Sites 243 and 244

Abstract

Long before, and since, development of the sea floor spreading hypothesis and of modern plate tectonic concepts, a variety of predrift paleopositions and subsequent movements has been proposed for Madagascar. Most of the authors concerned have displayed their unfamiliarity with (or have found it convenient to disregard) the significance of both the similarities and differences between the structure and geological history of eastern Africa and western Madagascar, which have been summarized by Dixey (1960) and Flores (1970). An important factor contributing to the confusion is the almost complete lack of published data on the bathymetry, crustal structure, and stratigraphy of the Mozambique Channel between Mozambique and Madagascar. Flores (1970 has published a concise account of the Phanerozoic geology of Mozambique and western Madagascar based upon his personal familiarity with both areas and access to data from deep boreholes. He summarizes the overall similarities of Karroo and Cretaceous stratigraphy in both areas but is careful to emphasize the time and space distribution of continental/marine facies transitions. In both southeastern Africa and western Madagascar, continental lacustrine, with perhaps occasional marine conditions of sedimentation was prevalent during the Permo-Carboniferous and Triassic periods, giving way during the Jurassic to widespread volcanism in the west with contemporaneous marine sedimentation in the east (Madagascar), south (Agulhas Bank), and coastal Kenya (Dixey, 1960). According to Flores (1970), during the Cretaceous the marine facies of western Madagascar changed to continental, culminating in volcanic activity during the middle Cretaceous (Turonian to Santonian/ Campanian). By contrast, the present wide Mozambique coastal plain was an area of marine sedimentation throughout the Cretaceous, becoming continental in the west. Upper Cretaceous and Tertiary marine sediments are found on both sides of the Mozambique Channel, with Late Cretaceous intrusives and late Tertiary to Quaternary volcanic activity chiefly in Madagascar and in the Mozambique Channel (Figure 1). The general morphological features of the Mozambique Channel are shown on Figure 1, which is partly based upon an unpublished compilation of bathymetric data by Langseth, Heezen, and Ewing. To the south lies the 4500-5000-meter-deep abyssal plain in the Mozambique Basin, which is bounded on the west and east, respectively, by the Mozambique and Madagascar ridges and becomes progressively shallower northward into the Mozambique Channel. The adjacent continental slopes of Africa and Madagascar are steeply inclined down to the average depth of about 3000 meters in the Channel. The floor of the channel is deeply incised by the Zambesi Canyon and several tributary canyons from Madagascar which effectively act as channels for the transport and dispersion of terrigenous sediment into the northern Mozambique Basin. The Mozambique Channel is divided into two distinct basin compartments by the asymmetric (steep to the west) and weakly seismic Davie Ridge. This ridge follows the 42°E meridian southward until it disappears as a positive morphological feature near 19°S, where its trend is continued southward by the lower Zambesi Canyon. Both features, each in its own way, act as effective barriers to the transport of terrigenous sediment across the channel. West of the Zambesi Canyon and the volcanic islands of Europa and Bassas da India lies the Europa Basin and abyssal plain (3500 meters), which is underlain by stratified sediments more than 1 sec DT (double way time) thick. Between the Davie Ridge and northwestern Madagascar and extending northward to the volcanic Comoro Islands (which mark the southern boundary of the Somali Basin) lies the Comoro Abyssal Plain (3500 meters), which is similarly underlain by stratified sediments more than 1 sec DT thick. A series of airborne magnetic profiles flown east-west across the Mozambique Channel and northern Mozambique Basin has been published by Green (1972). Except over volcanic islands, the magnetic anomalies over the channel between 10°S and 22° S (Europa Island) are below IOO7 with wavelengths between 15 and 60 km. These contrast with the much greater amplitude and reduced wavelength of the anomalies over the Mozambique Basin between 25°S and 30° S. Reconstructions of Gondwanaland have variously placed Madagascar in three possible positions at the end of the Paleozoic before the commencement of drift: 1) Adjacent to the coast of East Africa, off Somalia, Kenya and Tanzania, followed by drift to the south and slightly east. This reconstruction is favored by du Toit (1937); Fisher, Engel, and Hilde, (1968) (who deduce that it was 20° north of its present position during the Permian and that is has occupied its present position relative to Africa since the Cretaceous); Dietz and Holden (1970); McElhinny (1970); Smith and Hallam (1970); Heirtzler and Burroughs (1971) (who believe that the southward movement began since the early Cenozoic, 42-65 m.y. ago, and still continues at a rate of 2.9 cm/yr); and Sowerbutts (1972) (who suggests the initiation of breakup during the Cretaceous). 2) Adjacent to the pre-Cretaceous continental margin of Mozambique or Natal, with subsequent drift to the north and east. This paleoposition is advocated by Wellington (1954, 1955); Flores (1970) (commencement of drift during Early Jurassic, cessation during mid-Cretaceous- Turonian); Wright and McCurry (1970); Heirtzler (1971) (movement sometime since the Cretaceous); and Green (1972) (who believes that Madagascar separated from the Natal margin by east-west spreading about the Mozambique Ridge from the Late Triassic until Late Cretaceous/early Tertiary and that possible continued northward movement relative to Africa is indicated by present-day seismic activity of the Davie Ridge). 3) Present position relative to Africa has been maintained since the Paleozoic at least. This reflects the conclusion of Dixey (1956, 1960) (who proposed late Carboniferous subsidence of the Mozambique Channel area resulting in the formation of a "geosyncline" and deposition of the Karroo sequence with up to 14 km of sedimentary rocks); Pepper and Everhart (1963); Holmes (1965); Flower and Strong (1969); and Tarling (1971). Each of the above reconstructions has implications relating to the crustal composition and structure beneath the Mozambique Channel: 1) Movement of Madagascar southward along a transform fault structure implies, according to Heirtzler and Burroughs (1971), that basement on the east side of the Davie Ridge should be representative of the Madagascar plate and the stratigraphic section should be similar to that of eastern Kenya and Somalia. 2) Movement of Madagascar eastward away from Africa implies the presence of oceanic or thinned continental crust beneath the Mozambique Channel. 3) If no lateral movement of Madagascar relative to Africa has taken place since the Paleozoic, the channel should be underlain by a succession of terrestrial and shallow-marine sediments and volcanics transitional between the stratigraphic sections found in east southern Africa and western Madagascar. One of the major objectives of Deep Sea Drilling Project Leg 25 was to provide essential data which would assist in the solution of this enigmatic problem; and, with this end in view, two sites located in the Mozambique Channel were selected for drilling on the basis of available data and our present meager understanding of the structure and stratigraphy of this region. If Madagascar separated from Africa and drifted eastward, the Mozambique Channel should be underlain by oceanic or thinned continental crust covered by thick sediments shed from Africa and Madagascar. Depending on the timing of drift, the sedimentary succession should include the Tertiary, possibly also the Cretaceous, and perhaps even the Karroo, in which case thick lava successions should be included. If, on the other hand, Madagascar has not moved relative to Africa, the mean elevation of the Mozambique Channel could be explained by the presence of similarly thick succession of supracrustal rocks resting upon continental crust of attenuated thickness (Talwani, 1962). No seismic refraction measurement in the Mozambique Channel are available, but the Gallieni 6 flexotir profile (Figure 2) across the lower Zambesi Canyon clearly shows a prominent deep acoustic basement reflector at about 1.4 sec DT (below the canyon floor) which might be true basement or a particularly hard horizon such as the upper surface of interbedded lavas in the sedimentary sequence. Above this acoustic basement reflector, two other distinct reflecting horizons can be seen; one lies just below the floor of the 600-meter-deep Zambesi Canyon and the other is cut by the canyon. The airgun profiles made by Vema 19 and Thomas B. Davie 267 (Figure 3) show that the uppermost reflector is very prominent, particularly on the western side of the canyon. The position selected for Sites 243/244 was located in the deepest part of the Zambesi Canyon floor in order to achieve the deepest possible penetration of the sedimentary section. The objectives were ( l ) to sample the sediments in transit down the Zambesi Canyon some 800 km from the Zambesi River mouth, (2) to establish the deepest possible stratigraphic succession for possible correlation with the on-shore sections in Mozambique and western Madagascar and with epeirogenic events in both areas, and (3) to identify and date the prominent reflector at 1.4 sec DT beneath the canyon floor.