This paper presents an internally and globally consistent model of plate evolution in eastern Indonesia from Middle Miocene to Present time. It is centered on the Banda Sea region located in the triple junction area between the Pacific–Philippine, Australia and South–East Asia plates. The geological and geophysical data available from Indonesia were until recently insufficient to define a unique plate tectonic model. In this paper, the new data taken into account clearly restrict the possible interpretations. Owing to a great number of geological, geophysical and geochemical studies, the major plate boundaries (the Sunda–Banda subduction zone to the south, the Tarera–Aiduna Fault zone and the Seram Thrust to the east, and the Sorong Fault zone and Molucca Sea collision zone to the north) are now clearly identified. The age of the major tectonic structures is also better known. Geodetic measurements well constrain the Present time plate kinematics. We also consider the deformation history within eastern Indonesia, where numerous short-lived microplates and their related microcontinents successively accreted to the Asiatic margin. Moreover, magnetic anomalies identification of the North and South Banda Sea basins allows a precise kinematic reconstruction of the back-arc opening. We used the Plates software to test the coherency of our model, presented as a series of 4 plate reconstruction maps from 13 Ma to the present. Finally, the origin of oceanic domains restored by our reconstruction is discussed.

Southeastern Indonesia is located at a convergent triple junction of 3 plates : the Pacific (including the Caroline and Philippines plates), the Australian and the Southeast Asian plates (fig. 1). The age of the different basins : the North Banda Sea (Sula Basin), the South Banda Sea (Wetar and Damar Basins) and the Weber Trough has been debated for a long time, Their great depth was a reason to interpret them as remnants of oceanic domains either of Indian or Pacific ocean affinities. It has now been demonstrated from geochronological studies that these basins have formed during the Neogene [Rehault et al., 1994; Honthaas et al., 1998]. The crust has been sampled only in the Sula Basin, where basalts or trachyandesites with back-arc geochemical signatures have been dredged. Their ages range from 11.4 +/- 1.15 to 7.33 +/- 0.18 Ma [Rehault et al., 1994; Honthaas et al., 1998]. The study of the magnetic anomaly pattern of these basins confirms this interpretation and defines an age between 12.5 and 7.15 Ma for the North Banda Basin and between 6.5 to 3.5 Ma for the South Banda Basin [Hinschberger et al., 2000; Hinschberger et al., 2001]. Furthermore, the existence of volcanic arcs linked to subducted slabs suggests that these basins resulted from back-arc spreading and subduction slab roll-back. Lastly, the Weber Trough which exceeds 7 300 rn in depth and is one of the deepest non subduction basins in the world, remains enigmatic.A compilation of existing bathymetric data allows us to present a new bathymetric map of the region (fig. 2 and 3). A comparison with the previous published maps [Mammerickx et al., 1976; Bowin et al., 1982] shows numerous differences at a local scale. This is especially true for the Banda Ridges or in the Sula Basin where new tectonic directions are expressed.In the North Banda Basin, the Tampomas Ridge, which was striking NE-SW in the previous maps, is actually NW-SE parallel to the West Buru Fracture Zone and to the Hamilton Fault scarp (fig. 6). This NW-SE direction represents the initial direction of rifting and oceanic spreading. In this basin, only the southeastern rifted margin morphology is preserved along the Sinta Ridges. The basin is presently involved in an overall compressional motion and its buckled and fractured crust is subducted westwards beneath East Sulawesi (fig. 4a, 5 and 6). The northern border of the North Banda Basin is reactivated into sinistral transcurrent motion in the South Sula Fracture Zone continued into the Matano fault in Sulawesi.The South Banda Sea Basin is divided in two parts, the Wetar and Damar Basins with an eastward increase in depth. The Wetar and Damar Basins are separated by the NNW-SSE Gunung Api Ridge, characterized by volcanoes, a deep pull apart basin and active tectonics on its eastern flank (fig. 4b and 7). This ridge is interpreted as a large sinistral strike-slip fracture zone which continues across the Banda Ridges and bends towards NW south of Sinta Ridge.The Banda Ridges region, separating the North Banda Basin from the southern Banda Sea (fig. 5 and 7), is another place where many new morphological features are now documented. The Sinta Ridge to the north is separated from Buru island by the South Buru Basin which may constitute together with the West Buru Fracture Zone a large transcurrent lineament striking NW-SE. The central Rama Ridge is made of 2 narrow ridges striking NE-SW with an << en-echelon >> pattern indicating sinistral strike slip comparable to the ENE-WSW strike-slip faulting evidenced by focal mechanisms in the northern border of the Damar Basin [Hinschberger, 2000]. Dredging of Triassic platform rocks and metamorphic basement on the Sinta and Rama Ridges suggests that they are fragments of a continental block [Silver et al., 1985; Villeneuve et al., 1994; Cornee et al., 1998]. The Banda Ridges are fringed to the south by a volcanic arc well expressed in the morphology : the Nieuwerkerk-Emperor of China and the Lucipara volcanic chains whose andesites and arc basalts have been dated between 8 and 3.45 Ma [Honthaas et al., 1998].Eastern Indonesia deep oceanic basins are linked to the existence of 2 different subduction zones expressed by 2 different downgoing slabs and 2 volcanic arcs : the Banda arc and the Seram are [Cardwell et Isacks, 1978; Milsom, 2001]. They correspond respectively to the termination of the Australian subduction and to the Bird's head (Irian Jaya) subduction under Scram (fig. 5). Our bathymetric study helps to define the Seram volcanic arc which follows a trend parallel to the Seram Trench from Ambelau island southeast of Burn to the Banda Island (fig. 2 and 5). A new volcanic seamount discovered in the southeast of Buru (location of dredge 401 in figure 7) and a large volcano in the Pisang Ridge (location of dredge 403 in figure 7 and figure 8) have been surveyed with swath bathymetry. Both show a sub-aerial volcanic morphology and a further subsidence evidenced by the dredging of reefal limestones sampled at about 3000 in depth on their flank.We compare the mean basement depths corrected for sediment loading for the different basins (fig. 9). These depths are about 5 000 in in the Sula Basin, 4 800 in in the Wetar basin and 5 100 in in the Damar basin. These values plot about 1 000 in below the age-depth curve for the back-arc basins [Park et al., 1990] and about 2000 in below the Parsons and Sclater's curve for the oceanic crust [Parsons et Sclater, 1977]. More generally, eastern Indonesia is characterized by large vertical motions. Strong subsidence is observed in the deep basins and in the Banda Ridges. On the contrary, large uplifts characterize the islands with rates ranging between 20 to 250 cm/kyr [De Strict et al., 1989a]. Excess subsidence in the back-arc basins has been attributed to large lateral heat loss due to their small size [Boemer et Sclater, 19891 or to the presence of cold subducting slabs. In eastern Indonesia, these mechanisms can explain only a part of the observed subsidence. It is likely that we have to take into account the tectonic forces linked to plate convergence. This is supported by the fact that uplift motions are clearly located in the area of active collision.In conclusion, the bathymetry and morphology of eastern Indonesian basins reveal a tectonically very active region where basins opened successively in back-arc, intra-arc and fore-arc situation in a continuous convergent geodynamic setting.

The South Banda Basin is located within eastern Indonesia near the triple junction between the Eurasian, Pacific and Indo-Australian plates. It is underlain by oceanic crust, but its origin and age were not well established. It has been interpreted as a Mesozoic trapped piece of Indian ocean or as a Cretaceous-Eocene basin related to the Celebes and Sulu basins, but a Neogene back-are origin was also considered. Recent geochemical and geochronological studies strongly support the latter hypothesis.In this paper we present a new analysis of the magnetic data from the eastern part of the South Banda Basin, the Damar Basin. We used magnetic field measurements collected during nine oceanographic cruises from various institutions. Looking for magnetic correlation in the time span given from recent geochronological data, the comparison between measured profiles and theoretical profiles deduced from the reversals of Earth's magnetic field during Neogene time allows us to infer an opening of South Bands Basin during Late Miocene-Early Pliocene time, from 6.5 to 3.5 Ma. Magnetic lineations 2An, 2Ar, 3n, 3r, 3A, and possibly 3Ar are recognized, with an extinct spreading centre trending ENE-WSW. At least five segments are identified, each segment being separated by inactive transform faults perpendicular to the extinct spreading centre. A half spreading rate of about 3 cm/yr is calculated, based on spacing of magnetic lineations.The opening history of the basin is discussed. The cause of cessation of spreading is likely the are-continent collision dated at about 3 Ma. However the onset of opening is less determined. We suggest that Damar Basin began to open at about 6.5 Ma during magnetic period 3An as an intra-are basin, separating the Banda volcanic are to the south from the incipient Lucipara volcanic are to the north. The latter was probably created at the beginning of rifting of Damar Basin, as shown by both magnetic and geochronological data.The young age of the South Banda Sea Basin is contradictory with its great depth. We discuss this problem in view of thermal and tectonic considerations. We finally conclude that our magnetic model allows us to more precisely describe the opening of the basin. (C) 2001 Elsevier Science B.V. All rights reserved.

The North Banda Sea Basin is located in Eastern Indonesia. close to the triple junction between the Eurasian, Pacific and Indo-Australian plates, and opened during Late Miocene time in a back are setting. We use the magnetic and bathymetric data to depict this opening and the geodynamical evolution of the basin. We also take into account radiochronological datations available from some dredges of its basement. Sea floor spreading occurred from 12.5 to 7.15 Ma directed by three large NW-SE transform faults, namely the West Buru, Tampomas and Hamilton fracture zones. Finally, a schematic model of the North and South Banda basins evolution is presented. (C) 2000 Academie des sciences/ Editions scientifiques et medicales Elsevier SAS.

Dredgings conducted during the French–Indonesian cruises Banda Sea II and III collected volcanic rocks from several ridges of the Banda Sea area (Tukang Besi ridge, site 218; Lucipara ridge, sites 214 and 305; Nieuwerkerk–Emperor of China, sites 219 and 220). With the exception of one 46-Ma-old N-MORB type basalt, thought to belong to an ophiolitic complex, K–Ar and Ar–Ar datings indicate that all the dredged volcanics are Neogene. They range in age from ca. 10 Ma (Tukang Besi back-arc basalts) to 8–7 Ma (Nieuwerkerk–Emperor of China calc-alkaline andesites) and to 7–3 Ma (Lucipara OIB-type transitional basalts and cordierite-bearing andesites). Radiogenic isotopic signatures of andesites are consistent with an AFC (Assimilation coupled with Fractional Crystallization) process involving assimilation of continental crust. 8–3-Ma-old calc-alkaline volcanic activity is also recorded on the Wetar segment, an inactive part of the East Sunda arc, and corresponding isotopic compositions are also consistent with an AFC process involving continental crust. These features suggest that Lucipara–Nieuwerkerk–Emperor of China ridges and the Wetar segment were representing a single volcanic arc 8–7 Ma ago. The corresponding calc-alkaline activity was related to the subduction of the Indian oceanic lithosphere beneath continental blocks of Australian origin. Back-arc opening processes occurred from 6 to 3 Ma as a multi-rift opening for the Wetar basin and as a single-rift opening for the Damar basin while subduction-related magmatism was still active in the Wetar segment. Volcanic activity stopped at 7 Ma in the Nieuwerkerk–Emperor of China ridge. On Lucipara ridge, 6–3 Ma volcanic activity emplaced concomitantly transitional basalts and cordierite-bearing andesites. The mineralogical and chemical features of the latter are consistent with an AFC process involving assimilation of continental crust by mantle-derived basaltic magmas. The end of magmatic activity on both volcanic segments at 3 Ma is thought to result from the collision of Timor with the Wetar segment of the Sunda arc.

New biostratigraphical, geochronological and geochemical data from Kur island anti two near by dredgings allow us to show unknown events on the eastern margin of the Weber basin: an Early Oligocene magmatic arc; a tectono-metamorphic event between 24 and 17 Ma; and an Early Pliocene deformation, related to the collision between the Australian plate and the Banda arc. The Weber basin was created during Pleistocene with an uplift of its eastern margin. These events are briefly replaced within the geodynamic context of the Eastern Indonesia.

A fairly complete sedimentary succession (3000–4000 m thick) off the Australian margin east of Kai Besar island (Arafura Sea) is documented from micropalaeontological analysis of samples from eight dredges collected during the Banda Sea II cruise on board Baruna Jaya III. Sediments indicate shelf environment, with siliciclastic probably Late Carboniferous-Early Permian sediments, Jurassic-Cretaceous black shales, open-marine pelagic sediments from Late Albian to Mid-Late Eocene age and shallow water platform carbonate deposited during Early Miocene. From Late Miocene to Holocene times pelagic sediments were deposited. Because samples were dredged, stratigraphic gaps in the succession can only be inferred, but the general sedimentary sequence is similar to that of surrounding areas in the Arafura Sea, as known from exploratory wells.

Pillow-lavas have been dredged from the deep basement of the North Banda Sea, during the French-Indonesian Geobandut cruise. They are transitional basalts and trachyandesites displaying negative niobium and tantalum anomalies similar to those of lavas from back-arc basins. Their K-40-Ar-40 whole rock isotopic ages are 9 and 6.9 Ma respectively. Despite a large subsidence anomaly, the deep ''oceanic'' basement engaged in subduction beneath Sulawesi Island is thus not a trapped piece of Indian Ocean but an Upper Miocene back-arc basin floor.

Some quantities, like boundary layer parameters, turbulence intensity, and skewness and flatness of u and ?u/?t, were measured to specify the flow in which the effect of the wall proximity on hot-wire readings was studied and corrected. The relationship between skewness and flatness of ?u/?t established by Van Atta and Antonia (1980) is also valid in the flows investigated here.