ABSTRACT

During late Early Ordovician times an increase in the rate of subsidence in the Ghaba Salt Basin and western South Oman Salt Basin is suggested by the thick sequence of continental clastics of the Ghudun Formation. After a phase of rift-shoulder uplift and erosion, related to a renewed pulse of extension which may have initiated diapiric growth of salt structures in the Ghaba Salt Basin, sedimentation resumed again in the Mid Ordovician. During this period, the center of deposition shifted to the Saih Hatat area in North Oman. This paper documents seismic and well data, field investigations and petrological study of potassic mafic rocks from the Huqf area which were intruded in the eastern side of the Ghaba Salt Basin. A Mid Ordovician age of 461 ± 2.4 million years has been established for these rocks by the Argon-Argon step heating method. Analogy with the petrology and setting of similar potassic mafic rocks from the Rio Grande Rift in the western United States of America suggests that they were intruded into the shoulder of an intra-continental rift. The data provide the first clear evidence of a pulse of rift-shoulder uplift in the Huqf area during the Mid Ordovician. The 3-kilometer-thick Mid to Late Ordovician clastic sediments of the Amdeh Formation in North Oman, together with the occurrence of abnormally thick sedimentary sequences and volcanics in the Tabas Graben in Iran, are consistent with a period of break-up of eastern Gondwana. Together, the Ghaba-Saih Hatat and Tabas Basins are considered to be part of a failed rift arm. These observations further improve our regional knowledge of the Early to Late Ordovician tectonic setting of Oman and will assist in unlocking the hydrocarbon potential of classical rift-related structures consisting of early-rift Early Ordovician sand-prone reservoirs sealed by syn-rift Mid to Late Ordovician marine shales.

INTRODUCTION

In Central and South Oman, a series of northeast-southwest to north-northeast–south-southwest trending extensional basins are developed on Precambrian crystalline basement (Figure 1). Episodic basin formation in Gondwana was initiated in Late Precambrian times following the formation of the Rodinia supercontinent (Stern, 1994). In Oman, the western margin of these basins is marked by a structurally complex compressional zone of late Early Cambrian age defined as the “Western Deformation Front” (Loosveld et al., 1996). Deformation increases in intensity towards the southwest across the South Oman Salt Basin into Yemen. This tectonic pulse probably signaled the end of the indentation of West Gondwana into East Gondwana (Stern, 1994) and is associated with regional uplift and erosion over the entire Arabian Plate (Stump et al., 1995). The eastern margin of Oman’s salt basins in contrast is less tectonized and Phanerozoic sediments offlap, onlap onto, or are truncated against, a structural high known as the Huqf High (Gorin et al., 1982). This high is located close to the present east coast of Oman.

Subsequent phases of uplift of the Huqf High resulted in the present-day exposure of rocks ranging in age from the Precambrian granitoid basement to Tertiary sediments (Figure 2). The outcropping sequence contains both major and minor unconformities. Large sections of the stratigraphic column are absent, particularly in the Late Paleozoic due to lack of accommodation space and erosion during a Mid Carboniferous Hercynian event. Documented rift-shoulder uplift pulses in the Late Carboniferous to Early Permian and again in Late Triassic to Early Jurassic times preceded the break-up of Gondwana in Late Jurassic times (Blendinger et al., 1990; Immenhauser et al., 1998). In many cases, the exact nature of the unconformities, particularly the minor ones, remains to be elucidated. In this respect, igneous rocks are important as they provide petrological and geochronological evidence pertinent to the plate-tectonic setting and dating of the extensional structural events with which they are associated (Wilson, 1994).

In this paper, we initially review the regional geology and structural history of Oman during the Late Precambrian and Early Paleozoic. The paper then documents the petrology and geochronology of potassium-rich mafic intrusions exposed in the northern Huqf. These volcanics provide evidence for a subtle rift event. We integrate this information with the Early Paleozoic tectonic and sedimentological evolution of Oman and surrounding areas of the Arabian Peninsula.

For the Cambrian to base Silurian, the time scale of Gradstein and Ogg (1996) has been applied (Figure 2). For the younger Phanerozoic we use the time scale of Harland et al. (1990). The main change compared to the time scale of Harland et al. (1990) is the shift in time of the base Cambrian from 570 to 545 million years ago (Ma). Presently, the base Cambrian is assigned an absolute age of 543 Ma (Knoll et al., 1995; Brasier et al., 1997).

REGIONAL GEOLOGY

Onset of Sedimentation onto Basement up to the Angudan Event, Late Early Cambrian

About 560 Ma, a change took place from predominantly transpressional tectonics to a predominantly transtensional regime on the Arabian Peninsula. Sediments older than 560 Ma are present, but their distribution in time and space is poorly known in Oman. The increase in extension occurred at the end of a protracted continent-continent collision between East and West Gondwana (e.g. Stern, 1994; Le Métour et al., 1995; Grunow et al., 1996; Shackleton, 1993 and 1996; Blasband et al., 1997). Uplift and erosion of large areas of continental crust is suggested by the rapid and significant increase in seawater 87Sr/86Sr from 0.708 to 0.709 during deposition of the Huqf Supergroup (Burns et al., 1994). Intra-continental extension is attributed to an extended phase of indentation of West Gondwana into East Gondwana (Schmidt et al., 1979; Stern, 1994). The sinistral northwest-southeast Najd Trend with a cumulative strike-slip component of about 300 kilometers (km) (Moore and Al-Shanti, 1979; Al-Husseini, 1988) is evidence of this continental deformation. In Oman, Loosveld et al. (1996) described two major Infracambrian-Early Cambrian cycles of intracratonic rifting (Figure 2) which formed northeast-southwest to north-northeast–south-southwest trending basins. These rift basins are approximately perpendicular to the Najd Fault Zone and the orogenic collision front and were considered to be related to escape tectonism (Molnar and Tapponnier, 1975; Stern, 1994). An alternative, but not necessarily mutually exclusive, model to explain the Late Precambrian to Early Cambrian extension is based on continental failure and rifting of the Gulf salt basins following a northwest-southeast trend (Al-Husseini and Al-Husseini, 1990; Talbot and Alavi, 1996). Subsequent rifting of Paleo- and Neo-Tethys followed this early trend (Stampfli et al., 1991).

In Oman, Rift Cycle I took place in the Late Precambrian forming the Abu Mahara Basin. Geophysical data and the few well penetrations suggest the development of a northeast-southwest trending rift basin segmented by a series of north-south to northeast-southwest-trending basement highs such as the Mabrouk-Makarem High (Figure 1). Rifting was accompanied by volcanism and granite intrusions. Transitional and tholeiitic basalts are known from the Saih Hatat area (Le Métour, 1988), whereas more felsic igneous rocks occur in Central and South Oman. Rhyolitic volcanics at the base of the Abu Mahara Group in the northern Huqf area have Rb/Sr ages of 562 ± 42 Ma and red alkali granite intrusions in the Marbat area were dated at 556 ± 11 Ma (Dubreuilh et al., 1992; Platel et al., 1992).

The Abu Mahara Basin, referred to above, was re-activated during earliest Cambrian Upper Huqf times indicating renewed tectonic activity and rapid subsidence. Loosveld et al. (1996) refer to this stage as Rift Cycle II. However, ongoing evaluation of the western flank of the South Oman Salt Basin suggests that it may be more appropriately described as a foreland basin. (Immerz, personal communication, states that a foreland basin was earlier proposed by Boserio et al., (1995), but the development of the foreland basin followed the deposition of the Ara Group in their model.) Carbonates, shales, silicilytes and evaporites of the Ara Group are conformably to unconformably overlain by clastics of the Lower Cambrian Nimr Group, recently incorporated within the Huqf Supergroup (Hartkamp-Bakker et al., 1998) (Figure 2). The Angudan Unconformity, separating the Nimr Group from the Haima Supergroup (Figure 2), is considered to represent the end of the East African Orogeny, during which the indentation of West Gondwana into East Gondwana came to a halt, and escape tectonism ceased (Hartkamp-Bakker et al., 1998). Within the clastics of the Nimr Group, many major and minor erosional gaps reflect shifting alluvial fans and pulses of uplift of the hinterland.

The Pan-African cycle, of which the East African Orogeny formed the active part during its waning stages, ended in about Early to Mid Cambrian times and was followed by uplift and cooling of almost the entire Gondwana Continent. This is indicated by apatite fission track ages (AFTA) which do not exceed about 500 Ma versus AFTA ages as old as 1.2 billion years in Laurasia (Veevers, 1995). AFTA data to the southwest of the South Oman Salt Basin measured in mildly anchimetamorphic granitoid basement in Ghudun-1 follow this 500 Ma uplift trend. However, along the eastern flank of the South Oman Salt Basin, AFTA data in pre-Angudan Nimr sediments give ages older than 500 Ma, for example in Rima-1 they range from 670 to 550 Ma (Visser, 1991). For this area, it indicates continued “cooling” without major subsidence after sedimentation.

The large influx of clastics during the late extension phase reflects this last tectonic pulse of the Pan-African cycle and the initiation of salt diapirism in the South Oman Salt Basin. After the extension phase during Ara deposition and the Angudan event, the sag phase resulted in the deposition of the Mid Cambrian to Lower Ordovician Mahatta Humaid Group (Figure 2).

Basin Development following the Angudan Event

Significant overstepping of the Mahatta Humaid clastics from the South Oman and Ghaba Salt Basins onto the Huqf High took place during the late Early Cambrian to Early Ordovician. A flat topography is indicated by the lateral continuity of transgressive-regressive cycles over the entire Fahud-Ghaba Salt Basins (Figure 3). Continental clastics dominate the lower part of the Mahatta Humaid Group, whereas marine clastics characterize the upper part (Droste, 1997). Towards the top of the group, coarsening-upwards deltaic to fluvial clastic sediments of the Arenigian Ghudun Formation dominated once more. In the Ghaba and Fahud Salt Basins, the Ghudun Formation is characterized by thick monotonous proximal braid plain sandstones, thinning over Petroleum Development Oman (PDO) concessions from east (Andam-1, 1,020 meters (m) to west Makarem-1, 300 m, Figure 4). Dipmeter data show dips to the west-northwest to northwest, offlapping from the Huqf axis over Saih Nihayda field. This orientation is consistent with the observation on seismic data of west-northwest to northwest prograding clinoforms and downlap surfaces within the Ghudun Formation (Figure 5).

Ghudun sedimentation was terminated by uplift. The subsequent erosion was more severe in South Oman than in North Oman and more severe over the highs than the lows (Figure 6). Based on the truncation of regionally correlated intra-Ghudun flooding surfaces, up to several hundreds of meters of Ghudun Formation were removed from highs separating the two northern Oman salt basins (Droste, 1997) (Figure 6). The significant thickness of the Ghudun Formation in the lows and subsequent erosion over block-faulted highs is considered to be related to a rifting event. The lack of regional erosion documents the absence of a rift dome or a phase of regional compression. Unfortunately, detailed biostratigraphic control for the poorly fossiliferous Mahatta Humaid and especially for the barren sandy Ghudun is lacking and does not allow precise calculation of an increased rate of subsidence. However, a significant increase during deposition of the late Early Ordovician Ghudun Formation in the Ghaba Basin is evident.

In southwest Oman, west of the Ghudun-Khasfah High and Anzauz-Rudhwan Ridge, the Ghudun Formation onlaps onto the slightly deformed Abu Mahara indicating a westward tilting of the area as the depositional axis shifted away from the South Oman Salt Basin (Loosveld et al., 1996). This westward tilting during Arenigian times, opposite to the direction of subsidence along the western flank of the Late Precambrian-Early Cambrian rift basins, is considered another sign of renewed extension over a broad area. Overall tilting of the eastern flank of South Oman to the west-northwest continued during deposition of the Caradocian Hasirah Formation, onlapping onto the Ghudun Formation. In this area, the basal part of the Safiq Group (the Llanvirnian-Llandeilian Saih Nihayda Formation) was not deposited (Figure 7).

After uplift and erosion of the Ghudun Formation, a renewed period of extension took place in the Ghaba Salt Basin during Mid Ordovician times. Seismic evidence of rift-shoulder uplift and erosion over the Mabrouk-Makarem High is shown in Figure 8. This tectonic event is based on the observed truncation of the Ghudun Formation towards the high and it is followed by rapid Mid Ordovician subsidence (‘collapse’) as indicated by the drape of the anoxic Hanadir Shale over the continental Ghudun Formation. Clearly expressed on the seismic data is the downlap (‘outbuilding’) of the deltaic highstand deposits of the Saih Nihayda Formation onto the Hanadir flooding surface. The onlap of the Hanadir Shale onto the Ghudun Formation is suggested, but not expressed, on this line because of truncation by the Al Khlata Formation beyond about shot point 4,300.

In addition, analysis of the seismic and well data indicates that the rate of halokinesis increased during this period as shown by an increased rate of downbuilding and ponding of deepwater mass flows against diapir flanks. These mass flows belong to the Saih Nihayda and Hasirah formations (Figure 9). Notwithstanding continued halokinesis in the Ghaba Salt Basin during Cambrian to Early Ordovician times (Boserio et al., 1995), in the following period salt grew diapirically over reactivated basement faults related to the fault pattern of the former Huqf extension cycles. Regional extension has been shown to trigger the rise of diapirs by faulting the brittle overburden (Vendeville and Jackson, 1992). The trap formation of the Saih Nihayda field is related to this upper Ghudun-lower Safiq period of extension and is indicated by en-echelon faulting of its crest (Figure 10). The tilted fault blocks are onlapped by transgressive basal sands and eventually sealed by shoreface mudstones of the Saih Nihayda Formation during a renewed pulse of extension following rift-shoulder uplift and erosion (Figures 9 and 11).

The Mid Ordovician increase in the rate of subsidence is probably the result of the pulse of extension that initiated a basin-wide change in depositional environment. During deposition of the Arenigian Ghudun Formation in the Ghaba Salt Basin, the environment varied only slightly from alluvial plain to deltaic, whereas during Mid Ordovician to Early Silurian times the environment changed significantly from alluvial plain in the south to marine (outer shelf) in the north (Figure 9). This rise in relative sea-level on the Arabian Peninsula occurred at the onset of Safiq deposition during a period of a significant global sea-level fall (Stump et al., 1995; Vail et al., 1977). This is further evidence for an extensional tectonic pulse under the assumption of rather constant sediment supply.

The Safiq Group has been subdivided into three major transgressive-regressive (TR) cycles (Droste, 1997; Partington et al., 1998). A major flooding occurred at the base of each TR-cycle, followed by progradation of braid-plain deltas flowing into deep water, resulting in mass flow deposits in the Saih Nihayda and Hasirah formations. The youngest cycle, the Early Silurian Sahmah Formation, escaped erosion only along the west flank of the Ghudun-Khasfah High and Anzauz-Rudhwan Ridge, and locally in paleo-lows within the Fahud and Ghaba Salt Basins formed by salt withdrawal.

During the Caradocian, at the end of the deposition of the Hasirah Formation, the impact of a shortlived, intense pulse of polar glaciation was felt. The advance of ice was at its maximum in the Ashgillian (Vaslet, 1990; Al-Husseini, 1990). This glaciation was responsible for a eustatic sea-level fall, resulting in significant erosion along the basin flanks and a temporary return to coarse clastic shallow-marine sedimentation as observed in outcrops of the Saih Hatat area (Le Métour et al., 1986). Erosion of the Hasirah Formation due to the Late Ordovician glaciation appears not to be significant in the Ghaba Salt Basin (Figure 9), although incised channels have been observed on seismic sections. In South Oman along the East Flank, evidence for the deposition of the Safiq Group and the impact of the Late Ordovician glaciation is destroyed by Mid to Late Carboniferous Hercynian uplift and erosion, followed by the onset of rifting. During the latest Carboniferous-Early Permian, the East Flank was part of a pronounced rift shoulder of the Batain Basin, a proto-Indian Ocean Basin (Belushi et al., 1996; Immenhauser et al., 1998). Rift-shoulder uplift was accompanied by Late Carboniferous to Early Permian highland glaciation, again removing major amounts of the Paleozoic.

In North Oman, the Ghudun depocentre coincides approximately with the depocentre of the Early Cambrian Ghaba Salt Basin. The northern extension of this Ghudun Basin can be traced into the southeastern Oman Mountains. Here, the Early to Late Ordovician Amdeh Formation (Figure 12) was deposited in a shallow-marine to an outer-shelf environment, unconformably overlying the Lower Huqf Hiyam Formation which is equivalent to the Khufai Formation (Lovelock et al., 1981; Le Métour, 1988). The basal, 80-240 m thick Lower Siltstone Member Am1 of the Amdeh Formation is not dated, but is thought to be late Early Ordovician (Arenigian) in age. This interpretation is based on a similar geologic setting compared to the Ghaba Basin. The conglomerates (Figure 12) in the top part of the Lower Siltstone Member Am1 (Villey et al., 1986) indicate a phase of rift-shoulder uplift following initial extension and resulting in the formation of tilted fault blocks. The conglomerates, including carbonates of the Late Precambrian Hiyam Formation, were derived predominantly from the west. Evidence of rift-shoulder activity is also based on the architecture of the unconformably overlying Lower Quartzite Member Am2, wedging out completely to the northwest (Villey et al., 1986). In Saih Hatat, this event is the signal for increased subsidence resulting in the deposition of about 3,000 m of Mid to Late Ordovician clastic sediments (Figure 12). The Lower Quartzite Member Am2 above the unconformity represents the coarse transgressive systems tract before the major transgression of the

Llanvirnian Middle Shale Member Am3. A second major marine transgression took place in the Upper Siltstone Member Am5 during Early Caradocian times. Its timing is based on the presence of possible Caradocian trilobites (Le Métour, 1988) and a regional correlation with the Hasirah Formation in the Ghaba Salt Basin. The coarse conglomerates within the marine siltstones at the top of the Upper Siltstone Member Am5 could represent dropstones related to the advancing ice sheets and the accompanying eustatic sea-level fall.

In contrast to the Ghaba Basin where extension is primarily an Arenigian phenomenon, major extension in Saih Hatat took place predominantly in Mid Ordovician times, 10 to 15 million years (m.y.) later. Diachronous development of individual sub-basins belonging to one major extensional system is well known; for example, the diachronous Late Jurassic basin formation in the North Sea. Often, the boundaries between the (sub) basins are enigmatic structural zones consisting of transfer faults and accommodation zones (Chapin and Cather, 1994). The boundary between the Ghaba and Saih Hatat basins is located below the Hawasina-Semail Nappes without reliable seismic coverage and has thus not been studied.

Evidence for a Mid Ordovician extensional pulse was discussed by Loosveld et al. (1996). These authors noted that although the end of the sag phase should have been reached at the latest some 50 m.y. after Huqf rifting, extension resumed again with the deposition of shallow- to deep-marine sediments of the Safiq Group. This phase of extension was poorly understood due to lack of well data and poor seismic imaging. The recent drilling campaign for Haima gas yielded new well data, and seismic imaging improved considerably. This phase of late Early to Mid Ordovician extension, following a period of uplift and erosion, was also observed in other parts of the Arabian Peninsula (Stöcklin, 1968; Connally and Wiltse, 1996; Jones et al., 1996; Jones and Stump, 1999), and was hitherto poorly understood. The observation of potassium-rich mafic dikes, related to this renewed pulse of extension, enables a more complete picture of the geology to be obtained.

Petrology and Setting of the Mafic Rocks

Several outcrops of mafic intrusives occur in the Huqf area (Figure 13) (Dubreuilh et al., 1992). These rocks were described as alkali-olivine basalts or trachytes and in some cases were erroneously correlated with Tertiary basalts from Al-Ashkarah further north (Figure 1), which were intruded at the onset of the Gulf of Aden rifting some 38 to 40 Ma (Bott et al., 1992; Worthing unpublished 39Ar/40Ar data). During the course of fieldwork for this paper, three of the Huqf occurrences were sampled: Qalilat Suwaidi, Qarn Aswad, and a small body west of Sirab (Figure 13). The Sirab rocks are clinopyroxene and olivine-phyric basanites that petrographically and geochemically closely resemble the Tertiary basalts of Al-Ashkarah and are thus correlated with them. They will not be considered further here. Rocks from the other two outcrops are petrographically and geochemically distinct and are discussed below.

Qalilat Suwaidi (Little Black Hill)

The intrusive rocks cropout on three contiguous knoll-shaped hills about 400 m south of the Duqm to Sinaw road. They were mapped as basalt dikes by Dubreuilh et al. (1992) at locality 20°27.5'N, 57°52.0'E (Figure 13). The highest knoll rises about 40 m above the surrounding plain. The sub-circular shape of the knolls suggests a plug or plugs. However, outcrops on the northerly knoll suggest that the mafic rocks are underlain by purple shales of the late Precambrian Shuram Formation. Contact dips appear to be shallow and radial from the center of the outcrop and thus it is possible that the body was intruded as either a sill, dike or laccolith. The fine-grained, dense, black mafic rock is strongly fractured into sub-horizontal joints on a three to five centimeter scale. The joints dip outwards parallel to the contour of the slope on one knoll and are interpreted as cooling fractures. The shallow intrusions contain xenoliths of crustal material such as quartzite, red granite, and stromatolitic dolomite and, more rarely, euhedral crystals of a pink K-feldspar up to 2 centimeters long. No ultramafic xenoliths were seen. In coarser facies, phlogopite phenocrysts (phlogopite: the magnesium-rich variety of biotite) are visible to the naked eye, though generally the rock is aphanitic in hand specimen.

In thin section, the rocks contain abundant microphenocrysts of lath-like, red-brown sieved and castellated phlogopite, euhedral laths of colorless clinopyroxene, and sparse olivine microphenocrysts partly or completely altered to a pale-colored serpentine. The groundmass contains lath-like K-feldspar, abundant granular ilmenite and colorless glass. The rock also contains euhedral to rounded pseudomorphed phenocrysts of brown amphibole locally in reaction relationship with clinopyroxene. A more comprehensive petrographic description is given in the Appendix.

Qarn Aswad (Black Horn)

Here dense black mafic rocks form a roughly elliptical ridge some 300 m by 150 m (20°32.1'N., 57°53.8' E., Figure 13). The core of the Qarn Aswad structure is occupied by low ground in which there is no exposure. In the inner walls of the ridge, at the exposed lower contact, it is clear that the mafic intrusion is discordant to the bedding in the enclosing purple shales of the Shuram Formation. There is also a narrow zone of contact metamorphism. Where exposed, dips of the contact are shallow to moderate and appear to be radial to the core of the structure. The intrusion thus has an umbrella-shaped form similar to that seen at Qalilat Suwaidi. This distinctive form may have been caused by local updoming of competent horizons within the Shuram Formation by a rising plug of mafic magma accompanied by lateral intrusion into the updomed sediments. Petrographically the rocks are similar to those from Qalilat Suwaidi except that pseudomorphed olivine is more abundant, though still comprising only about 10% by volume of the rock and phlogopite is present in only minor amounts.

The long axis of the elliptically shaped Qarn Aswad outcrop is aligned northwest to southeast indicating the local principal horizontal stress direction during intrusion. Based on Landsat satellite maps, the Qarn Zaza-Qarn Aswad-Qalilat Suwaidi intrusions are aligned in a zone trending north-northeast to south-southwest (Figure 13). This direction is considered to represent the regional principal horizontal stress direction during the Mid to Late Ordovician, parallel to the Ghaba Basin axis (Figure 4).

Geochemical Analyses and Mineralogical Classification of the Mafic Rocks

Whole rock, trace element and rare earth element (REE) analyses were determined at the University of Copenhagen, Denmark, and the University of Tasmania, Australia. One sample (HG 10) was run for whole rock and trace elements at both laboratories with excellent agreement. Analytical methods are given in the Appendix and representative analyses are presented in Table 1. The mafic rocks are undersaturated (SiO2: 44.42 to 46.67 weight percentage [wt%]) and alkaline, with total alkalis ranging from 6.65 to 7.60 wt%. K2O values range from 3.63 to 5.02 wt% and Na2O from 2.51 to 3.45 wt%. K2O/Na2O ratios range between 1.31 and 1.94 (mean 1.51). They are nepheline normative (range 4.51 to 9.10%) and strongly enriched in large ion lithophile elements (LIL) such as Sr, K and Rb. MgO ranges from 5.12 to 6.31 wt% and the Mg number (100 (Mg/Mg + Fe)), from 48.73 to 50.27. Cr and Ni range from 55 to 103 parts per million (ppm) and 68 to 92 ppm, respectively. The high total alkalis and high K2O/Na2O ratios indicate that the rocks are potassic (e.g. Peccerillo, 1992).

Geochemically, the rocks from Qalilat Suwaidi and Qarn Aswad are almost identical, however the mineralogical differences require some comment. The presence of abundant phlogopite microphenocrysts at Qalilat Suwaidi suggests that the rocks may be lamprophyres (Carmichael et al., 1974). Streckeisen (1979) and Rock (1987) defined lamprophyres as hypabyssal, melanocratic igneous rocks with porphyritic textures carrying only mafic phenocrysts, essentially phlogopite and/or amphibole with minor olivine. According to the classifications of these two authors and Muller and Groves (1995), the Qalilat Suwaidi mafic rocks would be classified as minettes. It is interesting to note that Platel et al. (1992, 1993) have described as “semi-lamprophyric” a suite of northwest to southeast trending dikes intruded into the basement metamorphics in the Marbat area east of Salalah (Figure 1). These rocks contain phlogopite and/or hornblende as the main mafic minerals. They are also potassium-rich with a K2O/Na2O ratio of 1.71. The Qarn Aswad rocks plot in the basanite-tephrite field of a TAS (Total Alkalis Silica) diagram. The rarity of phlogopite and the presence of normative olivine at <10% indicates that they should be classified as potassic tephrites.

Geochronology

Potassium-Argon (K/Ar) and Argon-Argon (Ar/Ar) age determinations on the Huqf intrusive rocks are summarised in Table 2. Although the accuracy of K/Ar dates is generally regarded as suspect, the data set as whole points to a tectonomagmatic event in the Middle to Late Ordovician. The northwest to southeast trending “semi-lamprophyric” dikes (Platel et al., 1993) in the Marbat area (438 ± 22 Ma) are also included in Table 2 although sericitization as well as resetting of unstable minerals was noted. Based on overall setting, they are considered not to belong to the same Mid to Late Ordovician tectonomagmatic event as the Huqf suite. The Marbat mafic dikes were clearly eroded before the deposition of the Marbat Sandstone Formation which is regionally correlated with the Late Precambrian Abu Mahara Formation (personal communication of one of the reviewers). The Ordovician age may have been imposed on the Marbat dikes by resetting during the heat pulse associated with the Mid Ordovician extensional event described in this paper.

In the present study, one sample (HG 10, wt% K2O = 3.93) was selected for 39Ar/40Ar dating at Oregon State University(OSU), Corvallis, U.S.A. The sample was selected because of its freshness and relative absence of xenoliths. Rock chips 0.5-1 millimeter in size were irradiated for 6 to 8 hours in the core of the OSU TRIGA reactor and conversion from 39K to 39Ar by neutron capture was monitored with hornblende standard Mmhb-1 (Samson and Alexander, 1987). Six incremental heating steps were conducted. The isotopic composition of Ar released at each step was measured using an AEI-MSIOS mass spectrometer at OSU. Plateau age estimates are based on consecutive incremental heating steps that overlap within analytical uncertainties and cumulatively include at least 50% of the 39Ar released from the sample.

The whole-rock spectrum is presented in Figure 14 and indicates an age of 461 ± 2.4 Ma. According to the time-scale of Gradstein and Ogg (1996), the event took place during the Mid Ordovician (Llandeilian) (Figure 14). The excellent plateau suggests that the age is reliable and there is no clear evidence of a younger thermal event. The date is thus consistent with the tectonomagmatic event suggested by the data in Table 2. The low temperature step possibly represents degassing of the K-feldspar which is susceptible to Ar loss, particularly from crystal rims (closure temperature Tc = 125° to 350°C). The plateau age probably represents degassing of the biotite (Tc = 260° to 350°C) and thus indicates the point at which the rock cooled below Tc for biotite. The high temperature step possibly represents degassing of the amphibole (Tc = 450° to 525°C), (Spear 1993).

DISCUSSION

The Origin of Potassic Magmas in the Rio Grande Rift

A worldwide survey by Thompson et al. (1989) suggested that minettes are closely associated in space and time with heating and or thinning of the sub-continental mantle. This conclusion is partly based on their study of minettes occurring within a zone of intracontinental extension in Northern Colorado, U.S.A. This area is part of the northerly extension of the Oligocene-Miocene Rio Grande Rift and studies of the same area by Alibert et al. (1986) and Gibson et al. (1993) also record potassic rocks including minettes. The evolution of the Rio Grande Rift took place in two distinct phases; an early phase (32 to 27 Ma) characterised by the formation of broad, shallow basins and low-angle faulting. During this phase, crustal extension took place in a broad zone up to 170 km across. In the second phase, which took place in the Miocene (16 to 10 Ma), the zone of extension narrowed to about 50 km and faulting was high-angle resulting in the formation of narrow grabens that define the present morphology of the rift zone (Chapin and Cather, 1994). Gravity and teleseismic studies have shown that the width of thinned lithospheric mantle underlying the Rift is about 750 km and this distance corresponds to the surface expression of magmatism.

Gibson et al. (1993) undertook a 670 km geochemical traverse across the axis of the Rio Grande Rift, sampling volcanics from the flanks, shoulders and rift trough. Their work which includes K/Ar dates, showed that the earlier magmatic activity on the rift flanks and shoulders is dominated by potassic mafic rocks including minettes. In the trough on the other hand, the volcanics are tholeiitic and alkaline in character with OIB-like affinities (OIB = oceanic-island basalt) in both phases of activity. Potassic magmas are absent from the trough. They concluded that the potassic magmas were derived from a metasomatised layer in the lithospheric mantle that melted by decompression during the early stages of continental extension (McKenzie, 1989).

Comparison between Interior Oman and the Rio Grande Rift

The intracontinental Rio Grande Rift is of similar size to the rifted basin described in this paper and its geophysical setting, tectonomagmatic evolution, and igneous geochemistry are well-documented (e.g. Keller and Cather, 1994; Rogers et al., 1982; Alibert et al., 1986; Thompson et al., 1989; Gibson et al., 1993). Thus, in the following section the geochemical characteristics of potassic rocks from both areas will be compared with the aim of drawing possible conclusions about the tectonomagmatic setting of the Huqf area during the Middle Ordovician.

In order to compare the potassic rocks from both areas, analyses were abstracted from Rogers et al. (1982), Alibert et al. (1986), Thompson et al. (1989), and Gibson et al. (1993) and plotted on major element and chondrite normalized trace element and REE variation diagrams. As Gibson et al. (1993) do not clearly identify the rock types of all their samples, all analyses with K2O/Na2O>1 were plotted including two rocks specifically identified as minettes. Analyses from the other papers are identified by the authors as minettes. Selected examples of these diagrams are shown in Figures 15a to 15f and 16a and 16b. The diagrams show that the geochemical signatures of the Huqf intrusions are broadly similar to the minettes and other potassic rocks from the Rio Grande Rift, particularly the specimens from Gibson et al. (1993). However, there are some differences, particularly in the major element plots (Figures 15a to 15f). For example, compared with the Rio Grande suite, the Huqf mafic rocks are very depleted in MgO and slightly depleted in CaO (Figures 15e and 15f). Although many of the analyses lie close to the Rio Grande Trend, they plot at the low silica end of the diagrams suggesting that they are primitive. However the low MgO (Figure 15e) which is particularly anomalous does not support this. This will be discussed below. On the other hand, chondrite-normalised trace element and REE signatures (Figures 16a and 16b) are very similar to those from the Rio Grande Rift, the latter showing strong, light REE enrichment. The mild Nb enrichment shown in Figure 16b may be explained by the high modal content of ilmenite as the Nb is probably partitioned in this mineral.

The origin of lamprophyres and other potassic rocks is controversial. In igneous provinces where they occur in abundance, their composition may vary greatly. This compositional variability is explained as being the result of a variety of processes, such as melting of different mantle sources, magma mixing, assimilation and fractional crystallisation (e.g. Rock, 1987). Without further isotopic studies it is impossible to draw firm conclusions about the origin of the Huqf rocks. However, their broad geochemical similarity to those from the Rio Grande Rift suggests that they have a comparable origin. It is possible that some of the differences outlined above may be explained by fractionation. For example, the large phenocrysts of amphibole and clinopyroxene in a magmatic reaction relationship may represent a residual CaMg-rich phase that has largely fractionated out of the magma, thereby depleting these elements. Similarly, the low MgO may also be due to olivine fractionation. Thus, in summary, the Huqf rocks have a similar geochemical signature to the potassic rocks from the Rio Grande Rift and differences may be explained in terms of simple fractionation models, but the operation of other petrogenetic processes cannot be discounted.

Potassic igneous rocks occur in a wide variety of geological settings including continental arcs, post-collisional arcs, initial and late oceanic arcs and within-plate settings (Muller and Groves, 1995). Since the standard basalt tectonic discriminant diagrams are misleading for this suite, Muller and Groves (1995) have proposed a set of diagrams that discriminate between rocks from these different tectonic settings. Using their Zr/Al2O3 against TiO2/Al2O3, TiO2 against Al2O3 and Y against Zr diagrams, the

Huqf mafic intrusives clearly plot in the Within-Plate fields, confirming intrusion within an intracontinental setting (Figures 17a, b, c). Potassic rocks including minettes from the Rio Grande Rift are also plotted on this diagram.

The magmatic activity recorded in the Huqf appears to be very minor when compared with the Rio Grande Rift. However, more volcanics may still be buried or have been stripped off during subsequent erosion that may have taken place during Mid to Late Carboniferous times resulting in the removal of up to 2 km of sediments in the Huqf (Visser, 1991). It is probable that lithospheric stretching in Oman did not progress beyond the initial stages represented in the Rio Grande Rift by the early phase of magmatic activity that resulted in the intrusion of the potassic rocks on the rift flanks and shoulders (Gibson et al., 1993). This is also consistent with the style of Mid Ordovician rifting seen in central Oman that appears to have involved formation of a broad, shallow basin not a narrow, steep graben.

Thus in summary, minettes and other potassic rocks from the Rio Grande Rift were intruded into the flanks and shoulders of an intracontinental rift during an early phase of broad extension involving the crust and lithospheric mantle. Potassic rocks from the Huqf are geochemically similar and their geographical location suggests that they were intruded into the eastern rift shoulder of the intracontinental Ghaba Salt Basin during a pulse of extension in the Middle Ordovician. These observations are consistent with the known continental to deep-marine environment of the Mid to Late Ordovician sediments in Oman that lack any plate margin or island arc affinity. This is the first direct evidence based on shallow intrusions, that active rift extension took place during deposition of the Upper Ghudun Formation and the Safiq Group. A transtensional tectonic setting explains the faulting of the Ghudun Formation across the basin and the onlap of the unconformably overlying Safiq Group as observed on seismic sections. Tilted Ghudun fault blocks associated with the extensional event provide first-class reservoirs and are efficiently sealed by draping lower shoreface shales of the Saih Nihayda Formation.

Early Paleozoic Rifting on the Arabian Peninsula

Following the amalgamation of Gondwana in the protracted Pan-African Orogeny, crustal separation within the Arabian Peninsula had already started during the waning stages of the continent-to-continent collision. Rifting in the Late Precambrian-Early Cambrian can be considered as the first break-up attempt of the combined rift basins on the Arabian Peninsula, including Iran, and Greater India which have been defined as Proto-Tethys (Talbot and Alavi, 1996). The next phase of Gondwana break-up started during the Mid to Late Cambrian and led to the development of the Rheic Ocean, separating Western Gondwana (North Africa) from Baltica (Paris and Robardet, 1990). The Rheic Ocean was the continuation of Paleo-Tethys to the west. During Cambrian-Ordovician times, Northern Gondwana terrain slivers separated from Gondwana (e.g. Ziegler, 1988) and terrain slivers drifted away from Iran towards the north during the Ordovician (Stampfli et al., 1991). Such a terrain sliver is exposed in the Lesser Caucasus (Sengör, 1990). Rifting and drifting was accompanied by igneous activity. In northern Iran, the earliest phase of mafic volcanism started in the Ghelli region, some 350 km north of Tabas, during the Early Ordovician. Volcanic activity continued in the Talesh Mountains, southwest of the Caspian Sea, during the Mid Ordovician. It was followed by major basalt outpouring during the Late Ordovician to Early Silurian, especially in the northeastern Alborz Range and on the Lut Block (Berberian and King, 1981) (Figure 18).

Recent deep exploration drilling in Central Saudi Arabia also shows a Mid to Late Ordovician phase of extension (e.g. Jones et al., 1996). Following a phase of compression sub-parallel to the Huqf axis in the Early Ordovician, as expressed by uplift of the Central Arabian and Qatar Arches, subsequent extension was characterized by broad grabens filled with marine Mid to Late Ordovician siliciclastics belonging to the Qasim Group (Connally and Wiltse, 1995, 1996). So far, Ordovician volcanism has not been reported in Saudi Arabia. In this area, Mid Ordovician to Early Silurian sedimentation appears to have taken place in a rim basin related to rifting to the north in Iran and to the east in Oman (for the concept of rim basin or intracratonic sag, see Stampfli et al., 1991). The eastern flank of this rim basin is formed by the Ghudun-Khasfah and Mabrouk-Makarem Highs. The latter is the western rift shoulder of the Ghaba Salt Basin. This basin is considered to be an aulacogen, one arm of a triple-rift junction that failed to drift to an ocean. The other two rift arms successfully developed as part of the northwest to southeast trending Paleo-Tethys and were located to the north of the Alborz Range (Figure 18). Extensive mafic volcanism (Berberian and King, 1981) is observed in the area of the triple-rift junction (Figure 18).

Although the concept of a failed rift is conjectural, the northern extension of the Ghaba Salt Basin has been shown to be located in Saih Hatat, within the southeastern Oman Mountains (Le Métour, 1988), (Figure 1). The extension of the Ordovician basin to the north of Oman is not fully constrained and reflects the unresolved paleo-plate reconstruction of platelets such as the Central Iranian and Helmund Blocks (e.g. Scotese and McKerrow, 1990). The Central Iranian Block (CIB) consists of the Tabas Graben flanked to the east by the Lut Block and to the west by the Yazd Block or Nain Horst (CIB as defined by Sengör, 1990). Based on the presence of Ordovician carbonates on the Central Iranian Block (Stöcklin, 1968; Berberian and King, 1981), Le Métour (1988) suggested that this platelet must have been located further north, closer to the equator, when compared to non-carbonate conditions in Oman further south. This proposed location of the CIB is consistent with the extreme thickening of the Ordovician-Silurian sediments in the Tabas Graben (Figure 18) which was probably the extension of Oman’s aulacogen to the north. In Iran, the thickness of the Infracambrian to Middle Triassic is, with the above exception, remarkably constant (Stöcklin, 1968). The present paleogeography of the CIB agrees well with the locations proposed by Soffel and Förster (1984) and Le Métour (1988). Following Ordovician times, left-lateral strike-slip faulting during the Late Triassic-Early Jurassic Cimmerian event resulted in the displacement and counter-clockwise rotation of the Central Iranian Block with respect to Oman (Sengör, 1990).

The contrast in amount of volcanics on opposite sides of the Tabas Graben in Iran and the Ghaba Salt Basin (Figure 18) can be explained by the asymmetric rift configuration of a simple shear model (e.g. Stampfli et al., 1991; Talbot and Alavi, 1996). The upper plate above the asthenospheric upwelling is suggested to have been located in the Huqf and Lut areas, where significant amounts of volcanics are still present notwithstanding erosion (Figure 19). The paucity of tuffs both over Central Iran and within the Ghaba Basin may be partly due to westerly winds and partly due to subsequent erosion over Central Iran and Huqf as well as difficulties in detecting tuff mixed with other fine-grained lithologies at great depth in the Ghaba Salt Basin. In contrast, the lower plate, characterised by a lack of volcanics, contains the Ghaba Rift Basin and its western flank. The morphology of the western and eastern rift shoulders is consistent with the above model. The eastern flank, formed by the Huqf High, is morphologically expressed by a flexure as observed on seismic data, in contrast to the western flank where the Ghudun-Khasfah and Mabrouk-Makarem Highs are developed as a sharp rift shoulder representing the initial break-away (Figure 19).

The late Early to early Late Ordovician rift basins have similar locations and trends compared to those developed in the Late Permian-Early Triassic, when Neo-Tethys was taking shape to the northwest of Oman (Stampfli et al., 1991) and the Batain Basin, a Proto-Indian basin, to the east of Oman. It was not until the Late Jurassic that rifting in the southern rift arm progressed into drifting with the formation of oceanic crust (Immenhauser et al., 1998).

CONCLUSIONS

Rifting of the Arabian Peninsula during the late Early to early Late Ordovician has been suspected for some time based on an increase in rate of deposition in Central Oman, and a change in facies from fluvial to lower shoreface. Efforts to more clearly document this extensional phase however, were hampered by lack of well data and poor seismic imaging. Further north in Saih Hatat, rift-shoulder uplift and deposition of some 3 km of sediments during the Mid to early Late Ordovician was considered to be further evidence for such an extensional phase.

In this paper we have assembled geochemical and geochronological evidence based on potassic mafic dykes from the Huqf which suggest that they were intruded into the shoulder of an intracontinental rift during the Mid Ordovician. The igneous activity together with stratigraphic, sedimentological and tectonic data have produced a compelling picture of rift-related extension and rift-shoulder uplift in the late Early to early Late Ordovician. The diachronous development of the Ghaba Salt and Saih Hatat Basins indicates an extended time of rifting of not less than 15 million years. Together with the Tabas Graben in Iran, these two basins can be considered as sub-basins of a failed rift arm. This rift phase represents an early stage in the break-up between Africa/Arabia and India. Thus, an additional phase of tectonic activity prior to the Late Carboniferous to Early Cretaceous extensional and Alpine compressional events has been documented. Structures formed during such phases are important for the trapping of hydrocarbons. Therefore, the tilted fault blocks of the sand-prone Ghudun Formation sealed by onlapping marine shales of the Safiq Group are promising new targets for hydrocarbon exploration.

ACKNOWLEDGEMENTS

The authors wish to thank the Ministry of Oil and Gas, Sultan Qaboos University and Petroleum Development Oman (PDO) for their permission to publish this paper. Mike Worthing is grateful to Oregon State University, Corvallis, for supplying age data of the mafic dike and to Kent Brooks of the University of Copenhagen and Phil Robinson of the University of Tasmania for facilitating the X-ray fluorescence and REE analyses. Mike Worthing also thanks Ron Berry of the University of Tasmania, Australia for discussions on the geochronology and Samir Hanna who proposed the project on which this paper is based. The PDO authors would like to thank H. Droste, T. Faulkner, E. Hoogerduin Strating, A. McCoss, Ide van der Molen, M. Naylor, P. Richard and Paul Senycia for stimulating discussions and contributions. The comments of two referees have been valuable in improving this paper.

APPENDIX

Petrographic Descriptions

In order to verify petrographic mineral identifications, particularly of the finer-grained phases, qualitative EDS spectra were obtained using a LINK Isis 3 attached to a Jeol JSM 840A scanning electron microscope at Sultan Qaboos University.

Qalilat Suwaidi

In thin section the mafic rocks have a fresh appearance with a hypocrystalline and micro-porphyritic texture. They contain abundant microphenocrysts of lath-like red-brown to almost colourless sieved and castellated phlogopite ranging in length between 0.5 and 1.5 mm. The mineral is locally altered to a pale-colored chlorite. Euhedral laths of a colorless clinopyroxene about 0.1 mm long are also abundant. They have low birefringence and strongly oblique extinction and some hourglass structure. There is a faint flow structure in the rock defined by alignment of the clinopyroxene laths. The rock also contains sparse olivine phenocrysts up to 0.5 mm in diameter which are now partly or completely pseudomorphed with a pale-colored serpentine showing typical meshing after olivine. Rare phenocrysts of clinopyroxene also occur which enclose phlogopite laths. Coarser facies contain K-feldspar with Carlsbad twinning that poikilitically enclose groundmass minerals. In finer grained facies the groundmass minerals are difficult to resolve but include lath-like K-feldspar, phlogopite, abundant granular ilmenite and a colorless glass. Rounded xenoliths of strained quartz up to 0.25 mm are present. These may be derived from the underlying gneissic basement or the Abu Mahara Formation.

Another notable feature of the rock is euhedral to rounded pseudomorphed phenocrysts composed mainly of aligned blebs of ilmenite and a colourless mineral. The largest is 2.5 mm across. Isolated grains of brown K-Ti-rich amphibole within the phenocryst are in optical continuity suggesting that the area was once occupied by a single large K-Ti-rich amphibole crystal. In one case, this crystal is in contact with a grain of pale-colored clinopyroxene that appears to be in equilibrium with the amphibole and is interpreted as a primary inclusion inside the amphibole probably in a magmatic reaction relationship. The remaining area is filled by blebs of granular ilmenite and primary phlogopite laths plus a pale-colored mineral that possibly represents the breakdown of the primary amphibole to a low-temperature amphibole plus ilmenite.

Qarn Aswad

These rocks are similar to those from Qalilat Suwaidi except that pseudomorphed olivine is more abundant, though still comprising only about 10% of the mode, and phlogopite is present in only minor amounts.

Geochemical Analytical Methods

At Copenhagen University, an automatic PW 1606 x-ray fluorescence machine was used. Samples were first ignited to 950oC and volatile loss measured and corrected for oxidation (Fe2+ to Fe3+). Ignited samples were then mixed with sodium tetraborate and fused. Sodium was determined by atomic absorption after solution in HF. FeO was determined on a glass disc.

At the University of Tasmania, samples were run on a PW 1480 machine. Samples were heated to 1,000oC overnight to determine volatile loss. Ignited samples were then mixed with Norrish flux and lithium nitrate and fused at 1,050oC for ten minutes and poured into Pt moulds. Na was determined on the fusion disc and Fe was determined as Fe2O3. One sample (HG10) was run at both laboratories with excellent agreement. Pressed pellets for trace elements analysis were prepared at 32 tonnes cm-2 using 6 gms of powder and PVA binder. The pellets were backed with boric acid.

REE determinations were done on an ELEMENT HR-ICP-MS Funnigan MAT machine at the University of Tasmania. Deionised water was used for rinsing and solution preparation. Samples were digested in HF and BDH Analar ® HNO3 purified by double distillation. One hundred milligrams of powdered sample were weighed into Sarvillex vials and moistened with ultrapure water. Two milliliters of HF and 0.5 ml of HNO3 were added and the sealed vial placed on a hotplate for 48 hours at 130oC. The vials were removed twice in every 24 hours, cooled and placed in an ultrasound bath for 2 hours. After 48 hours the vials were opened and evaporated to incipient dryness and the residuum dissolved in 2 ml HNO3 followed by 10 to 20 ml of water. The resulting solution was transferred to a polycarbonate container and diluted to 100 ml. An Indium internal standard was also added to give a final concentration of 10 mg gm-1. At least two reagent blanks for each sample were prepared.

ABOUT THE AUTHORS

W. Heiko Oterdoom joined Petroleum Development Oman in 1995 as a Senior Exploration Geologist in diverse functions ranging from operational stratigraphy and geological studies to frontier exploration. Heiko has 18 years exploration experience as Regional Geologist with Shell in oil shales, the North Sea – Norwegian Sea and the Far East. Heiko was educated at Berkeley and Zürich and holds a PhD in Petrography from the Federal Institute of Technology (ETH), Zürich. He is a licensed guide of the Dutch tidal flats.

Mike A. Worthing is Head of the Earth Science Department at Sultan Qaboos University, Oman where he has worked since 1992. He completed a PhD in Structural Geology and Metamorphic Petrology at London University in 1971 based on fieldwork in Arctic Norway. Since then Mike has worked as an academic in the UK, Uganda, Papua New Guinea, Australia and Oman. His main research interests are mineralogy and the petrology and geochemistry of mafic igneous and metamorphic rocks. He enjoys sports, wadi-bashing and violin music.

Mark Partington is a Senior Seismic Interpreter in the Exploration Department at Petroleum Development Oman. Mark’s main responsibilities lie in the field of prospect evaluation, applied stratigraphy and integration with seismic data. Before joining PDO in 1996, he was employed as Stratigrapher by Shell International and Shell Research in The Netherlands. He has a MSc from University College London and a PhD from Aberdeen.