|Environmental Changes on the Coasts of Indonesia (UNU, 1980, 53 pages)|
Although there has been geomorphological research on several parts of the Indonesian coastline, the coastal features of Indonesia have not Yet been well documented. The following account-based on studies of maps and charts, air photographs (including satellite photographs), reviews of the published literature, and our own traverses during recent years-is a necessary basis for dealing with environmental changes on the coasts of Indonesia. Coastal features will be described in a counter-clockwise sequence around Sumatra, Java, Kalimantan, Sulawesi, Bali and the eastern islands, and Irian Jaya. Inevitably, the account is more detailed for the coasts of Java and Sumatra, which are better mapped and have been more thoroughly documented than other parts of Indonesia. In the course of description, reference is made to evidence of changes that have taken place, or are still in progress.
Measurements of shoreline advance or retreat have been recorded by various authors, summarized and tabulated by Tjia et al. (1968). Particular attention has been given to changes on deltaic coasts, especially in northern Java (e.g, Hollerwoger 1964), but there is very little information on rates of recession of cliffed coasts. Measurements are generally reported in terms of linear advance or retreat at selected localities, either over stated periods of time or as annual averages, but these can be misleading because of lateral variations along the coast and because of fluctuations in the extent of change from year to year.
Our preference is for areal measurements of land gained or lost or, better still, sequential maps showing the patterns of coastal change over specified periods. We have collected and collated sequential maps of selected sites and brought them up-to-date where possible.
Coastal changes can be measured with reference to the alignments of earlier shoreline features, such as beach ridges or old cliff lines stranded inland behind coastal plains. In Sumatra, beach ridges are found up to 150 kilometres inland. The longest time scale of practical value is the past 6,000 years, the period since the Holocene marine transgression brought the sea up to its present level. Radiocarbon dating can establish the age of shoreline features that developed within this period, and changes during the past few centuries can be traced from historical evidence on maps and nautical charts of various dates.
These have become increasingly reliable over the past century, and can be supplemented by outlines shown on air photographs taken at various times since 1940. Some sectors have shown a consistent advance, and others a consistent retreat; some have alternated. A shoreline sector should only be termed "advancing" if there is evidence of continuing gains by deposition and/or emergence, and "retreating" if erosion and/or submergence are still demonstrably in progress (Fig. 4).
Coastal changes may be natural, or they may be due, at least in part, to the direct or indirect effects of Man's activities in the coastal zone and in the hinterland. Direct effects include the building of sea walls, groynes, and breakwaters, the advancement of the shoreline artificially by land reclamation, and the removal of beach material or coral from the coastline. Indirect effects include changes in water and sediment yield from river systems following the clearance of vegetation or a modification of land use within the catchments, or the construction of dams to impound reservoirs that intercept some of the sediment flow. There are many examples of such man-induced changes on the coasts of Indonesia.
Reference will also be made to ecological changes that accompany gains or losses of coastal terrain, and to some associated features that result from man's responses to changes in the coastal environment.
Incidental references to some of the coastal features of Sumatra were included in Verstappen's (1973) geomorphological reconnaissance, but there has been no systematic study of this coastline. Verstappen's geomorphological map (1:2,500,000) gives only a generalized portrayal of coastal features: it does not distinguish cliffed and steep coasts, the extent of modern beaches, fringing reefs, or mangrove areas, but it does indicate several sectors where Holocene beach ridge plains occur.
Sumatra is 1,650 kilometres long and up to 350 kilometres wide, with an anticlinal mountain chain and associated volcanoes bordered to the east by a broad depositional lowland with extensive swamp areas along the Straits of Malacca. Off the west coast the Mentawai Islands constitute a "non-volcanic arc," consisting of uplifted and tilted Tertiary formations, their outer shores being generally cliffed -facing the predominant south-westerly swell transmitted across the Indonesian Ocean-while the inner shores are typically lower and more indented, with embayments fringed by mangroves. There are emerged coral reefs and beach ridges, especially on the outer shores, and the possibility of continued tilting is supported by the disappearance of islets off the coast of Simalur even within the present century (according to Craandijk 1908: quoted by Verstappen 1973). There are, however, contrasts between the islands, the relatively high island of Nias (summit 886 metres) being encircled by emerged reef terraces suggestive of uplift with an absence of tilting, while Enggano is tabular, steep-sided, and reef-fringed. Much more detailed work is needed to establish the evolution of these island coasts, and the effects of recurrent earthquakes and tsunami. At this stage, no information is available on rates and patterns of shoreline changes taking place here.
The south-west coast of mainland Sumatra is partly steep along the fringes of mountainous spurs, and partly low-lying, consisting of depositional coastal plains. Swell from the Indonesian Ocean is interrupted by the Mentawai Islands and arrives on the mainland coast in attenuated form. It is stronger to the north of Calang, where there are surf beaches bordering the blunted delta of the Tuenom River, and south-east of Seblat, where there are steep promontories between gently curving sandy shorelines backed by beach ridges and low dunes, interrupted by such blunted deltas as the Mana, the Seblat, and the Ketuan.
Coral reefs are rare along the central part of the south-west coast of Sumatra because of the large sediment yield from rivers draining the high hinterland, but to the south there are reef-fringed rocky promontories. Pleistocene and Holocene raised beaches and emerged coral reefs are also extensive, especially on headlands near Krui and Bengkulu, where reefs raised 30 metres above the present sea level have been truncated by the recession of steep cliffs. Farther south the coast shows the effects of vulcanicity on the slopes of Rajabasa. The Krakatau explosion of 1883 generated a tsunami that swept large coral boulders onshore and produced a fallout of volcanic ash that blanketed coastal features and augmented shore deposits. Near Cape Cina the steep coasts of Semangka Bay and Tabuan Island are related to en echelon fault scarps that run north-west to south-east, and the termination of the coastal plain near Bengkulu may also result from tectonic displacement transverse to this coastline. Farther north, the Indrapura River turns parallel to the coast to follow a swale behind beach ridges before finding an eventual outlet to the sea with the Batang River.
Padang is built on beach ridges at the southern end of a coastal plain that stretches to beyond Pariaman. The extensive shoreline progradation that occurred here in the past has evidently come to an end, for there are sectors of rapid shoreline erosion in Padang Bay, where groynes and sea walls have been built in an attempt to conserve the dwindling beach. North of Pariaman the cliffed coast intersects the tuffs deposited from the Manindjau volcano, and farther north there is another broad swampy coastal plain, with associated beach ridges built by wave action reworking fluvially supplied sediment derived from the andesite cones, Ophir and Malintang, in the hinterland. Towards Sirbangis this plain is interrupted by reef-fringed headlands of andesite on the margins of a dissected Pleistocene volcano. Beach erosion has become prevalent in the intervening embayments between here and Natal, and Verstappen (1973) suggested that the swampy nature of the coastal plain here could be due to recent subsidence, which might also explain the present recession of the coast. Broader beach ridge plains occur farther north, interrupted by Tapanuli Bay, which runs back to the steep hinterland at Sibolga. Musala Island, offshore, is another dissected volcano. Next comes the broad lowland on either side of the swampy delta of the Simpan Kanang, in the lee of Banyak Island, and beyond this the coast is dominated by sandy surf beaches, backed in some sectors by dune topography, especially in the long, low sector that extends past Meulaboh.
At the northern end of Sumatra the mountain ranges break up into steep islands with narrow straits scoured by strong tidal currents. Weh Island is of old volcanic rocks, terraced and tilted, with emerged coral reefs up to 100 metres above sea level. Uplifted reefs are also seen on some of the promontories of the northern Sumatran mainland. At Kutaraja the Aceh River has filled an intermontane trough, but the deltaic shoreline has been smoothed by waves from the north-west, coming down the Bengalem Passage between Weh and Peunasu islands, so that the mouths of distributary channels have been deflected behind sand spits and small barrier islands. Beach ridges built of fluvially supplied sediment form intersecting sequences near Cape Intem, where successive depositional plains have been built and then truncated, and there is an eastward drift of beach material along the coast towards Lhokseumawe.
Within this sector Verstappen (1964a) examined the coastal plain near the mouth of the Peusangan River. He concluded that a delta had been built out north of Bireuen, only to be eroded after the Peusangan was diverted by river capture 8 kilometres to the south (Fig. 5). Following this capture, the enlarged river has built a new delta to the east. Patterns of truncated beach ridges on the coastal plain commemorate the shorelines of the earlier delta, which also retains traces of abandoned distributary channels and levees on either side of a residual creek, the Djuli. At the point of capture the Peusangan valley has since been incised about 20 metres, but the old delta was clearly built with the sea at its present level, and so piracy must have taken place within the past 6,000 years, after the Holocene marine transgression had brought the sea up to this level. The new delta has developed in two stages (A, B in Fig. 5), the first indicated by converging beach ridges on either side of an abandoned river channel, the second farther east, around the present mouth. Dating of these beach ridges could establish rates of coastal advance and retreat in this area, and show when the river piracy took place.
South-east from Cape Diamant the low-lying swampy shores of the Straits of Malacca have sectors of narrow sandy beach interspersed with mudflats backed by mangroves, which also fringe the tidal creek systems to the rear. As the Straits narrow the tide ranges increase, and river mouths become larger, funnel-shaped estuaries bordered by extensive swamps instead of true deltas. The widest estuary is that of the Kampar River, where the tide range is sufficient to generate tidal bores that move rapidly upstream. The river channels are fringed by natural levees, and patterns of abandoned levees may be traced throughout the swamps. Locally there has been tectonic subsidence- marked by the formation of lakes amid the swamps-as on either side of the Siak Kecil River and south of the meandering Rokan estuary where lakes which formed along an abandoned river channel as it was enlarged by subsidence are now shrinking as the result of swamp encroachment.
In the narrower part of the Straits of Malacca there are elongated shoal and channel systems, and some of the shoals have developed into swampy islands, as on either side of the broad estuary of the Mampar. Verstappen (1973) suggested that the Bagansiapiapi Peninsula and the islands of Rupat, Bengkalis, and Tebingtinggi may be due to recent tectonic uplift, and the Rupat and Pajang Straits to alignments of corridor subsidence. The islands have extensive swamps, but their northern and western coasts are fringed by beach ridges possibly derived from sandy material on the sea floor during the shallowing that accompanied emergence. Farther south the Indragiri and Batanghari estuaries traverse broad swamp lands, in which they have deposited large quantities of sediment derived from the erosion of tuffs from volcanoes in their headwater regions. These very broad swamp areas have developed in Holocene times with the sea at, or close to, its present level. The rapidity of their progradation may be related to several factors: an abundance of fluvial sediment yield derived from the high hinterland by runoff under perenially warm and wet conditons; the luxuriance of swamp vegetation which has spread rapidly forward to stabilize accreting sediment, and has also generated the extensive associated peat deposits; and the presence of a broad, shallow, shelf sea, on which progradation may have been aided by tectonic uplift.
In eastern Sumatra, progradation appears to have been very rapid within historical times, but there is not yet sufficient information to permit detailed reconstruction and dating of the shoreline sequences. Studies of early maps, the accuracy of which is uncertain, and interpretations of descriptions by Chinese, Arab, and European travellers led Obdeijn (1941) to suggest that there had been progradation of up to 125 kilometres on the Kuantan delta since about 1600 AD. In further papers, Obdeijn (1942a, 1942b, 1943, 1944) found supporting evidence for extensive shoreline progradation along the Straits of Malacca and in southern Sumatra. In the fifteenth century Palembang, Djambi, and Indragiri were ports close to the open sea or a short distance up estuarine inlets (Van Bemmelen 1949). More recently, the shoreline of the Djambi delta prograded up to 7.5 kilometres between 1821 and 1922, while on the east coast the fishing harbour of Bagansiapiapi has silted up, and the old Sri Vijayan ports are now stranded well inland (Verstappen 1960, 1964b).
Witkamp (1920) described hillocks up to 4 metres high occupied by kitchen middens containing marine shell debris and now located over 10 kilometres inland near Serdang, but these have not been dated. Tjia et al. (1968) quoted various reports of beach ridges up to 150 kilometres at Air Melik and Indragiri, which were former shorelines, but such features are sparser on these swampy lowlands than on the deltaic plains of northern Java. Commenting on the rarity of beach ridges, Verstappen (1973) suggested that the sandy loads of the rivers are largely deposited upstream, so that only finer sediment reaches the coast to be deposited in the advancing swamp lands. Some beach ridges were derived from sediment eroded from the margins of drier "red soil"-ta/ang-particularly around former islands now encircled by swamps, as in the Mesuji district. If progradation has been aided by emergence one would expect beach ridges to be preserved as surface features, for where progradation has been accompanied by subsidence (as on most large deltas) the older beach ridges are found buried as sand lenses within the inner delta stratigraphy. The Holocene evolution of the lowlands of eastern Sumatra still requires more detailed investigation, using stratigraphic as well as geomorphological evidence.
Patterns of active erosion and deposition alongside the estuaries north of Palembang have been mapped by Chambers and Sobur (1975). The changes are due partly to estuarine meandering, with undercutting of the outer banks on meander curves as the inner banks are built up. Towards the sea there has been swamp encroachment, for example along the Musi-Banjuasin estuary, which is bordered by low natural levees breached by orthogonal tributary creeks. The shoreline on the peninsula north of Sungsang is advancing seawards, and there is active progradation along much of the southern coast of Bangka Strait.
Bangka Island rises to a steep-sided plateau with a granite interior: like the Riau and Lingga islands to the north it is geologically a part of the Malaysian Peninsula. Pleistocene terraces occur up to 30 metres above present sea level on Bangka, and its northern and eastern shores have coralfringed promontories and bays backed by sandy beach ridges, but the southern shores, bordering Bangka Strait, are low and swampy, with mangrove-fringed channels opening on to shoaly seas. Belitung is morphologically similar, but has more exposed coasts, with sandy beach-ridge plains extensive south of Manggar on the east coast, facing the south-easterly waves from the Java Sea. Both islands have tin-bearing alluvial deposits in river valleys and out beneath the sea floor, where such valleys extended and incised during glacial low sea-level phases and were submerged and infliled as the sea subsequently rose.
South of Bangka the east-facing coast of Sumatra consists of beach ridges backed by swamps and traversed by estuaries. Lobate salients such as Cape Menjangan and Cape Serdang are beach-fringed swamps rather than deltas, but beach ridges curve inland behind swamps on either side of the Tulangbawang River, where progradation has filled an estuarine gulf. At Telukbetung the lowlands come to an end as mountain ranges intersect the coast in steep promontories bordering Sunda Strait.
In 1883 the explosion of Krakatau, an island volcano in Sunda Strait (Fig. 6), led to the ejection of about 18 cubic kilometres of pumice and ash, leaving behind a collapsed caldera of irregular outline, up to more than 300 metres deep and about 7 kilometres in diameter (Fig. 7).
The collapse caused a tsunami up to 30 metres high on the shores of Sunda Strait and surges of lesser amplitude around much of Java and Sumatra (Verbeek 1886). Marine erosion has cut back the cliffs produced by the explosive eruption: at Black Point on Pulau Krakatau-Ketjil, cliffs cut in pumice deposited during the 1883 eruption had receded up to 1.5 kilometres by 1928 (Umbgrove 1947). Since 1927 a new volcanic island, Anak Krakatau, has been growing in the centre of the caldera, with phases of rapid enlargement and outward progradation in the 1940s and early 1960s (Zen 1969).
Sunda Strait is bordered by volcanoes, the coast consisting of high volcanic slopes, with sectors of coral reef, some of which have developed rapidly in the century since the Krakatau explosion destroyed or displaced their predecessors. Panaitan Island consists of strongly folded Tertiary sediments, with associated volcanic rocks, and has a sandy depositional fringe around much of its shoreline. Similar rocks form the higher western part (Mount Payung) of the peninsula of Ujong Kulon, the rest consisting of a plateau of Mio-Pliocene sedimentary rocks. This peninsula is a former island, attached to the mainland of Java by a depositional isthmus (Verstappen 1956). It is cliffed on its south-western shores, but the southern coast has beaches backed by parallel dune ridges up to 10 metres high, covered by dense Pandanus scrub, the beach curving out to attach a coral island at Tereleng as a tombolo. The northwest coast has cliffs up to 20 metres high, passing into bluffs behind a coral reef that lines the shore past Cape Alang and into Welkomst Bay. Volcanic ash and negro heads on and behind this reef date from the Krakatau explosion, when a tsunami washed over this coast. Verstappen (1956) found that notches up to 35 centimetres deep had been excavated by solution processes and surf swash on the coral boulders thrown up onto this shore in 1883. This is rapid compared with solution notching measured by Hodgkin (1970) at about 1 millimetre/year on tropical limestone coasts. Within Welkomst Bay there are mangrove sectors, prograding rapidly on the coast in the lee of the Handeuleum reef islands. The geomorphological features of Sunda Strait deserve closer investigation, with particular reference to forms that were initiated by catastrophic events almost a century ago (cf. Symons 1888).
An island about 1,000 kilometres long and up to 250 kilometres wide, Java is threaded by a mountain range which includes several active volcanoes. To the north are broad deltaic plains on the shores of the Java Sea; to the south steeper coasts, interrupted by sectors of depositional lowland, face ocean waters.
The west coast of Java is generally steep, except for the Bay of Pulau Liwungan, where the Ciliman River enters by way of a beach-ridge plain. Near Merak the coast is dominated by the steep slopes of the Karang volcano which descend to beachfringed shores. Panjang and Tunda islands, offshore, are of Miocene limestone, but the shores of Banten Bay are lowlying and swampy, with some beach ridges, widening to a deltaic plain of the Ciujung River. This marks the beginning of the extensive delta coastline built by the silt-laden rivers of northern Java. There are protruding lobes of deposition around river mouths and intervening sectors of erosion, especially where a natural or artificial diversion of the river has abandoned earlier deltaic lobes, or sediment yield has been reduced by dam construction. A patchy mangrove fringe persists although there has been widespread removal of mangroves, in the course of constructing tambak (brackishwater fishponds), and in places these are being eroded. Some sectors are beach-fringed and the prevalence of northeasterly wave action generates a westward drifting of shore sediment. Fig.8 shows the pattern of change on the north coast of West Java detected from comparisons of maps, drawn between 1883 and 1885, and 1976 Landsat imagery: there has been seaward growth of land in the vicinity of river mouths, and smoothing and recession of the shoreline in intervening sectors.
There was rapid progradation of the Ciujung delta after the diversion of its lower course for irrigation and flood-control purposes. Growth of the new delta led to the joining of Dua, a former island, to the Javanese mainland, and this has raised problems of wildlife management, for the island had been declared a bird sanctuary in 1973, before it became so readily accessible from the land. Immediately to the west there have been similar changes on the Cidurian delta since 1927, when an irrigation canal was cut, and a new outlet established 4.5 kilometres west of the old natural river mouth. Comparison of outlines on air photographs showed that over an 18-year period the new delta built up to 2.5 kilometres seawards at the mouth of the artificial outlet, while the old delta lobe to the east was cut back by wave action which removed the mangrove fringe and eroded fishponds to the rear (Verstappen 1953a).
Changes have also taken place on the large and complex delta built by the Cisadane River. Natural breaching of levees by floodwaters led to the development of a new outlet channel, and when delta growth began at the new outlet the delta preciously built around the old river mouth began to erode, the irregular deltaic shoreline being smoothed as it was cut back (Verstappen 1953a).
Numerous coral reefs and coralline islands (the Thousand Islands) lie off Jakarta Bay, and many of these have shown changes in configuration during the past century. As a sequel to the studies by Umbgrove (1928,1929a,1929b). Zaneveld and Verstappen (1952) traced changes with reference to maps made in 1975,1927, and 1950.
Haarlem have grown larger as the result of accretion on sand cays and shingle ramparts, but there are also sectors where there has been erosion or lateral displacement of such features on island shorelines. In general the shingle ramparts have developed around the northern and eastern margins, exposed to relatively strong wave action, while the sand cays lie to the south-west, in more sheltered positions. Verstappen (1954) found changes in the position of shingle ramparts before and after 1926, on these islands, which he related to climatic variations. In the years 1917-1926 easterly winds predominated, with the ITC in a relatively northerly position because the Asian anticyclone was weak, and wave action built ramparts on the northern and eastern shores; after 1926 westerly winds became dominant, with the ITC farther south because of stronger Asian anticylonicity, and waves built new ramparts of shingle on the western shores (Verstappen 1968).
There is evidence of subsidence on some of the coral islands, such as Pulan Pugak, where nineteenth-century bench-marks have now sunk beneath the sea, while others have emerged: Alkmaar Island, for example, has a reef above sea level undergoing dissection. Some of the islands have been modified by the quarrying of coral limestone for use in road-making and buildings in Jakarta. This quarrying augmented the supply of gravel to shingle ramparts, but several islands that were quarried have subsequently been reduced by erosion: Umbgrove (1947) quoted the example of Schiedam, a large low-wooded island on a 1753 chart, reduced to a small sand cay by the 1930s.
The features of Jakarta Bay were described in a detailed study by Verstappen (1953a). The shores are low-lying, consisting of deltaic plains with a mangrove fringe inter rupted by river mouths and some sectors of sandy beach. Between 1869 and 1874 and 1936 and 1940 as much as 26 square kilometres of land was added to the bay shores by deltaic progradation, mainly on the eastern shores (Fig. 9). Detailed comparisons of maps made between 1625 and 1977 show the pattern of advance at Sunda Kelapa, Jakarta (Fig. 10). Inland, patterns of beach ridges mark earlier alignments of the coast during its irregular progradation, the variability of which has been related to fluctuations in the position of river mouths delivering sediment (Fig. 11). The beach ridges diverge from an old cuspate foreland at Tanjung Priok, across the deltaic plains of the Bekasi-Cikarang and Citarum rivers to the east.
Pardjaman (1977) published a map based on a comparison of nautical charts made in 1951 and 1976 which showed substantial accretion along the eastern shores of the bay, especially alongside the mouths of the Bekasi and Citarum rivers. This was accompanied by shallowing off the mouths of these rivers. Along the southern shores at Jakarta a fringe of new land a kilometre wide has been created artificially for recreational use by reclaiming the mangrove zone and adjacent mudflats. On the other hand, removal of lorry-loads of sand from Cilincing Beach resulted in accelerated shoreline erosion. In the 65 years between 1873 and 1938 the shoreline retreated about 50 metres but in the 24 Years between 1951 and 1975, with sand extraction active, it went back a further 600 metres (Pardjaman 1977).
East of Jakarta the Citarum River drains an area of about 5,700 square kilometres, including mountainous uplands, plateau country, foothills, and a wide coastal plain, with beach ridges up to 12 kilometres inland. It has built a large delta (Fig. 12), which in recent decades has grown northwestwards at Tanjung Karawang, with subsidiary growth northwards and southwards at the mouths of the Bungin and Blacan distributaries. At present the river heads north from Karawang, and swings to the north-west at Pingsambo, but at an earlier stage it maintained a northward course to build a delta in the Sedari sector. This has since been cut back, leaving only a rounded salient, along the shores of which erosion is continuing (Plate 1). The shores are partly beach-fringed, the beaches showing the effects of westward longshore drifting, which builds spits that deflect creek mouths in that direction. Eroding patches of mangrove persist locally and, north of Sungaibuntu, there is erosion and dissection of fishponds (PIate 21.
According to Verstappen (1953a) the Citarum delta prograded by up to 3 kilometres between 1873 and 1938, although sectors of the eastern shore of Jakarta Bay retreated by up to 145 metres. After the completion of the Jatiluhur Dam upstream in 1970 a marked slackening of the rate of progradation of the deltaic shoreline was noted at the mouth of the Citarum. BY contrast, growth on the neighbouring Bekasi delta accelerated after 1970. It was decided that dam construction had diminished the rate of sediment flow down the Citarum River because of interception of silt in the impounded reservoir whereas the sediment yield from the undammed Bekasi River had increased. Such reduction of the rate of progradation has been widely recognized on many deltaic shorelines, following dam construction within their catchments, and the onset of delta shoreline erosion is a phenomenon that has also been documented widely around the world's coastlines (Bird 1976). There is little doubt that the rate and extent of delta shoreline progradation will diminish and that shoreline erosion will accelerate and become more extensive as further dams are built in the catchments of the rivers of northern Java. This will be accompanied by increasing penetration of brackish water into the river distributaries and the gradual spread of soil salinization into deltaic lands.
East of the Citarum delta are the extensive depositional plains built up by the Cipunegara River (Fig. 13). The Cipunegara has a catchment of about 1,450 square kilometres, with mountainous headwater regions, carrying relics of a natural deciduous rain forest and extensive tea plantations; a hilly central catchment with teak forest, rubber plantations, and cultivated land; and a broad coastal plain bearing irrigated ricefields. The river meanders across this plain, branching near Tegallurung, where the main stream runs northwards and a major distributary, the Pancer, flows to the north-east. An 1865 map shows the Cipunegara opening through a large lobate delta, the Pancer having a smaller delta to the east, but when topographical maps were made in 1939 the Pancer had developed two large delta lobes extending 3 to 4 kilometres out into the Java Sea while the Cipunegara delta had been truncated, with shoreline recession of up to 1.5 kilometres. Aerial photographs taken in 1946 showed further advance on the Pancer delta, and continued smoothing of the former delta lobe to the west (Hollerwoger 1964). Tjia et al. (1968) confirmed this sequence with reference to the pattern of beach ridges truncated on the eastern shores of Ciasem Bay and the 1976 Landsat pictures show that a new delta has been built out to the north-east (Fig. 14). Along the coast the mangrove fringe (mainly Rhizophora) has persisted on advancing sectors but elsewhere has been eroded or displaced by the construction of fishponds.
Third in the sequence of major deltas east of Jakarta is that built by the Cimanuk River (Fig. 15). The Cimanuk and its tributaries drain a catchment of about 3,650 square kilometres, the headstreams rising on the slopes of the Priangan mountain and the Careme volcano, which carry rain forest and plantations. There has been extensive soil erosion in hilly areas of the central catchment following clearance of the forest and the introduction of grazing and cultivation, particularly in the area drained by the Cilutung tributary (Van Dijk and Vogelzang 1948). The Cimanuk thus carries massive loads of silty sediment down to the coast: of the order of 5 million tonnes a Year (Tjia et al. 1968). The broad coastal plain bears extensive rice-fields, with fishponds and some residual mangrove fringes along the shoreline to the north. The river meanders across this plain, the distributary Rambatan diverging north-westwards near Plumbon.
Hollerwoger (1964) traced changes on the delta shoreline with references to maps made in 1857, 1917, and 1935, and air photographs taken in 1946. Examination of beach-ridge patterns, marking successive shorelines, shows that before 1857 the Cimanuk took a more northerly course and built a delta lobe (Fig 16). BY 1857 this was in course of truncation, and the Cimanuk mouth had migrated westwards to initiate a new deltaic protrusion. Between 1857 and 1917 large delta lobes were built by the Cimanok and the Rambatan but an irrigation channel, the Anyar Canal, had been cut from Losarang to the coast, diminishing the flow to the Rambatan, and a new delta began to grow at the canal mouth, out into the embayment between the Cimanuk and Rambatan deltas. BY 1935 this embayment had been filled, the shoreline having advanced about 6 kilometres in 17 years, while erosion had cut back the adjacent Rambatan delta. Continued growth occurred at the mouth of the Anyar Canal and the Cimanuk between 1935 and 1946, by which time the Rambatan delta shoreline had retreated up to 300 metres.
During a major flood in 1947 the Cimanuk established a new course north-east of Indramayu, and a complex modern delta has since grown here (Plate 3). Stages in the evolution of this modern delta are shown in Fig. 17. At first there was only a single channel, but three main distributaries-the Pancer Balok, Pancer Payang, and Pancer Song-have developed as the result of levee crevassing, and each of these shows further bifurcations resulting from channel-mouth shoal formation, as well as the cutting of artificial lateral outlet channels (Tjia 1965; Hehanussa et al. 1975; Hehanussa and Hehuwat 1979). Since 1974 the Pancer Balok has replaced the Pancer Payang as the main outlet. Erosion has continued on the northern lobe where the present coastline shows an enlargement of tidal creeks, probably the result of compactionsubsidence.
On the east coast, south of Pancer Song, there has been erosion in recent decades. Sand drifting northwards has been intercepted by the oil terminal jetty at Balongan, and the shoreline north of the jetty is retreating rapidly. According to Purbohadiwidjojo (1964) Cape Ujung, to the south, was an ancient delta lobe, but there is no evidence that any channel led this way. Tjia (1965) suggested that it might be related to a buried reef structure, but there is no evidence of this either. In fact, the cuspate promontory is situated where one of the earlier beach ridges has been truncated by the present shoreline. Patterns on the 1976 Landsat picture suggest that the cape is at the point of convergence of two current systems in the adjacent sea area, but it is not clear whether the pattern is a cause or a consequence of present coastal configuration.
Although it has only a relatively small catchment (250 square kilometres) the Bangkaderes has built a substantial delta (Fig. 18) on the coast south-east of Cirebon. This is because of its large annual sediment load, derived from a hilly catchment where severe soil erosion has followed forest clearance and the introduction of farming. An 1853 map showed a small lobate delta but by 1922 two distributary lobes had been built, advancing the shoreline by up to 2.7 kilometres. Air photographs taken in 1946 show continued enlargement of the eastern branch, extended by up to 1.8 kilometres seawards, and erosion of the western branch, which no longer carried outflow (Hollerwöger 1964).
A few kilometres to the east are the Sanggarung and Bosok deltas (Fig. 19). The Sanggarung has a catchment of 940 square kilometres, and rises on the slopes of the volcanic Mt. Careme. The headwater regions are steep and forested and partly farmed land, and the coastal plain consists largely of rice-fields, with fishponds to seaward and some mangrove fringes. An 1857 survey showed a delta built out northeastwards along the Bosok distributary, and between 1857 and 1946 deposition filled in the embayment to the east, on either side of the Sebrongan estuary, and there was minor growth on the Bosok delta; to the north west the Sanggarung built out a major deltaic feature, with several distributaries leading to cuspate outlets. The coastal lowland here has thus shown continuing progradation of a confluent delta plain without the alternations that occur as the result of natural or artificial diversion of river mouths (Hollerwoger 1964).
The Pemali delta (Fig. 20) also showed consistent growth between an 1865 survey, 1920 mapping, and 1946 air photography (Hollerwöger 1964). The river drains a catchment of about 1,200 square kilometres, with forested mountainous headwater regions and extensive hilly country behind the swampy coastal plain. The delta grew more rapidly between 1920 and 1946 than it had over the 56 years preceding the 1920 survey, possibly because of accelerated soil erosion in hilly country as the result of more intensive farming.
The growth of the Comal delta to the east has shown fluctuations (Fig. 21).When it was mapped in 1870 the Comal (catchment area of about 710 square kilometres) was building a lobate delta to the northwest but by 1920 growth along a more northerly distributary had taken place. The river then developed an outlet towards the north-east, leading to the growth of a new delta in this direction by the time air photographs were taken in 1946. The earlier lobes to the west had by then been truncated. In this, as in the other north Java deltas, growth accelerated after 1920, probably as a result of increasing soil erosion due to intensification of farming within the hilly hinterland (Hollerwöger 1964).
The Bodri delta (Fig. 22) is the next in sequence. The Bodri River rises on the slope of the Prahu volcano, and drains a catchment of 640 square kilometres. Again the mountainous headwater region backs a hilly area, with a depositional coastal plain, mainly under rice cultivation. An 1864 survey shows the Bodri opening to the sea through a broad lobate delta which had grown northwards to Tanjung Korowelang at the mouths of two distributaries when it was remapped in 1910. Thereafter a new course developed, probably as the result of canal-cutting to the north east, and by 1946, when air photographs were taken, a major new delta had formed here, prograding the shoreline by up to 4.2 kilometres. Meanwhile, the earlier delta at Tanjung Korowelang had been truncated and the shoreline smoothed by erosion (Hollerwoger 1964).
East of Semarang large-scale progradation is thought to have taken place in recent centuries. Demak, a sixteenth-century coastal port, is now about 12.5 kilometres inland behind a prograded deltaic shoreline. Continuing progradation is indicated by the small delta growing at the mouth of a canal cut from the River Anyar to the sea, but otherwise the coastline north to Jepara is almost straight at the fringe of a broad depositional plain. According to Niermeyer (1913: quoted by Van Bemmelen 1949) the Muria volcano north-east of Demak was still an island in the eighteenth century, when seagoing vessels sailed through the strait that separated it from the Remang Hills, a strait now occupied by marshy alluvium. This inference, however, needs to be checked by geomorphological and stratigraphical investigations.
The shoreline of the Serang delta, south of Jepara, changed after the construction of the Wulan Canal in 1892, which diverted the sediment yield from the Kedung River to a new outlet, around which a substantial new delta has been formed. In 1911 this was of cuspate form, but by 1944 it was elongated, and by 1972 it had extended in a curved outline northwards, branching into three distributaries (Fig. 23). Between 1911 and 1944 the new delta gained 297 hectares, and from 1944 to 1972 a further 385 hectares, including beach-ridge systems and a seaward margin adapted for brackish-water fishponds.
Beyond Jepara the coast steepens on the flanks of Muria, but the shores are beach-fringed rather than cliffed. To the east the Juwana River opens on to the widening deltaic plain behind Rembang Bay, but at Awarawar the coast consists of bluffs cut in Pliocene limestone. Tuban has beaches and low dunes of quartzose sand, supplied by rivers draining sandstones in the hinterland, but otherwise the beaches on northern Java are mainly of sediments derived from volcanic or marine sources. Hilly country continues eastwards until the protrusion of the Solo River delta.
The modern Solo delta (Fig. 24) has been built out rapidly from the coast at Pangkah since a new artificial outlet from this river was cut at the beginning of the present century (Verstappen 1977). Comparisons of the outlines of the Solo delta shown on 1 :50,000 topographical maps made in 1915 and 1936 and on air photographs taken in 1943 and 1970 indicated #award growth of 3,600 metres in 1915 to 1936, a further 800 metres between 1936 and 1943, and 3,100 metres between 1943 and 1970; in real terms the delta increased by 8 square kilometres in the first period, 1 square kilometre in the second, and a further 4 square kilometres in the third (Verstappen 1964a, 1977). The rate of progradation of such a delta depends partly on the configuration of the sea floor, for as the water deepens offshore a greater volume of sediment is required to produce the same increase in surface area. It also depends on the rate of fluvial sediment yield, which has here increased following deforestation and intensified land use within the catchment, so that larger quantities of silt and clay have been derived from the intensely weathered volcanic and marry outcrops in the hinterland: the average suspended sediment load is 2.75 kilograms per cubic metre. Much of the silt has been deposited to form levees, while the finer sediment accumulates in bordering swamps.
The features of this delta include a relatively smooth eastern shoreline backed by parallel beach ridges and fronted by sand bars, the outlines determined by northeasterly wave action during the winter months. As this is also the dry season, there has been a tendency for distributaries and creeks formed on the eastern side of the Solo to be blocked off by wave deposition and silted up, the outcome being that the channels opening north-westwards have persisted to carry the bulk of the discharge and sediment yield from the Solo in the wet season, so that the delta has grown more rapidly in this direction. Mangroves (mainly Rhizophora spp.) are patchy and eroded on the eastern shore, but broad and spreading seawards between the distributary mouths on the more sheltered western shore. The tide range is small (less than 1 metre), but at low tide the mudflats exposed on the western shores are up to 200 metres wide. The rapid growth of such a long, narrow delta, protruding more than 20 kilometres seawards, is related partly to the shallowness of the adjacent sea and the consequent low-wave energy conditions and partly to the predominance of clay in the deltaic sediment, which is sufficiently cohesive to form persistent natural levees projecting out into the Java Sea.
Between 1915 and 1936 there was some lateral migration of the Solo River, marked by undercutting of banks on the outer curves of meanders, and a new outlet channel (3 in Fig. 24) was initiated, probably as the result of flood overflow and levee crevassing on the meander curve. A small delta formed here, but by 1970 it had been largely eroded leaving only a minor protuberance on an otherwise smoothly prograded eastern coast. The effects of canal construction are well illustrated where a channel, cut between 1936 and 1943 from a distributary (2 in Fig. 24) to irrigate rice-fields, increased drainage into an adjacent creek (5 in Fig 24) which then developed levees that grew out seawards. However, by 1970 this, too, had been cut back. A similar development farther south (4 in Fig. 24) converted a creek into a minor distributary of the Solo, with its own sub-delta lobe by 1943, but progradation of mangrove swamps (largely replaced by fishponds) has proceeded rapidly on this part of the western coastline, and by 1970 the distributary, although lengthened, protruded only slightly seawards. In the course of its growth, the Solo delta has incorporated the former island of Mangari, which consists of Pliocene limestone (Verstappen 1977).
East of the broad funnel-shaped entrance to Surabaya Strait the Bangkalan coast of north-west Madura Island shows several small mangrove-fringed deltas on a muddy shoreline. The north coast of the island of Madura is remarkably straight, with terraces that show intermittent emergence as the result of tectonic uplift. The hinterland is steep, with areas of Pliocene limestone, but the shore is generally beachfringed, with some minor dunes to the east. The southern coast of the island is depositional, with beaches of grey volcanic sand that culminate in a recurved spit at Padelegan. Coastal waters are muddy, but outlying islands, such as Kambing, have fringing reefs and derived beaches of pale coralline sand. Tide range increases westwards, and the Baliga River enters the sea by way of a broad, mangrovefringed tidal estuary, bordered by swampy terrain, with a narrow beach to the west.
Surabaya Strait shows tidal mudflats, scoured channels, and estuarine inlets indicative of relatively strong current action, and there has been extensive reclamation for fishponds along the mangrove-fringed coast to the south. In the fourteenth century ships could reach Mojokerto, now 50 kilometres inland on the Brantas delta, which continues to prograde around its distributary mouths. The southern shores of Madura Strait are beach-fringed, the hinterland rising steeply to the volcanoes of Bromo and Argapura. Beach sediments are grey near the mouth of rivers draining the volcanic hinterland, pale or cream near fringing coral reefs, and white in the Jangkar sector, where quartzose sands are found.
The east coast of Java is steep, with streams radiating from the Ijen volcano, but to the south a coastal plain develops and broadens. This consists of low beach ridges built mainly of volcanic materials derived from the Ringgit upland. The Sampean delta is fan-shaped, accreting on its western shores as erosion cuts back the eastern margin. The Blambangan Peninsula is of Miocene limestone, and has extensive fringing reefs backed by coralline beaches, with evidence of longshore drifting on the northern side, into the Straits of Bali.
The south coast of Java is dominated by wave action from the Indonesia Ocean, and receives a relatively gentle southwesterly swell of distant origin and stronger locally generated south-easterly waves that move shore sediments and deflect river outlets westwards, especially in the dry winter season. It is quite different from the north coast of Java, being dominated by steep and cliffed sectors and long, sandy beaches rather than protruding deltas. There is very little information on the extent of shoreline changes in historical times, and we cannot accept the statement of Tjia et al. (1968, p. 26) that abrasion rates along the south coast must have been much higher than those on the deltaic northern shoreline because of the more powerful wave action from the Indonesian Ocean: changes on this rocky and sandy coast will have been relatively slow.
The Bay of Grajagan is backed by a sandy barrier enclosing a river-fed estuarine lagoon system with an outlet to the sea at the western end, alongside the old volcanic promontory of Capil. Farther west the coast becomes indented, with cliffed headlands of Miocene sedimentary rock and irregular embayments, some with beaches and beach ridges around river mouths. Nusa barung is a large island of Miocene limestone with a karstic topography and a cliffed and isletted southern coast; its outlines are related to joint patterns and southward-tilting (Tjia 1962). It modifies oceanic wave patterns on the sandy shores of the broad embayment to the north in such a way as to generate longshore drifting from west to east so that the Bondoyudo River has been deflected several kilometres eastwards to an outlet behind a barrier spit leading to a cuspate foreland with multiple beach ridges (Fig. 25).
The coastal plain then narrows westwards and gives place to a steep indented coast on Miocene sedimentary formations, including the limestones of Kendeng, with bolder promontories of andesite near Tasikmadu. At Puger and Meleman there are beach-ridge systems surmounted by dunes up to 15 metres high, with a thick vegetation cover, in sequence parallel to the shoreline. These interrupt the predominantly karstic limestone coast (Plates 4, 5, and 6), with cliffed sectors and some fringing reefs, that continues westwards to Parangtritis. Near Baron the limestone cliffs are fronted by shore platforms exposed at low tide and flat-floored notches, cut in the base of cliffs and stacks, testify to the importance of solution processes in the shaping of these features (Plate 4). Locally, beaches of calcareous sand and gravel occupy coves, and where these occur an abrasion ramp may be seen at the rear of the shore platform. At Baron a river issues from the base of a cliff and meanders across a beach of black sand that has evidently been washed into the valley-mouth inlet by ocean waves (Plate 5), the sand having come from sea-floor deposits supplied by other rivers draining the volcanic hinterland.
At Parangtritis the cliffs end, and the broad depositional plain of central south Java begins. The Opak and Progo rivers, draining the southern slopes of the Merapi volcano, are heavily laden with grey sands and gravel derived from pyroclastic materials. During floods these are carried into the sea to be reworked by wave action and built into beaches with a westward drift (Plates 7 and 8). The coastal plain has prograded, with the formation of several beach ridges separated by swampy swales. No measurements of historical changes are available, but our reconnaissance in November 1979 found evidence of sequences of localized progradation at the river mouths followed by westward distribution of part of the prograded material. It appears that the alignment of the shore is being maintained, or even advanced seawards, as the result of successive increments of fluvial sand supply. Finer sediment, silt and clay, is deposited in bordering marshes and swales, or carried into the sea and dispersed by strong wave action.
On some sectors, especially near Parangtritis, the beach is backed by dune topography, typically in the form of ridges parallel to the shoreline and bearing a sparse scrub cover (Plate 9). At Parangtritis there are mobile dunes up to 30 metres high, driven inland by the south-easterly winds (Plate 10). The presence of mobile dunes, unusual in this humid tropical environment, may be due to a reduction of their former vegetation cover by sheep and goat grazing, and by the harvesting of firewood (Verstappen 1957).
Whereas the present beach and dune systems consist of incoherent grey sand, readily mobilized by wind action in unvegetated areas, the older beach-ridge systems farther inland are of more coherent silty sand which can be used for dry-land cultivation. The silt fraction may be derived from airborne (e.g., volcanic dust) or flood-borne accessions of fine sediment, or it may be the outcome of in situ weathering of some of the minerals in the originally incoherent sand deposits.
At Karangtawang the depositional lowland is interrupted by a high rocky promontory of andesite and limestone, the Karangboto Peninsula. There are extensive sand shoals off the estuary of the Centang River, which washes the margins of the rocky upland, and there appears to have been rapid progradation of the beach to the east-and also in the bay to the west-where sand has built up in front of a former sea cave which used to be accessible only by means of ropes and ladders when men descended the cliff to collect birds' nests. Rapid accretion may have been stimulated here by the catastrophic discharge of water and sediment that followed the collapse of the Sempor Dam in the hinterland in 1966.
The sandy and swampy coastal plain resumes to the west of the Karangboto Peninsula, and extends past the mouth of the Serayu River. In this sector it has been disturbed by the extraction of magnetite and titanium oxide sands; in places, the beach ridges have been changed into irregular drifting dunes, while dredged areas persist as shallow lagoons.
On either side of the mouth of the Serayu River the coastal plain has prograded by the addition of successive sandy beach ridges separated by marshy swales. The sediments are of fluvial origin, reworked and emplaced by wave action, and progradation has enclosed a former island as a sandstone hill among the beach ridges. According to Zuidam et al. (1977) the coastal plain shows a landward slope at a number of places where the streamlets flow land-wards instead of seawards, and this is presumed to be due to very recent differential tectonic movements.
The geomorphological contrast between the irregular deltaic coast of northern Java and the smooth outlines of depositional sectors on the south Java coast is largely due to contrasts in wave-energy regimes and sea-floor topography. The sediment loads of rivers flowing northwards and southwards from the mountainous watershed are similar, but the finer silt and clay, deposited to form deltas in the low-energy environments of the north coast, are dispersed by high-wave energy on the south coast. The coarser sand fraction seen in beach ridges associated with the north coast deltas is thus concentrated in more substantial beach and dune formations on the south coast. The contrast is emphasized by the shallowness of coastal waters off the north coast, which reduces wave energy, as opposed to the more steeply shelving sea floor off the south coast, which allows larger waves to move into the shoreline. Nevertheless, silt and clay carried in floodwaters settles in the swales between successively built beach-ridge systems along the southern coast, and in such embayments as Segara Anakan, and, as we have noted, it may also have been added to the sandy deposits of older beach ridges inland.
The nature and rate of sediment yield from rivers draining to the south coast vary with the size and steepness of the catchment, with geological features such as catchment Ethology, and with vegetation cover. In the Serayu River basin, deforestation has accelerated sediment yield and increased the incidence of flooding in recent years. Meijerink (1977) found that the annual sediment yield from catchments dominated by sedimentary rocks was ten times that of catchments with similar vegetation and land use on volcanic formations, the contrast being reflected in the nature and scale of depositional features developed at the river mouth.
West of the Serayu River the sandy shoreline, backed by beach ridges, curves southwards to Cilacap, in the lee of Nusakambangan, a high ridge of limestone and conglomerate, with precipitous cliffs along its southern coastline. Extensive mangrove swamps threaded by channels and tidal creeks border the shallow estuarine embayment of Segara Anakan (Fig. 26), which receives large quantities of silty sediment from the Citanduy River. At the eastern end, strong tidal currents maintain a navigable inlet for the port of Cilacap, which stands on the sandy barrier behind a shoaly bay. A meandering channel persists westwards, leading through the mangroves to Segara Anakan, which has a larger outlet through a steep-sided strait to Penandjung Bay. Changes in the configuration of Segara Anakan between 1900 and 1964 were traced by Hadisumarno (1964), who found evidence for rapid advance of mangroves into the accreting intertidal zone. He reported surveys made in 1924, when the average depth (ignoring deeper tidal channels) was 0.5 to 0.6 metres, and 1961, when it had shallowed to 0.1 to 0.2 metres, the tidal channels having deepened. Mangrove advance is exceptionally rapid here, and much of the shallow lagoon is expected to disappear as mangroves encroach further in the next two decades. The features and dynamics of Segara Ankan are being studied in Phase II of the UN University LIPI Indonesian coastal resources management project in 1980-81.
West of Segara Anakan the beach ridge plain curves out to the tombolo of Pangandaran, where deposition has tied an island of Miocene limestone (Panenjoan) to the Java mainland (Fig. 27), and continues on to Cijulang, where the hinterland again becomes hilly. Beaches line the shore, and many of the rivers have deflected and sand-barred mouths. At Genteng a beach-ridge plain develops, curving out to a tombolo that attaches a former coralline island, and beach ridges also thread the depositional lowlands around the mouths of the Ciletuh and Cimandiri rivers flowing into Pelabuhanratu Bay. The beach ridges indicate past progradation, but no information is available on historical trends of shoreline change in this region. West of Pelabuhanratu the coast steepens, but is still fringed by surf beaches, some sectors widening into depositional coastal plains with beach and dune ridges and swampy swales, including the isthmus which ties Ujong Kulon as a peninsula culminating in Java Head
The Indonesian coasts of Kalimantan have received very little attention from geomorphologists, and there is no information on rates of shoreline change in historical times. The western and southern coasts are extensively swampy, with mangroves along the fringes of estuaries, inlets and sheltered embayments. The hilly hinterland approaches the west coast north of Pontianak, where there are broad tidal inlets, and to the south depositional progradation has attached a number of former volcanic islands as headlands. The Pawan and the Kapuas rivers have both brought down sufficient sediment to build substantial deltas (Tjia 1963) but in general the shoreline consists of narrow intermittent sandy beaches backed by swamps, with cuspate salients in the lee of islands such as Gelam, or reefs as at Tanjung Sambar. South of Kendawagan a ridge of Triassic rocks runs out to form the steep-sided Betujurung promontory and the hills on Bawat and Gelam islands.
The south coast is similar, with a number of cuspate and lobate salients, most of which are swampy protrusions rather than deltas. Sand of fluvial origin has drifted along the shoreline east and west from the mouth of the Siamok, to form the straight spit of Tanjung Bandaran to the east, partly enclosing mangrove-fringed Sampit Bay, and the recurved spit of Tanjung Puting to the west. Near Banjarmarsin, ridges of Cretaceous and Mio-Pliocene rock run through to form the promontory of Selatan where the swampy shores give place to the more hilly coastal country of eastern Kalimantan.
The east coast has many inlets and swamp-fringed embayments, the chief contrast being the large Mahakam delta, formed downstream from Samarinda (Fig. 28). Coarse sandy sediment derived mainly from ridges and valleys in the Samarinda area is prominent in the delta, which has numerous distributaries branching among the swampy islands (Magnier et al. 1975, Allen et al. 1976). Other rivers draining to the east coast open into funnel-shaped tidal estuaries, as at Balikpapan and Sangkulirang, and Berau, Kajau, and Sesayap in the north-east (Tjia 1963); as has been noted, tide ranges are higher on the east coast of Kalimantan than on the south and west coasts. At Balikpapan, shoreline erosion has resulted from the quarrying of a fringing coral reef, but the rate and extent of this erosion have not been documented.
The coasts of Sulawesi have also received little attention from geomorphologists but it is known that this island has been tectonically active. In contrast with the low-lying swampy shores of Kalimantan there are long sectors of steep coast, often with terraced features indicating tectonic uplift or tilting, especially where coral reefs have been raised to various levels up to 600 metres above present sea level, some of them transversely warped and faulted. Rivers are short and steep, with many waterfalls and incised gorges, and there are minor depositional plains around the river mouths. Fringing and nearshore coral reefs are extensive, and along the shore there are sectors of beach sand, with spits and cuspate forelands, especially in the lee of offshote islands, as at Bentenan. It is likely that progradation is taking place where rivers drain into the heads of inlets and embayments, especially on the east coast, where mangrove fringes are extensive, but no details are available. Volcanic activity has modified coastal features locally, for example on Menado-tua -the active volcano off Menado in the far north of the island- and erosion has been reported at Bahu, but again there are no detailed studies. South and south-east of Sulawesi there are many uplifted reef patches and atolls, as well as islands fringed by raised reef terraces Binongko, for example, has a stairway of 14 reef terraces, the highest 200 metres above sea level (Kuenen 1933), and Muna is a westward-tilted island with reef terraces up to 445 metres above sea level (Verstappen 19601.
Bali and Nusatenggara
The northwestern coast of Bali consists of Pliocene limestone terrain, the shores having yellow beach sands and some fringing reefs. A lowland behind Gilimanuk becomes a narrowing coastal plain along the northern shore, giving place to a steeper coast on volcanic rocks near Singaraja. Out to the north, the Kangean islands include uplifted reefs and emerged atolls.
In the eastern part of Bali the coast is influenced by the active volcanoes, specially Agung, which generate lava and ash deposits that move downslope and provide a source of sediment that is washed into the sea by rivers, particularly during the wet season (December to April). These sediments are then distributed by wave action to be incorporated in grey beaches. Sanur beach is a mixture of fluvially supplied grey volcanic sand and coralline sand derived from the fringing reef (Tsuchiya 1975,1978). At Sengkidu the destruction of a fringing reef by collecting and quarrying of coral has led to erosion of the beach to the rear, so that ruins of a temple now stand about 100 metres offshore, indicating that there has been shoreline erosion of at least this amount in the past few decades following the loss of the protective reef (Praseno and Sukarno 1977). Similar erosion is in progress on Kuta and Sanur beaches.
South of Sanur, in the lee of the broad sandy isthmus that links mainland Bali to the Bukit Peninsula (of Miocene limestone) to the south, spits partly enclose a broad tidal embayment with patches of mangrove on extensive mudflats. This peninsula has a cliffed coast with caves and notches, stacks rising from basal rock ledges and extensive fringing reefs; beaches occupy coves and south of Benoa beach deposition has resulted in the attachment of a small island to the coast as a tombolo (Plate 11).
West of the isthmus, ocean waves determine the curvature of beach outlines, and there has been erosion in recent decades on either side of the protruding airport runway at Denpasar. The beach here, in the lee of a fringing coral reef, is of pale coralline sand, backed by low dunes. It gives way northwards to grey sand of volcanic origin, with beaches interrupted by low rocky promontories and shore benches. Longshore drifting to the north-west is indicated by spits that deflect stream mouths in that direction (Plate 12), and as wave energy decreases, in the lee of the Semenanjung promontory of south-eastern Java, the beaches become narrower and gentler in transverse gradient.
At the north-western end of Bali the Gilimanuk spit shows several stages of growth from the coast at Cejik, to the south, interspersed with episodes of truncation. Verstappen (1975b) suggested that growth occurred during phases of dominance of westerly wave action and truncation when south-easterly waves were prevalent, the variation being due to wind regimes associated with long-term migrations of the ITC, but stages in the evolution of this spit have not Yet been dated.
Many of the features found on Bali are also found on the similar Lesser Sunda islands to the east, but few details are available. Cliffs of limestone and volcanic rock extend along the southern coasts of Lombok, Sumbawa, and Sumba but elsewhere the coasts are typically steep rather than cliffed and often have fringing coral reefs. There are many volcanoes, some of them active: Inerie and Iya in southern Flores and Lewotori to the east have all erupted in recent times and deposited lava and ash on the coast, as has Gamkonora on Halmahera. Rivers have only small catchments, and depositional lowlands are confined to sheltered embayments, mainly on the northern shores. Terraces and emerged reefs indicative of uplift and tilting are frequently encountered on these eastern islands (Davis 1928). On Sumbawa uplifted coral reefs are up to 700 metres above sea level, attached to the dissected slopes of old volcanoes and, on Timor, reef terraces-much dissected by stream incision-attain 1,200 metres above sea level, the higher ones encircling mountain peaks that were once islands with fringing reefs or almost-atolls with reefs enclosing lagoons that had a central island.
Chappell and Veeh (1978) have examined raised coral terraces on the north coast of Timor and on the south coast of the adjacent volcanic island of Atauro, where they extend more than 600 metres above sea level. Dating by Th230-U234 established a sequence of shoreline features and fringing reefs developed during Quaternary oscillations of sea level on steadily rising land margins. On Atauro, where the stratigraphy is very well displayed in gorge sections cut through the terrace stairway, the shoreline of 120,000 Years BP is 63 metres above present sea level. Correlation with other such studies, notably in Barbados, New Guinea, and Hawaii, suggests that the world ocean level was then only 5 to 8 metres above the present, which indicates a mean uplift rate of about 0.5 metres per 1,000 years in Atauro. At Manatuto, Baucau, and Lautem in north-east Timor, dating of similar terraces indicates a similar uplift rate but at Hau, just east of D;l;, the shoreline of 120,000 years BP is only 10 metres above sea level indicating a much slower rate of land uplift, only 2 to 4 centimetres per 1,000 years.
Another emerged almost-atoll is seen on Rot;, southwest of Timor, where the enclosing reefs have been raised up to 200 metres, the highest encircling interior hills of strongly folded sedimentary rocks. Kissar, north-east of Timor, has a stairway of five reef terraces, the highest at 150 metres above sea level. Leti, to the east of Timor, has been uplifted in two stages to form reef terraces 10 metres and 130 metres above sea level, and similar features are seen at 10 to 20 metres and 200 to 240 metres on the nearby island of Moa. Yamdena is an island bordered by high cliffs of coral limestone cut into the outer margins of a reef that has been raised 30 metres out of the sea.
North of the Banda Sea, Seram has a coral reef 100 metres above sea level, and Ambon, which consists of two islands linked by a sandy isthmus, has reefs at heights of up to 530 metres. Gorong, south-east of Seram, is an atoll uplifted in several stages to 300 metres, and now encircled by a lagoon and a modern atoll reef Obi and Halmahera also have upraised reef terraces up to 300 metres above sea level. In the Aru Islands, Verstappen 11960) described cliffs fronted by shore platforms that had been submerged as the result of tectonic subsidence, but uplifted atolls also occur in this region.
A great deal of research is required to establish the nature of coastal features in eastern Indonesia. Some of the reconnaissance accounts are misleading: cliffs have been taken as evidence of recent uplift, and mangrove-fringed embayments as indications of recent subsidence; and it is possible that too much emphasis has been given to catastrophic events, such as earthquakes, volcanic eruptions, and tsunami, in the interpretation of coastal features.
Tectonic movements have undoubtedly influenced coastal changes in parts of Irian Jaya, both of steep sectors, mainly in the north, and in the extensive swampy lowlands to the south. Verstappen (1964a) compared a 1903 map of Frederik Hendrik island (Yos Sudarso), near the mouth of the Digul River on the south coast, with maps based on air photographs taken in 1945, and found evidence of substantial progradation, which he attributed to recent uplift in a zone passing through Cape Valsch (Fig. 29). Frederik Hendrik Island is mainly low-lying, with extensive reed-swamps, and its bordering channels are scoured by strong tidal currents but the Digul River opens into a broad estuary, and under present conditions it relinquishes most of its sediment load upstream as it traverses extensive swamps and recently subsided areas between Gantentiri and Yondom. In consequence it is not now building a delta into the Arafura Sea.
On the north coast of Irian Jaya the Mamberamo has built a substantial delta, but in recent decades this has shown little growth; indeed, the western shores show creek enlargement and landward migration of mangroves, while the eastern flank is fringed by partly submerged beach ridges with dead trees, all indicative of subsidence (probably due to compaction) and diminished sediment yield from the river. Verstappen (1964a) related this diminished yield to an intercepting zone of tectonic subsidence that runs across the southern part of the delta, marked by a chain of lakes and swamps, including an anomalous mangrove area (Fig. 30). The largest of the lakes, Rombabai Lake, is adjacent to the levees of the Mamberamo, and at one point the subsided levee has been breached during floods and a small marginal delta has grown out into the lake.
The islands west of Irian Jaya show evidence of tectonic movements, Waigeo being bordered by notched cliffs of recently uplifted reef limestone, while Kafiau is essentially an upraised almost-atoll with hills of coral limestone ringing an interior upland.
In September 1979 a major earthquake (force 8 on the Richter scale) disturbed the islands of Yapen and Biak, north of Irian Jaya, initiating massive landslips on steep coastal slopes, especially near Ansus on the south coast of Yapen. According to the United States Geological Survey it was the strongest earthquake in Indonesia since the August 1977 tremor on Sumbawa which had similar effects. Tsunami generated by these and other earthquakes were transmitted through eastern Indonesia but there have been no detailed studies of their geomorphological and ecological consequences.
This review of Indonesian coastal features has indicated the variety of forms that exist within this archipelago, the bestdocumented sectors being the north-eastern coast of Sumatra and the north coast of Java, both of which show evidence of substantial changes within historical times. It is hoped that geomorphological studies will soon provide much more information on the other sectors, which at this stage are poorly documented and little understood.