended by mid-Miocene (Taylor and Hayes 1980,1983), along with the northward drift of the Philippines (Uyeda and McCabe 1983).

The early Miocene was a time of major marine transgression (i.e., high sea levels), relative tectonic stability in most areas, and general warming of the world's climate (Petroconsultants Australasia 1991). Major volcanic activity was occurring in the Bali-Flores portion of the island arc at the southern margin of the Sunda craton, parts of Sulawesi and the Philippines, and the area offshore from the northern margin of the Australia Plate that was subsequently accreted to form the northern part of New Guinea (Petroconsultants Australasia 1991).

During the mid-Miocene, sea levels were moderately high, and tectonic activity relatively localized. A major collision zone extended from northwest Borneo into the area now occupied by Palawan involving the northwest Palawan Plate and the Philippine Plate. Another major collision was taking place along the northern margin of the Australian Plate where crustal breakup was occurring to form the complex of microplates that later formed the Moluccas and western New Guinea (Irian Jaya). The Melanesian Arc, now extinct, lay north of the Australian Plate, and was active during the mid-Miocene. Java was mostly below sea level, except for a string of volcanoes along its southern margin (Petroconsultants Australasia 1991).

Late Miocene to Recent (12 Ma to Present)

Australia collided with the eastern edge of the Banda Arc. Timor formed along the northern edge of the collision zone. The collision between the Sula microcontinent and eastern Sulawesi was in its later stages (Visser and Hermes 1962; Hamilton 1979; Silver and Smith 1983). This collision resulted in ophiolitic bodies thrust eastward toward the microcontinent, forming the Southeast and East Arms of Sulawesi (Hamilton 1979). The polarity of the subduction zone changed from east-facing to north-facing, and the North Arm was formed (Silver et al. 1983) and the Banda and Celebes Seas isolated. The Sulu Basin was isolated from the Celebes Basin by the late Neogene subduction which produced the Sulu Archipelago (McCabe and Cole 1989).

By early Pliocene, the positions of the islands were essentially similar to their present positions. Major deformation and folding was occurring in the central Philippines, Barisan Mountains of Sumatra, Java, and in Irian Jaya along the Lengguru Foldbelt. Volcanism was common in these areas, especially along the Sumatra-Java-eastern Indonesia volcanic arc system. Java was still underwater except for the southern volcanoes.

Postscript

At the Twenty-Fourth Annual Convention of the Indonesian Petroleum Association, held in Jakarta October 1995, Robert Hall presented a new plate tectonic model for the Indonesian region. Hall incorporated new paleomagnetic data from eastern Indonesia, recent revisions of the South China Sea, and known paleomagnetic and geological information from Southeast Asia to produce his model. Figures 2.21 to 2.26, reproduced here courtesy of the Indonesian Petroleum Association, depict the evolution of Indonesia during the Tertiary. For further details on this topic, Hall (1995) and Hall (in press) are recommended readings.

Figure 2.21. Principle tectonic features of the Indonesian region at present. (Hall 1995.)

Courtesy of the Indonesian Petroleum Association.

Figure 2.22. Reconstructions of the region at 50 and 45 Ma. (Hall 1995.)

Courtesy of the Indonesian Petroleum Association.

Figure 2.23. Reconstructions of the region at 40 and 35 Ma. (Hall 1995.)

Courtesy of the Indonesian Petroleum Association.

Figure 2.24. Reconstructions of the region at 30 and 25 Ma. (Hall 1995.)

Courtesy of the Indonesian Petroleum Association.

Figure 2.25. Reconstructions of the region at 20 and 15 Ma. (Hall 1995.)

Courtesy of the Indonesian Petroleum Association.

Figure 2.26. Reconstructions of the region at 10 and 5 Ma. (Hall 1995.)

Courtesy of the Indonesian Petroleum Association.