Abstract The evolution of the global oceanic and atmospheric circulation systems has been affected by several forcing processes, with orbital variations being dominant on shorter geological time scales. Over longer periods of time (>10 Ma) the tectonic evolution of the solid Earth has been recognized as the major control on the development of the global climate system. Tectonic activity acts in one of two different ways to influence regional and global climate. The earliest solid Earth-climatic interaction recognized was the effect that the opening and closure of oceanic gateways had on the circulation patterns in the global ocean. Major effects on regional and sometimes global climate have been attributed to such changes, e.g. closure of the Isthmus of Panama ( Driscoll & Haug 1998 ). Since the late 1980s a second form of climate-tectonic interaction has been recognized, involving the growth and erosion of oro-genic belts. In this second category the Arabian Sea region must be considered the global type example.

Abstract We review previous models for the Paleogene tectonic evolution of the Arabian and Eastern Somali basins and present a model based on a new compilation of magnetic and gravity data. Using plate reconstructions, we derive a self-consistent set of isochrons for Chron 27 to Chron 21 (61-46 Ma). The new isochrons account for the development of successive ridge propagation events along the Carlsberg Ridge, leading to an important spreading asymmetry between the conjugate basins. Our model predicts the growth of the outer and inner pseudo-faults associated with the ridge propagation events. The location of outer pseudo-faults appears to remain very stable despite a drastic change in the direction of ridge propagation before Chron 24 (c. 54 Ma). The motion of the Indian plate relative to the Somalian plate is stable in direction through Paleogene time; spreading velocities decrease from 6 to 3 cm a -1 . Our reconstructions also confirm that the Arabia-India plate boundary was located west of the Owen Ridge along the Oman margin during Paleogene time; some compression is predicted at about Chron 21 (47 Ma) between the Indian and Arabian plates.

Abstract In the northern Arabian Sea the Arabian, Eurasian and Indian Plates are in tectonic interaction with one another. We present interpretations of multichannel seismic profiles across the Makran sub-duction zone (which is part of the Eurasian-Arabian Plate boundary) and the transtensional Murray Ridge and Dalrymple Trough (which are part of the Arabian-Indian Plate boundary). We distinguish four megasequences in the sedimentary succession, which we correlate over the entire study area. Regional unconformities separate the megasequences and enable us to establish a common history of the region before Late Miocene time (c. 20 Ma). The Early Pliocene (c. 4.5 Ma) reopening of the Gulf of Aden caused a reorganization of the plates and subsequent tilting of the oceanic crust of the Arabian Plate toward the Makran subduction zone. This event is documented by the regional M-unconformity. Since that time, sedimentation on the Oman Abyssal Plain has been permanently separated from the Indus Fan by the Murray Ridge, on the northern end of which there has been no significant sedimentation.

Abstract The margins of the Gulf of Oman Basin range from convergent at the north to translation at the west and east, and passive at the south. The basin's northern margin has been a site of continuous subduction since Cretaceous time, which has led to the creation of an 800 km long and 650 km wide accretionary wedge, most of which is above sea level. Strata in the centre of the Gulf of Oman Basin display minor deformation resulting from the northward tilting of oceanic crust. A basin-wide unconformity dividing these strata in two was the result of erosion during Early Oligocene time when bottom water circulation was enhanced during a climatic deterioration. The morphology of the basin's south margin is due to Early Triassic rifting, deposition during Jurassic-Early Cretaceous time, early Late Cretaceous ophiolite obduction and Late Cretaceous-Cenozoic deposition. The western side of the accretionary wedge, along the north side of the Gulf of Oman Basin, is in sharp contact with the western translation margin. Structures along this margin are the result of post-Eocene convergence of the Lut and Central Iran microplates. The eastern end of the accretionary wedge, however, is not in contact with the eastern transform margin, but is separated from it by a north-trending trough. The landward extension of this trough is defined by the north-trending Las Bela Valley. The eastern side of the accretionary wedge turns northward at 65°30'N along the west side of the trough and becomes aligned with the north-trending Ornach-Nal Fault along the west side of the Las Bela Valley. Similarly, the Murray Ridge complex turns northward at 25°N and becomes aligned with the north-trending Surjan Fault on the Las Bela Valley's east side. The Ornach-Nal and Surjan faults merge at the apex of the Las Bela Valley with the north-trending Las Bela-Chaman Structural Axis. Differences between the eastern and western sides of the accretionary wedge may be due to the presence of the Ormara microplate on the eastern end of the wedge, a plate that is being pushed ahead of the Arabian plate. The morphology of the Murray Ridge complex is the result of transtension and secondary compression along the Indian-Arabian plate boundary. We infer that most of the relief of the Murray Ridge complex resulted from a change in plate geometry in Early Miocene time. Subsequent tectonic Pliocene-Quaternary events have enhanced this relief.

Abstract We present a revised magnetic isochron map of the conjugate Arabian and Eastern Somali basins based on an up-to-date compilation of Indian, French, and other available sea-surface magnetic data. We have used the magnetic anomaly and the modulus of the analytical signal computed from the magnetic anomaly to identify and precisely locate the young and old edges of magnetic chrons in both basins. In addition to the major, well-defined anomalies, we have also used correlatable second-order features of the magnetic anomalies, the ‘tiny wiggles’, to strengthen the interpretation. The resulting isochrons and tectonic elements have been validated using the stochastic method of palaeogeographical reconstruction. The magnetic anomaly pattern in both basins depicts clear oblique offsets, characteristics of pseudofaults associated with propagating ridge segments. Our tectonic interpretation of the area revealed: (1) a complex pattern of ridge propagation between Chrons 28n (c. 63 Ma) and 25n (c. 56 Ma), with dominant eastward propagation between Chrons 26n (c. 58 Ma) and 25n; (2) numerous, systematic westward propagations between Chrons 24n (c. 53 Ma) and 20n (c. 43 Ma); (3) asymmetric crustal accretion (caused by ridge propagation and asymmetric sea-floor spreading) in the conjugate basins during the whole period; (4) a slowing of India-Somalia motion after c. 52 Ma.

Abstract The Holocene record on the coasts of the Arabian Sea provides information on the nature and rate of deformation generated by the interaction between the Indian, Arabian and Eurasian Plates. Holocene marine terraces show that the southern Makran has been subject to the infrequent but vigorous coseismic uplift (蠄2 m) that characterizes other subduction settings and they indicate landward rotation of the imbricate faults among which shortening is distributed. The lack of significant Holocene deformation on the SE coast of the Arabian peninsula is consistent with its position parallel to a transform, although there is evidence for large-scale buckling driven by convergence at the Strait of Hormuz in the NE. Geomorpholo-gical and tide-gauge evidence for localized uplift on the southwestern coast of India may represent compressional buckling here too in response to Himalayan collision. Bathymetric and geodetic data can help to bridge these sequences and thus enhance their value for quantifying plate rheology and dynamics, notably by linking variations in plate-margin displacement with major sea-floor strike-slip structures and by eventually confirming transitory as well as sustained compressive buckling on land.

Abstract The Indus Group is a Paleogene, syntectonic sequence from the Indus Suture Zone of the Ladakh Himalaya, India. Overlying several pre-collisional tectonic units, it constrains the timing and nature of India's collision with Eurasia in the western Himalaya. Field and petrographic data now allow Mesozoic-Paleocene deep-water sediments underlying the Indus Group to be assigned to three pre-collisional units: the Jurutze Formation (the forearc basin to the Cretaceous-Paleocene Eurasian active margin), the Khalsi Flysch (a Eurasian forearc sequence recording collapse of the Indian continental margin and ophiolite obduction), and the Lamayuru Group (the Mesozoic passive margin of India). Cobbles of neritic limestone, deep-water radiolarian chert and mafic igneous rocks, derived from the south (i.e. from India), are recognized within the upper Khalsi Flysch and the unconformably overlying fluvial sandstones of the Chogdo Formation, the base of the Indus Group. The Chogdo Formation is the first unit to overlie all three pre-collisional units and constrains the age of India-Eurasia collision to being no younger than latest Ypresian time (>49 Ma), consistent with marine magnetic data suggesting initial collision in the Arabian Sea region at c. 55 Ma. The cutting of equatorial Tethyan circulation north of India at that time may have been a trigger to the major changes in global palaeoceanography seen at the Paleocene-Eocene boundary. New 40 Ar/ 39 Ar, apatite fission-track and illite crystallinity data from the Ladakh Batholith and Indus Group show that the batholith, representing the old active margin of Eurasia, experienced rapid Eocene cooling after collision, but was not significantly reheated when the Indus Group basin was inverted during north-directed Miocene thrusting (23-20 Ma). Subsequent erosion has preferentially removed 5-6 km (c. 200 °C) over much of the exposed Indus Group, but only c. 2 km from the Ladakh Batholith. Reworking of this material into the Indus fan may complicate efforts to interpret palaeo-erosion patterns from the deep-sea sedimentary record.

Abstract Petrographic and geochemical data for new basalt and peridotite samples recovered from sampling sites at the Southern Murray Ridge help to constrain models for the evolution of the Owen-Murray Ridge system, which forms the northwestern boundary of the Indian plate. Trace elements immobile during alteration (Ti, Zr, Nb, Y and rare earth elements) suggest that the altered microphyric metabasalt has affinities to magmatism of active margins (island-arc tholeiite sensu lato). It is distinctly different from mid-ocean ridge basalt, back-arc-basin basalt, or intra-plate Deccan Trap basalt. The sampled serpentinized harzburgite or clinopyroxene-poor lherzolite was deformed under mantle conditions and is similar to the mantle section of nearby ophiolite sequences. This association of rocks suggests that an ophiolite melange was sampled. However, results from sampling station 462 NIOP indicate that the Murray Ridge complex also contains igneous rocks with Deccan Trap affinity. For the emplacement of the island-arc tholeiite we assume an origin in a convergent supra-subduction setting, related to the closing of a Late Cretaceous Neo-Tethyan ocean basin between the Arabian and Indian plates to the south and the Eurasian plate to the north. Since Neogene time, the Murray Ridge-Dalrymple Trough has been underlain by attenuated (?)continental crust and characterized by extensional rift tectonics.

Abstract We present a numerical model of the geothermal field of the Makran accretionary prism and of the slab being subducted below it. Calculated heat flow density values for the sea floor of the abyssal plain and the shelf slope are compared with in situ measured and bottom simulating reflector (BSR)-derived heat flow density values. The result suggests a predominance of conductive heat transport within the accretionary complex. Little evidence is found to suggest that fluid flow or frictional heat modifies the observed geothermal field to any great extent. We also studied the geothermal field associated with the decay of the potential gas hydrate layers (indicated by the presence of BSRs), as gas hydrate layers are being tectonically uplifted out of the gas hydrate stability field into shallower and warmer sea water. Theoretical considerations suggest a complete disappearance of gas hydrates at a water depth of about 750 m. The observed presence of numerous gas seeps almost exclusively at water depths of less than 800 m suggests that gas hydrate layers in the Makran accretionary prism act as a very effective cap rock to upward-directed flow of fluids containing notable amounts of dissolved gas from within the prism to the sea floor.

Abstract The Iranian Makran has been entirely mapped geologically on a scale of 1:250 000, except for a narrow coastal strip, which exposes the very youngest Cenozoic sediments of the main Makran accretionary prism. The geology of the Makran is less widely known than the geology of Oman, because it has been published in detail only in reports of the Geological Survey of Iran. There is no extension of the geological formations of Oman into the Makran, the only extension of Oman ophiolitic formations into Iran being at Neyriz and Kermanshahr, hundreds of kilometres to the NW. This summary is based on field mapping, photo-interpretation being used only to connect traverse lines. The oldest rocks are metamorphic rocks, which form the basement to the Bajgan-Dur-kan microcontinental ‘sliver’, a narrow block that extends hundreds of kilometres from the Bitlis Massif in Turkey, through the Sanandaj-Sirjan Block of the Zagros, to north of Nikshahr in the east of the Makran. Other metamorphic rocks form the Deyader Complex near Fannuj on the southern margin of the Jaz Murian Depression. These include blueschists, and are thought to form the tip of the Tabas Microcontinental Block, largely exposed north of the depression. There is also a small microcontinental block to the east, the Birk Block, which exposes only Cretaceous platform limestones and Permian sediments. The Bajgan Metamorphic Series are overlain, with a tectonized unconformable contact, by highly deformed and disrupted platform carbonates of Early Cretaceous to Early Paleocene age (Dur-kan Complex), containing tectonic inliers of Carboniferous, Permian and, rarely, Jurassic age. Ophiolites occur in two structural positions. South of the Bajgan-Dur-kan Block, the tectonic Coloured Melange of the Zagros continues eastwards inland of the Bashakerd Fault; this includes two layered ultramafic complexes, one with chromitites. The blocks forming the melange include radiolarites and deep-water limestones of Jurassic to Early Paleocene age. Ophiolites developed north of the microcontinental block form three distinct igneous complexes, two layered and one with intermediate sheeted dykes. Intercalated in the volcanic rocks of these ophiolites are radiolarites and deep-water limestones ranging in age from Jurassic to Paleocene time. There are small developments of Cretaceous sediments carrying rudists in the extreme NW of the inner ophiolite tract. In the NE, ophiolites are developed in the Talkhab Melange. All these ophiolites represent former, largely Cretaceous, tracts of deep ocean. The Cenozoic rocks form two immense accretionary prisms. The main Makran prism includes Eocene-Oligocene and Oligocene-Miocene flysch turbidite sequences, estimated as individually > 10 000 m thick. Above these sequences, there is an abrupt passage up without any apparent unconformity, through reefal Burdigalian limestones, and locally a harzburgite conglomerate development, into neritic sequences with minor turbidites, extending into the Pliocene units. The Saravan accretionary prism to the east repeats tectonically three thick flysch turbidite sequences of Eocene-Oligocene age, but younger sediments are restricted here to minor Oligocene-Miocene conglomerates, unconformable on the above sequences. There is a line of OHgocene(?) granodiorite bodies within the Saravan accretionary prism. Intense folding and development of schuppen structure, dislocation and melanging of the sediments affected the entire region in Late Miocene-Early Pliocene time. Post-tectonic uplift was followed by scattered developments of fanglomerates beneath the fault scarps. The Neogene deformation has obscured earlier deformational events. There is unconformity beneath Eocene sediments representing a mid-Paleocene disturbance. There is also evidence of a discontinuity in mid-Oligocene time. Pliocene-Pleistocene fanglomerates are unconformable on folded rocks. There are discontinuous developments of Eocene-Oligocene neritic sediments unconformably above the older rocks (ophiolites, platform limestones, metamorphic rocks), and to the north of the southern edge of the Jaz Murian Depression, the northern limit of the Makran, there is evidence of the survival here of a very shallow sea through Neogene time and the formation of small patches of reefal Oligocene-Miocene limestones, and Eocene to Pliocene shallow-water clastic sediments. A 150 km wide tract separates the coast from the trench, the total Cenozoic accretionary prism being 500 km wide. Extension from the Murray Ridge affects the extreme east of the region. The Saravan accretionary prism, it is suggested, faced a gulf, comparable with the Gulf of Oman, and this Saravan Gulf filled up and closed up by Early Oligocene time. Seismological evidence suggests that there is now active continental collision continuing along this suture.

Abstract The Zagros Mountains are situated along the NE margin of the Arabian plate and are the product of complex deformation which began in Late Cretaceous time as a result of the collision between the Arabian and Central Iranian plates. During Pliocene time, deformation increased when plate convergence was accelerated by the opening of the Red Sea. This stimulated the migration of a deformation front from the collision zone towards the SW into the undisturbed Zagros basin and led to the creation of the Zagros Mountain Belt. The type and distribution of the deformation in the Zagros are controlled mainly by plate velocity, which is linked to the anticlockwise rotation of the Arabian plate around a pole in Syria, and the regional stratigraphy. The sedimentary cover and the underlying metamorphic basement decouple along an important detachment horizon, the Hormuz Salt Formation, and the uneven thickness and distribution of this salt plays a key role in determining the geometry of the deformation belt. Analysis of the distribution and geometry of the folds provides evidence for the southwestwards migration of the deformation front into the Arabian plate. The analyses are consistent with field evidence for serial folding, which indicates that each fold takes c. 600 ka to develop fully, and with the model of a southwestward advancing deformation front driven by the processes of serial folding and footwall collapse.

Abstract The Makran slope-apron system is a stepped convergent margin across an active subduction complex. Shallow penetration piston cores have been recovered from the upper-slope region, three mid-slope basins and the abyssal plain. At most sites the upper 5-14 m of cored section is dominated by fine-grained, thin- to medium-bedded turbidites, averaging 5-10 turbidite events per metre of section. Oxygen isotope stratigraphy yields mean sedimentation rates of 50-95 cm ka −1 and a turbidite frequency of one event per 200-300 a. The upper-slope site has fewer turbidites and a greater proportion of hemipelagic mud. Fine-grained turbidite sequences are common, with top-cut-out and base-cut-out sequences most evident. Markov chain analysis of the transition between turbidite divisions confirms the normal T0-T8 order of sequence divisions. In some cases there is an upward gradation into a hemiturbidite facies. The range of turbidite bed thicknesses can be approximated by both power-law and log-normal distributions, typical of seismic triggering on an active margin, or of frequent river-flood sediment input. Small-scale vertical variations of turbidite bed thickness recognized by autocorrelation techniques can be interpreted as the result of bed-relief compensation effects (compensation cycles). The lateral distribution of both turbidites and hemipelagites is influenced by sediment focusing along pathways between slope basins. At a larger scale, climate, sea-level and tectonic effects have all played an important role in shaping margin sedimentation.

Abstract The Indus River system is one of the largest rivers on the Asian continent, but unlike the Ganges–Brahmaputra system, the drainage of the Indus is dominated by the western Tibetan Plateau, Karakoram and tectonic units of the Indus Suture Zone, rather than the High Himalaya. The location of the river system relative to the Indus Suture Zone explains the deep exhumation north of that line in the Karakoram, compared with the modest erosion seen further east in Tibet. The modern Indus cuts Paleogene fluvial sedimentary rocks of the Indus Group located along the Indus Suture Zone in Ladakh, northern India. After the final marine incursion within the Indus Group in the early Eocene (<54.6 Ma), palaeo-current indicators changed from a north-south flow to an axial, westward pattern, synchronous with a marked change in sediment provenance involving erosion of South Tibet. The Indus probably was initiated by early Tibetan uplift following the India-Asia collision. The river has remained stationary in the suture since Early Eocene time, cutting down through its earlier deposits as they were deformed by northward folding and thrusting associated with the Zanskar backthrust at c. 20 Ma. The Indus appears to have been located close to its present position within the foreland basin since at least Mid-Miocene time (c. 18 Ma), and to have migrated only c. 100 km east since Early Eocene time. In the Arabian Sea Paleogene fan sedimentation was significant since at least Mid-Eocene time (c. 45 Ma). Sediment flux to the mid fan and shelf increased during Mid-Miocene time (after 16 Ma) and can be correlated with uplift of the Murray Ridge preventing sediment flow into the Gulf of Oman, tectonic uplift and erosion in the Karakoram and western Lhasa Block, and an enhanced monsoon triggered by that same uplift. Sedimentation rates fell during Late Miocene to Recent time. The Indus represents 18% of the total Neogene sediment in the basins that surround Asia, much more than all the basins of Indochina and East Asia combined (c. 11%). Unlike the rivers of East Asia, which have strongly interacted as a result of eastward propagating deformation in that area, the Indus has remained uninterrupted and represents the oldest known river in the Himalayan region.

Abstract In 1997 Lasmo Oil Pakistan Ltd (Lasmo) gained a significant position in the offshore Indus Basin with the award of the Indus A and B Blocks. The main hydrocarbon play comprises Miocene shelf-delta sands interbedded with intraformational shale seals and sourced by gas-prone offshore equivalents. Approximately 12000 km of seismic data have been interpreted in the detailed evaluation of these blocks. However, only four wells have tested the preferred play type and no core or rock data were available to provide further insights into facies or age dating. Log data from two key wells in the offshore Indus area record the initial infilling of the basin by shale-dominated basinal or outer shelf sediments, followed by stacked thin-bedded sandstone-shale sequences of a shelf-delta nature. A zone of progradational sequences marks the transition between the two, but no other workable stratigraphic divisions were apparent. Regional seismic correlation established the diachronous nature of the prograding shelf package and this was matched by distinct bands of seismic progrades. A series of simple palaeogeographies of the prograding shelf margin were developed showing initial sediment input from the north and rapid progradation towards the south and west. The Oligo-Miocene basin fill of the offshore Indus Basin appears to be a 'one-step' fill process of a significant depocentre created between the Karachi Platform and the Murray Ridge. Canyons are a very distinct feature on seismic profiles and two main phases of development are apparent. The earlier phase is interpreted to be of Early Miocene age. Downcutting at this time rarely exceeded 400 m. The second phase of canyon development occurred during Plio-Pleistocene time, and these younger canyons often dominate the shallow section, with multiple phases of downcutting sometimes exceeding 1000 m. Where drilled, canyons of both ages have been found to be shale prone. These drilled canyons are interpreted to be on the palaeo-slope where erosion and sediment by-pass occurred during the active phase, and were subsequently filled by fine-grained deposits after abandonment. The two phases of canyoning are considered to relate to phases of tectonic activity in the collision zone between the western margin of the Indo-Pakistan plate and the Eurasian plate.

Abstract The southern Kirthar Fold Belt (KFB) and the contiguous Middle Indus Basin (MIB) constitute a major oil and gas province on the southern Pakistan foreland. In the Middle Indus Basin, gas is reservoired in Early Cretaceous marginal marine sandstones sealed by Paleocene shales. Reservoir quality of the Early Cretaceous sediments deteriorates towards the fold belt, but recent discoveries of gas in Upper Cretaceous sandstones sealed by Paleocene shales have highlighted its potential. To understand better the petroleum systems of this region and provide a potential correlation scheme for comparison with other areas of the country, a sequence stratigraphic interpretation was carried out on the Jurassic to Recent sediments of the KFB and MIB. On the basis of outcrop and well data, 23 depositional sequences have been identified: five in the Jurassic units, 10 in the Cretaceous units and eight in the Tertiary units. Sequence boundaries have been defined according to the Exxon method of identifying unconformities and their correlative conformities. However, equal importance has been given to identifying the potentially more chronostratigraphically significant maximum flooding surfaces between these sequence boundaries so as to define accurately the component systems tracts of each sequence. The depositional systems are described in terms of their relationship to the existing lithostratigraphic framework and interpreted in terms of sedimentary responses to external (eustatic) or local (tectonic) events. Notwithstanding the presence of a eustatic signature on some sequences, the majority appear to be tectonically driven and can be related to plate margin events affecting the NW margin of the Indo-Pakistan Plate during its rift-drift-collision history.

Abstract The distribution of sand dunes over the southern half of Arabia conforms to the influence of two wind systems: the northern Shamal, which is a strong wind that blows to the SSE down the Persian (Arabian) Gulf and then swings to the SW across the hyperarid Rub al Khali towards North Yemen; and the strong winds of the SW Monsoon system, which were responsible for forming linear dunes that trend north-south in the Wahiba Sands of eastern Oman and SW-NE in the Thar Desert (NW India). In the Thar Desert, the SW Monsoon alternates with the weaker NE Monsoon. The dating of exposures of older dune systems by isotopic, radiometric and optically stimulated luminescence (OSL) analyses has shown that the Shamal was active throughout the latter part of the Quaternary period, and probably as long ago as Mid-Miocene time ( c. 15 Ma). At times of glacial maxima, when global sea level was some 100-120 m or more lower than now, siliciclastic and carbonate grains were deflated from the exposed surface of the Persian Gulf and transported into the NE Rub al Khali within the United Arab Emirates. It is suspected that occasionally the Shamal also transported some quartz sands from the NW onto the exposed narrow continental shelf of SE Arabia, with silt-size particles being carried into the Arabian Sea. The SW Monsoon, on the other hand, was re-established over the coast of SE Arabia several millennia after the last glacial maximum and was fully established near the coast of SE Arabia during the early Holocene interglacial after the atmospheric high-pressure system associated with the glacial period had become weaker. Early during the Holocene interglacial periods when the SW Monsoon dominated, a combination of quartz and carbonate sands was deflated from the exposed continental shelf and transported to the north into the Wahiba Sands. Aeolian activity in the Thar Desert also peaked during this period of transition from full glacial to interglacial conditions. The dune systems of SE Arabia overlie the distal edges of older alluvial fans that in Oman date back at least 350 ka. The sediments of some of these fluvial sequences in Oman reached the Arabian Sea via Wadi Batha, only to be removed by along-shore currents driven by the SW Monsoon. In the Thar Desert, the supply of aeolian sediment is mostly from fluvial sources. Marine sediments from the Arabian Sea between Arabia and Thar record the contrasting effects of the Shamal and the SW Monsoon: the former mostly as a source of wind-blown dust from Arabia and the latter by causing upwelling of nutrient-rich waters leading to organic blooms.

Abstract Absolute and relative abundances of calcareous dinoflagellate cyst species in surface sediment samples from the Arabian Sea are compared with environmental parameters of the upper 100 m of the water column to gain information on their largely unknown autecology. Ten species or morphotypes were encountered of which four occurred only as accessories. On the basis of the distribution patterns of the six more abundant species or morphotypes, the studied area is subdivided into three provinces, demonstrating a clear relationship to monsoon-controlled upper-ocean conditions. The two dominant species, Thoracosphaera heimii and Orthopithonella granifera, show opposite trends in distribution of both their absolute and relative abundances. In the NE Arabian Sea, low absolute and relative abundances of T. heimii are mainly attributed to enhanced dissolution of the small tests in this region, whereas elevated concentrations of O. granifera seem to be related to higher water temperatures and the influence of the Indus River. Sphaerodinella albatrosiana and Calciodinellum operosum are most abundant in the open ocean, associated with lower nutrient levels, relatively high temperatures and low seasonality. Spiny cysts (mainly represented by Scrippsiella trochoidea), in contrast, exhibit a more shelf-ward distribution and are most abundant in regions that are influenced by coastal upwelling, characterized by eutrophic and rather unstable conditions with seasonally lower temperatures and a shallow thermocline. A generally negative correlation of calcareous dinoflagellate cysts with primary productivity or high nutrient concentrations, as proposed by other workers, cannot be confirmed. Cyst accumulation rates off Somalia show that strong turbulence and high current speeds are unfavourable for calcareous dinoflagellates, suggesting that these organisms are more successful under rather stratified conditions.

Abstract We present a multi-proxy study of sediment Core 905 from the Arabian Sea offshore Somalia to assess the validity of a number of proxies for productivity, temperature and wind strength, to reconstruct the monsoon history in the western Arabian Sea. The present-day seasonal variation in productivity in the modern Arabian Sea off Somalia reflects the change from the high-productivity SW monsoon to the low-productivity NE monsoon seasons. Annual productivity is therefore largely controlled by SW monsoon driven upwelling. The geochemical records of Core 905 document millennial-scale variations, for example, in Ba/Al and C org content. The Younger Dryas and the time equivalent period to Heinrich event 1 show low annual productivity whereas the early Holocene and Bølling-Allerød periods are characterized by high productivity. The upwelling-productivity peaked during Early Holocene time and was followed by a decrease toward the modern values. The total flux of planktic foraminifera and the concentration of the planktic foraminifera G. bulloides are not always controlled by the total productivity. Variations in calcite dissolution, the advection of expatriate fauna or a seasonal decoupling of primary and secondary production appear to hamper straightforward interpretations of those foraminifera records. We conclude that at significantly changed climatic boundary conditions compared with the present day, bulk-sediment-related proxies of productivity more consistently record the local upwelling history than foraminifer-based productivity proxies.

Abstract Settling nitrogen fluxes intercepted by sediment traps on the mid-slope and in the deep basin off Somalia show a consistent annual range of 3.4 ± 0.2‰ in their stable isotope composition. Seasonal minima in δ 15 N of 3.7‰ are associated with the moderate N fluxes derived from coastally upwelled water, which is rapidly carried offshore along eddy margins passing over the mooring sites during the SW monsoon (June-September). Coastal upwelling, offshore transport and deep wind mixing cease at the end of the SW monsoon, leading to enhanced utilization of the up to 20 μm of NO 3 − in the photic layer, maxima in the N export flux, and an increasing δ 15 N by Rayleigh distillation. Yet as stratification develops, nutrient exhaustion follows and export production collapses as the δ 15 N increases to over 7‰. Cyanobacterial N 2 fixation probably diminishes the δ 15 N by 0.4-1.6‰ during the autumn intermonsoon (November-December) when settling N fluxes are lowest. Nutrient utilization remains high during the NE monsoon (January-March), when nutrient entrainment by deep wind mixing results in enhanced N export with maxima in δ 15 N of up to 7.4‰. Annual N fluxes have virtually the same δ 15 N of 6.0‰ in all traps despite considerable differences in both N flux and δ 15 N between the traps during the year and at different depths. In comparison with the annual δ 15 N of 6.0‰ arriving on the sea floor, core-top sediments are enriched by +0.6‰ on the upper slope (at 487 m) increasing to + 2.9‰ in the deep basin (at 4040 m), whereas the N sediment burial efficiency declines from about 17% to 3%. It appears that the extent of oxic decomposition at the sediment-water interface is the most likely cause of such isotope enrichment. Similar positive gradients in δ 15 N with bottom depth have been reported from other continental margin transects and are generally attributed to increased nutrient utilization in the photic ocean with distance offshore. As for Somalia, nitrogen isotope fractionation as a result of oxic decomposition on the bottom rather than nutrient utilization at the ocean surface may account for the observed increase of sedimentary δ 15 N down continental margins in general.