A. S. Gaur* and K. H. Vora Marine Archaeology Centre, National Institute of Oceanography,
Lothal, believed to be the oldest dockyard in the world, is located at the head of the Gulf of Khambhat, now situated about 23 km away from the shoreline and about 12 m above the mean sea-level, on the left bank of river Bhogawa. The discovery of Persian Gulf seal13, terracotta models of African mummy, guerrilla, and boat model clearly demonstrate Lothals maritime connection and its relation with Mesopotamia and Egypt. A massive brick structure measuring 213 m ´ 36 m at Lothal is identified as a dockyard and some stone anchors in the vicinity suggest that Lothal was an important maritime trading centre. Moreover there are evidences of marine salinity in the soil of the tank14. Boats might have been reaching the dockyard through a channel of the river Bhogawa and Lothal must have been one of the major Harappan trading centres. Lothal thus provides a clear evidence of southward shifting of shoreline by about 23 km. A recent study15 in the valley of Mahi river suggests that seismic events had taken place between 3320 ± 90 and 2850 ± 90 year BP which may have played a major role in the evolution of the Mahi Basin in particular and Gujarat alluvial plains in general.
Padri is another recently excavated site by the Deccan College, Pune. It is located in the Gulf of Khambhat about 60 km south of the Bhavnagar port and just 1 km away from the shoreline. According to the excavating team, the site belongs to a fairly mature Harappan period, datable to 2200 BC16. There are numerous evidences of exploitation of marine resources at this site. One of the major findings is a large Harappan copper fish hook, which is 14 cm long with barbed point and loop on the other end17 and weighs 45 g. It is in a very good state of preservation. Such large fish-hooks were probably used to catch large marine fish weighing more than 50 kg. This indicates that the Harappans of Padri had perhaps mastered the technique of deep sea fishing. It is logical to presume that, for this purpose, they also used a big boat. The excavator of Padri site also inferred that the people there were producing salt and supplying it to the surrounding Harappan settlements18.
Dholavira is yet another Harappan site located in the Great Rann of Kachchh on a Khadir Bet. At present the Rann is a dry area but during good monsoon it gets flooded. Dholavira is supposed to have witnessed the earliest habitation of protohistoric period in Gujarat. Excavation has revealed a long cultural sequence which commences from the beginning of the third millennium BC, when perhaps a group of people from Makran coast arrived on the island through Kori creek. This assumption is based on the ceramic feature resembling those from the Amerian culture (datable to 3000 BC). Similar pottery has also been reported from other Harappan sites of Kachchh. The base of Rann of Kachchh might have been under 10 m deep water for 3000 years19 and the Khadir Bet could have served as an island in the shallow sea. It is not clear why people settled on an island rather than on the mainland, where agriculture and other commodities, including marine resources could be better exploited. The inference is that Dholavira was an active port and the Harappans must have found that this port was a safe harbour for anchoring boats. The long habitational history of the area highlights the importance of the location and of maritime activities. The location of the site seems to be favourable even for riverine navigation, in case one is not inclined to consider this a port site.
Kuntasi is another site locally known as Bibino Timbo located about 3 km south-east of the Kuntasi village. It is about 5 km inlandward from the present shoreline. A large stone structure measuring 9.5 m ´ 4.1 m outside the fortification, according to Dhavalikar et al.20 was used as landing platform for going down to the creek for loading-unloading cargo in boats. They further suggested that it was not an agricultural settlement but was a centre for acquiring and processing raw materials for manufacturing articles for export. Discovery of two cylindrical stone anchors with tapering sides, and large holes bored vertically throughout the length and the overall geomorphology of the area corroborate that Kuntasi could have been a port situated at the creek mouth during the Harappan period.
The protohistoric site is located at Bet Dwarka, Okhamandal, Jamnagar located south-east of the island near Siddi Pir Baba21. Onshore explorations in and around this site since 1979 have yielded a large quantity of protohistoric pottery including perforated jar, sharp carinated dish, dish on stand and bowls. Other important findings include two inscriptions, one seal and a few chert blades. These antiquities enable us to assume that Bet Dwarka was established at the beginning of the second millennium BC. Discovery of a large number of conch shell collumela and shell bangles in the region suggests that this site was a small port during the Harappan period. It is observed that the protohistoric habitation in the cliff section of Bet Dwarka-I (Figure 2) is being destroyed by the sea. Similarly, in the past the protohistoric habitation was destroyed near Balapur BDK-VI by an advancing sea. A historical period22 site located in intertidal zone (Figure 3) gets submerged during high tide. Obviously, the shoreline has been advancing since the Harappan Period.
Excavations at Malvan in south Gujarat23 suggest that this was a post-Harappan estuarine port, dating back to 1400 BC. The site was located on the banks of an oxbow lake formed by the Dumas branch of the Tapti river. Studies of sediment samples from this lake suggest that during the post-Harappan time tides would have been two and a half m above the dark grey clay bed. Hegde24, after analysing the sediment samples, reported that during the post-glacial climatic optimum (between 5000 BP and 3000 BP) the warmest period since the ice age, the eustatic sea-level was higher than the present. It was probably during this period that barges were floated into the lake with the tide. This established the point that Malvan was a port. Since silting was sufficiently advanced and the sea-level retreated, the site could not have functioned as a port.
Shifting of shoreline due to one or other reason is not a new phenomena. Dales25, who discovered a few Harappan sites along the Makran coast, believes that these sites must have served as a sea port during the third millennium BC and now these sites are located as far as 30 miles inland. These displaced ports reveal that the coastline in this part of Pakistan had risen considerably during the past 4000 years. Similarly, several workers have demonstrated that the sea-level was higher than the present around 6000 BP (4000 BC) at the Saurashtra coast10,26 (Table 1) as well. In view of this, it can be surmised that the above-mentioned archaeological sites on Kachchh and Saurashtra coast, which happened to be contemporary, had been connected with maritime activities, and therefore, can also be used as indicators of palaeo-shorelines for the region between 3000 BC and 1500 BC. However, it is also necessary to discuss other geological aspects connected with this.
Archaeological evidences suggest that the behaviour of shoreline was different at different places. Lothal is located about 26 km from the present shoreline and about 12 m above the MSL (Table 2). The gulf is narrow and a few major rivers debauch here such as Narmada, Tapti, Mahi, Luni, etc. Offshore currents are also responsible for depositing the sand in the Gulf, ultimately resulting in the shallowing of the Gulf. Earlier studies mentioned that the Gulf of Khambhat and the Gulf of Kachchh were connected during the early mid-Holocene period10. Bruckner27 had suggested that the river-generated sediments may have played a vital role in the seaward movement of the shoreline.
The Great Rann and Little Rann are unique examples of Holocene sedimentation. The two Ranns represent filled-up gulf and mark the site of accumulation in an estuarine delta environment that was marked by a fluctuating strandline since the advent of Holocene. Gupta28 mentioned that Holocene sediments of the Little Rann and Nal Lake were contemporaneous. The lowermost sandy clay horizon extends from about 9000 BP to about 4200 BP overlain by the silty clay horizon dating from 4200 to 1500 BP and then again by the most recent silty clay horizon. He further suggested that even as late as 2000 years ago, Little Rann was about 4 m deep and thus was inundated throughout the year.
Another important factor of shifting of the shoreline may be due to tectonic effect. Recent studies on sea-level changes clearly suggest that neotectonic activity had also played a vital role in sea-level fluctuations, particularly in Saurashtra and Kachchh region, during the Late Pleistocene/ Holocene. Studies on sea-level along the Mumbai coast29 suggest that until 8300 BP shallow sea-level conditions prevailed on the carbonate platform located at about 8090 m depth off Bombay. This is in contrast with the glacial eustatic sea-level position which was at about 22 m at 8300 years BP.
There are also evidences of earthquake from phase III of Dholavira in Khadir Bet of Rann of Kachchh30 which may be datable to around 2200 BC. Perhaps the effect of earthquake such as collapse of houses and diversion of drainage system, and finally the uplift of Rann were responsible for the decline of Harappans at Dholavira. There are evidences of navigation in the Rann of Kachchh during the historical period. The author of Periplus of Erythraean Sea writes Beyond the river Sinthus there is another gulf, not navigable, running in towards the north; it is called Eirinon; its parts are called separately the small gulf and the great; in both parts the water is shallow, with shifting sand-banks occurring continually and a great way from shore; so that often when the shore is not even in sight, ships run aground, and if they attempt to hold their course they are wrecked31.
In the case of Bet Dwarka, one can observe that archaeological material is spread all over the intertidal zone of BDK-I and II. The earliest date of Bet Dwarka could be put around the beginning of second millennium BC on the basis of archaeological findings. Remains of perforated jar, two inscriptions and one unique seal clearly indicate the settlement during the Harappan period in the area. A settlement of second century BC at Dwarka (about 15 km south of Bet Dwarka by sea route) was destroyed by the sea32, which again is an indication of sea-level rise in the past.
Archaeological sites provide evidences on shoreline fluctuation along the Gujarat coast at about 50004000 years BP. If a possible shoreline (Figure 4) of Harappan period is drawn touching the Harappan ports, the following conclusions can be arrived at:
(i) The evidences indicate that in the Gulf of Khambhat, a distinct shifting of shoreline offshoreward is observed. On the other hand, south and western Saurashtra coasts appear to be stable, at least for the last 4000 years. (ii) If Dholavira is established as a port by more convincing findings than what is presented here, it will have significant importance in the palaeoclimatic study as well as archaeology and anthropology of the region. It would then mean that the Great Rann was navigable during the Harappan period (4500 years BP) and its subsequent siltation may have even triggered the decline of Harap-pan civilization in the area. (iii) The study clearly demonstrates the use of archaeology in establishing palaeo-shoreline. While the tectono-eustacy can only indicate the possibility of occurring of an event, evidences like this clearly support it.
Dhavalikar, M. K., Rawal, M. R. and Chitalwala, Y. M., Kuntasi
ACKNOWLEDGEMENTS. We thank Dr E. Desa, for constant encouragement. We are grateful to Dr V. P. C. Rao for a critical review of the manuscript and for valuable suggestions. We thank all our colleagues in the Marine Archaeology Centre for assistance during the field work. S. B. Chitari, U. S. Sirsat and S. A. Karim helped at various stages. This is a NIO contribution No. 2641.
Received 19 January 1999; revised accepted 29 March 1999