INTRODUCTION

The Chitarwata and overlying Vihowa Formations span an important interval of tectonic history recording the character of Himalayan foreland basin sedimentation during the obduction of the Indian and Asian plates in western Pakistan, with regional transition from marginal marine to fluvial environments, and the rise of the high Himalayas. Previous investigation of the lithostratigraphy in the foothills of the Sulaiman Range at Dalana in Zinda Pir Dome (Downing et al. 1993) indicates the Chitarwata Formation is dominated by coastal paleoenvironments. In ascending order, the three units are associated with estuary, strandplain and tidal flat environments. The Vihowa Formation represents the prominent development of fluvio-deltaic sedimentation related to the southward progradation of the Indus River system.

On the basis of paleomagnetic data, the age span of the Chitarwata Formation in the Zinda Pir Dome has been evaluated as early Miocene from 22.3 –18.6 Ma (Friedman et al. 1992) more recently from 22–17.4 Ma (Lindsay and Downs 2000) and currently Oligocene at its base and earliest Miocene at the contact with the Vihowa Formation (Lindsay et al. this issue). In this age assessment, the Chitarwata Formation temporally overlaps with the Murree Formation exposed in northwest regions of the Potwar Plateau (Abassi and Friend 1989; Najman et al. 2003), the uppermost Shaigalu member of the Khojak Formation in the Katawaz Basin (Qayyum et al. 2001) as well as the Choksti Formation in the Indus suture zone (Clift et al. 2001a) and the Kasauli Formation in the Indian Himalayan foreland basin (Najman and Garzanti 2000).

To the southwest of Zinda Pir Dome in the Bugti Hills, the Chitarwata Formation also has been assigned a basal Bugti Member (Raza and Meyer 1984). The Bugti member has yielded possible early Oligocene vertebrate faunas (Welcomme and Ginsburg 1997, Marivaux et al. 1999; Welcomme et al. 2001), although an early Miocene radiometric date at 22.6 + 2.9 Ma (Tabbutt et al. 1997) purportedly in its lower portion conflicts with this biostratigraphic assessment. A chronostratigraphic record has not yet been established for Bugti Hills so no direct comparison to Zinda Pir Dome sections can be made in this regard. Hence, the age correlation between the two regions relies on the biostratigraphy and homotaxis of mammals.

There has been an ostensible incongruence between the biostratigraphy and corresponding age assessments for lower parts of the Chitarwata Formation in Zinda Pir Dome and Bugti Hills (early Miocene versus ?early Oligocene). This disparity has been attributed to a time-transgressive facies in the Chitarwata Formation with older deposits preserved to the south (Raza et al. 2002) or an undetected but notable hiatus calling into question the previous chronostratigraphic assessment of the Chitarwata Formation at Zinda Pir Dome (see Welcomme et al. 2001). In this latter case, biostratigraphic succession may be interrupted by a significant hiatus between the Oligocene vertebrate faunas from the lower member (e.g., locality Z108 described by Flynn and Cheema 1994) and later faunas in the upper member. The latest biostratigraphic and chronostratigraphic findings by Lindsay et al. (this issue) suggest that the faunas from the classic Dera Bugti localities and lower member of the Chitarwata Formation at Zinda Pir Dome are probably both Oligocene and closer in age than previously recognized. Yet there are still consequential differences between the two faunas. Among these differences at Zinda Pir Dome are the absence of primitive muroids, a paucity of remains of large indricothere rhinos, and no demonstrated stratigraphic overlap of large indricotheres with early proboscideans (Lindsay et al. this issue).

An explanation for the observed biostratigraphic differences in each area is suggested by the consequential lithostratigraphic, taphonomic, and paleoenvironmental differences between the Chitarwata Formation in the Zinda Pir Dome and Bugti Hills. Most notably, the coastal marine facies at Bugti Hills are thin, only 30 m thick, and thereafter directly evolve into a fluvio-deltaic floodplain with coal beds and numerous paleosol horizons interpreted as tropical to subtropical densely forested environments (Welcomme et al. 2001, Antoine et al. 2003). In the Zinda Pir area coastal facies are over 300 m thick and evolve into a fluvio-deltaic system much higher in the section with the deposition of the Vihowa Formation (Downing et al. 1993). Vertebrates from Zinda Pir area are sparse through its marginal marine facies while the Bugti Member at Bugti Hills has for the past 150 years produced one of the richest fossil vertebrate faunas in Asia. While a notably attenuated coastal facies of the Chitarwata Formation at Bugti Hills is a possibility it is also soundly viable that the Bugti Member is geologically distinct and a direct lithostratigraphic correlation to the Chitarwata Formation is unwarranted. If so, the Bugti Member could be related to an entirely different depositional and drainage system and assignable to a different formation than the Chitarwata Formation.

Paleocurrents and Paleodrainage in the Western Himalayan Foreland

In order to characterize Himalayan foreland basin sedimentation and paleogeography in relation to the character and sequence of India-Asia tectonics, investigations of pre- and post-collisional paleocurrents and paleodrainage have been conducted in the Indian Himalaya, Potwar Plateau, Kohat Plateau, Katawaz Basin, Sulaiman Range (Zinda Pir Dome), and Balochistan (Figure 1). Pre-collisional paleocurrent data is available from the late Cretaceous Pab Sandstone and supports northwest flow off of the Indian plate (Waheed and Wells 1990). Beginning in the early Eocene, paleocurrent indicators reflect post-collisional paleodrainage associated with the development of the Himalayan foreland and obduction along the western edge of the Indian plate (summarized in Table 1).

An early manifestation of the paleodrainage shift due to obduction and shelf reversal is recorded in the broadly distributed and marginal marine early Eocene Ghazij Formation in western Pakistan, where southeastward paleoflow from a northwestern highland is dominant (Waheed and Wells 1990; Clyde et al. 2003). A similar pattern is observed in the early Eocene Ganguri Sandstone of the Kohat Basin (Wells 1984; Pivnik and Wells 1996). Conversely in the Katawaz Basin, the early Eocene Khojak Formation indicates the Indus River system delta-fan complex is initiated with longitudinal and westward flow into the Katawaz remnant ocean (Qayyum et al. 1996; Qayyum et al. 2001).

Paleocurrent information in the western Himalayan foreland for the middle to late Eocene and the early Oligocene is sparse as this interval is marked by a return to open marine conditions represented by the Kirthar Formation in the Sulaiman Range and the Kohat Formation in the Kohat Plateau (Hemphill and Kidwai 1973; Pivnik and Wells 1996). This marine deposition has been suggested to have occurred in a north-south oriented foredeep proximal to the south flank of a peripheral bulge with a southeast trending axis (Pivnik and Wells 1996). From the early Oligocene through the early Miocene and to the west of this foredeep in the Katawaz basin, the Shaigalu Member of the Khojak Formation, with its sandstone-rich subaqueous to subaerial facies, supports paleodrainage from the early Himalayan orogenic highlands to the Katawaz delta and Khojak fan (Qayyum et al. 2001). Hence, drainage of the Indus River into the Katawaz basin during the Oligocene-early Miocene is a key component to understanding regional terrestrial fluvio-deltaic sedimentation, including the interval encompassing the Bugti Member assigned to the Chitarwata Formation.

Paleodrainage trends spanning the Oligocene to Pliocene strata in the Himalayan foreland in the Sulaiman Range are provided in a study by Waheed and Wells (1990) at Rahki Nala and Chaudwan Zam in the Zinda Pir Dome. Although these authors did not correlate their data strictly to defined formations or to major lithological and paleoenvironmental shifts in their section through the Chitarwata Formation, such as the subunits observed at Dalana, their study indicated the general dominance of southeastward and southwestward-directed paleodrainage directions through their sections. On the Potwar Plateau, the fluvial system of the middle Miocene Kamlial Formation was dominated by southeastward and eastward-directed paleodrainage associated with a large river system (Johnson et al. 1985; Hutt 1996) or alluvial fans (Willis 1993), and was considered by Najman et al. (2003) to represent the first stratigraphic evidence of the Indus River diversion to its current position at about 18 Ma.

Paleodrainage has also been assessed through analysis of provenance data in many parts of the Himalayan Suture Zone. A broad analysis of detrital sandstone modes from the Himalayan suture belt by Garzanti et al. (1996, figure 6) suggested that during the earliest collisional stage (early Eocene) sediment dispersal was to both the eastern and western areas of the suture zone involving a roughly arcuate pattern of drainage about the Nanga Parbat syntaxis. In this scenario, southeastward and eastward drainage prevail on the western side of the syntaxis, and southwestward drainage prevails on the eastern side with components of northward and southward drainage expected as flow off a forebulge. During the subsequent early collisional stage (Oligocene-Miocene), these authors suggested a greater southwestwardly directed component of paleodrainage on the western side of the syntaxis possibly along the proto-Chaman fault zone. This westward drainage model is consistent with the subsequent findings for the Khojak Formation and Himalayan drainage in that region by Qayyum et al. (2001). Additional support for the early evolution of westward drainage is found in the paleocurrent directions in the Eocene Nurla and Oligocene to early Miocene Choksti formations (Clift et al. 2001a). These formations record a shift to a dominant southwestward paleocurrent trend from the dominant northward and northeastward directed flow expressed in the underlying Paleocene Chogdo Formation in the Indian Suture Zone. The corresponding provenance of K-feldspars within the Nurla and Choksti formations indicate a detritus source from the Lhasa block of the Asian plate and support structural control of Indus River paleodrainage by the suture zone since the early Eocene. Moreover, provenance of the siliciclastic Khojak Formation in the Katawaz Basin is consistent with the idea of east-west structural control of the Indus Drainage by the suture zone with a detrital source from the early Himalayan orogen to the east (Qayyum et al.1996; Qayyum et al. 2001).

Our study examined late Oligocene-early Miocene paleodrainage in the subunits of the Chitarwata Formation and lowermost sandstones of the Vihowa Formation at Dalana in the Zinda Pir Dome. This investigation was undertaken: 1) to provide a framework of paleodrainage for the Chitarwata Formation linked to biostratigraphy and chronstratigraphic control, 2) to better understand the relationship of the Chitawata Formation’s marginal marine environments to the broader patterns of concurrent Himalayan tectonism, paleodrainage, and paleogeography, and 3) to investigate how age and preservation differences of vertebrate localities at Zinda Pir and Bugti Hills relate to paleodrainage and paleoenvironmental differences in the western Himalayan Foreland. In particular, the paleodrainage character indicated by the Chitarwata Formation provides an opportunity to investigate models of Indus River flow during the late Oligocene-early Miocene and evaluate whether the Zinda Pir Dome region was structurally and depositionally isolated from the main pulse of Indus River deposition (i.e., to the Katawaz Basin to the west).