Tuesday, January 15, 2013

K.S Valdiya On The Glacial Saraswati In Current Science

In the latest issue of Current Science, geologist K.S. Valdiya has written a long response to a paper   by  Giosan et.al 2012 which concluded that the Yamuna and Sutlej rivers of northwest India stopped flowing into the Ghaggar river by early Holocene. There is a companion paper by Clift et al 2012 on this topic which Valdiya does not elaborate on.

The Ghaggar floodplains formed the agricultural heartland of the Harappan civilization. This finding by Giosan et al and Clift et al if true, meant that during the Harappan civilization the Ghaggar was not a glacial fed river but a monsoon fed river, but likely a perennial one due to a wetter climatic regime in the Ghaggar catchment areas of the Siwaliks. That in turn had implications for Harappan water use and agriculture methods. This question has also fed the controversy about the origin of the Aryans and the relationship between Aryans and the Harrappan civilization, since some say that a glacial Ghaggar was the Vedic Saraswati mentioned in the Rig Ved, an ancient collection of hymns in the Sanskrit language composed perhaps around 1500 B.C. or so.

K.S. Valdiya strongly disagrees with Giosan et al's findings that Holocene climate imparted a characteristic geomorphology to large glacial rivers of this region and that their finding suggests that there was no glacial river flowing on the plains of Haryana and Punjab during mid late Holocene. Goisan et al in due course may provide a detailed reply to Valdiya's arguments. I do have several comments on the way in which K.S. Valdiya has presented his evidence.

1) Validya refers to the work of Saini et al 2009 who analyzed paleochannels near Sirsa in Haryana. Their work shows two main phases of fluvial activity. A major paleochannel complex in the late Pleistocene developed more than 20 thousand years ago, a timeline much before the Harappan civilization.  A much minor fluvial regime was also active between 6000 BP and 3000 BP i.e. during the Harappan civilization. They however draw no conclusion on whether this younger channel during the mid Holocene was glacial fed or monsoon fed. In another recent paper Saini and Mujtaba 2010 carry out more detailed OSL dating of a channel in the same paleochannel complex. They say that a lack of micaceous grains in channel sands of age 6000 BP to 3000 BP point to not a high Himalayan origin but a piedmont (Siwalik) origin. Valdiya does not mention this paper.

Valdiya referring to Saini et al 2009 says that the presence of minerals like tourmaline, hornblende, garnet, kyanite, biotite etc. in sands of the paleochannel would suggest a source in the high metamorphic Himalayas thereby implying a glacial connection. a) Again, the relevant question here is the mineralogy of sediment of mid late Holocene age and Saini et al don't present a detailed analysis of changes in mineralogy with age.  The mineralogy of the younger fluvial regime made up mostly of silty sand and clay is not described in any detail. b) Siwalik sandstones derived from erosion of the metamorphic Himalayas contain these minerals. These metamorphic minerals may also be then recycled from the Siwaliks into younger deposits.

2) Valdiya takes issue on a lack of large scale incised channels in the interfluve around the Ghaggar region. He points out the Holocene landforms if present would have been buried by sand blown in from the Thar desert making it difficult to recognize relict landforms. There are sandy deposits in this region especially in the western parts of Haryana  but if glacial rivers were flowing during the Holocene across the Ghaggar interfluve then why are incised landforms not preserved in the near vicinity where Sutlej and Yamuna come out on to the plains? These areas are quite far away from the influence of major eolian activity. Old terraces of Sutlej and Yamuna along their present day course can be recognized, so why not in the adjacent interfluve, oriented towards the Ghaggar?  Besides, the paleochannel that Saini and Mujtaba  2010 have worked on has a distinct topography outlined by bluffs represented by 5 meters scarps separating the channel from the uplands. Which means that relict landforms are preserved at places.

3) Valdiya emphasizes giving examples that sand has covered many parts of this region and therefore ancient landforms may not be visible to satellite borne instruments. He gives an interesting example which I want to use to make a point. He says that monuments in eastern U.P. are concealed under heaps of sand that look like knolls or small hillocks (p.43). So, sand may blanket and hide the actual feature but the relief may not get smoothened. And SRTM (Shuttle Radar Topographic Mission), - the technology that Giosan et al have used- measures the relief i.e it makes a high resolution map of elevation difference.  Similarly, large incised valleys may get draped by sand over time but unless sand has completely obliterated the elevation difference between the valley bottom and the upland and created a flat surface, that feature may be recognizable on an elevation map created by SRTM.

4) Stream terraces which exclusively drain the Siwaliks contain pebbles derived from erosion of metamorphic rocks. Valdiya writes of these pebbles: Very similar pebbles occur in the higher terraces lining the two banks of the anomalously wide and straight course of the east-flowing petty Bata stream, within the Siwalik  terrane. The Siwalik is wholly made up of softer sandstones, maroon claystones and shales. The remnant of the terraces can be seen at Garibnath and Sudanwala (Figure 11). I have seen the pebbles of metamorphic rocks at Sudanwala. One cannot imagine occurrence of pebbles of Lesser Himalayan rocks in a small tributary of stream originating in and flowing through exclusively in the Siwalik terrane unless the Himalayan-born Yamuna had once flowed through its channel which is as wide as that of the Yamuna. [emphasis mine]

I am surprised that he left out another Siwalik lithology exposed in this area, the conglomerate facies which was deposited in Pliocene-Early Pleistocene. These gravel beds deposited in braided streams within alluvial fan complexes were then deformed along with the other sediments to become the Siwalik hills.  The pebbles in Siwalik conglomerates were derived from a variety of source rocks including metamorphic rocks of the Lesser and Higher Himalayas.

As the Siwalik hills arose, a newer drainage system was initiated. The terraces that Valdiya mentions mark the level of the former river beds of this drainage system but which now lie abandoned at higher elevations as the river progressively cut into its alluvium and bedrock. The terraces will therefore contain material which has been reworked from the Siwalik bedrock and so one can easily imagine pebbles of Lesser Himalayan rocks in a small stream flowing exclusively in the Siwalik terrain. The pebbles may ultimately be of Lesser Himalayan origin but they may be recycled via the Siwaliks into streams that have no connection to the Lesser Himalayas. Valdiya does not consider this alternative explanation.

He further suggests that the Tons which is a glacial tributary of  the Yamuna once flowed westwards through the Poanta valley and joined the Markand river, a tributary of the Ghaggar. Is there sedimentological evidence for this taking place in the mid late Holocene? Are there sedimentary facies and sedimentary structures indicating westerly paleocurrent of a large river in the Poanta valley? Valdiya does not provide any compelling evidence for this scenario.  I suspect that the metamorphic pebbles Valdiya mentions have most likely been sourced from the Siwalik conglomerates.

5) Valdiya misquotes a paragraph from Giosan et al (p.51):

Interestingly, Giosan et al.1 concede that ‘the Yamuna may have contributed sediment to this region…’ (Hakra–Ghagghar) ‘before the Mature Harappan Phase. For we recovered 5400-year-old sandy flood deposit at Fort Abbas (in Cholistan) Pakistan’… ‘And on the upper interfluve, fine-grained floodplain deposition continued until the end of the Late Harappan Phase’

This makes it seem that Giosan et al are suggesting that the Yamuna was flowing in this region around 5400 years ago. Except that this is not at all what Giosan et al write and imply. Here is the full paragraph from Giosan et al.:

Provenance detection (32) suggests that the Yamuna may have contributed sediment to this region during the last glacial period, but switched to the Ganges basin before Harappan times.  The present Ghaggar-Hakra valley and its tributary rivers are currently dry or have seasonal flows. Yet rivers were undoubtedly active in this region during the Urban Harappan Phase. We recovered sandy fluvial deposits approximately 5;400 y old at Fort Abbas in Pakistan (SI Text), and recent work (33) on the upper Ghaggar-Hakra interfluve in India also documented Holocene channel sands that are approximately 4;300 y old. On the upper interfluve, fine-grained floodplain deposition continued until the end of the Late Harappan Phase, as recent as 2,900 y ago (33) (Fig. 2B). This widespread fluvial redistribution of sediment suggests that reliable monsoon rains were able to sustain perennial rivers earlier during the Holocene and explains why Harappan settlements flourished along the entire Ghaggar- Hakra system without access to a glacier-fed river (5, Fig. 3A). 

Valdiya, by quoting only fragments of sentences and adding "before the mature Harappan Phase" and  a "For" in the second line has given the paragraph a different meaning than what the authors intended. Perhaps he can explain why he did not faithfully reproduce the text from Giosan et al.

To summarize:

The Ghaggar was a much larger river in the past. Most geologists won't dispute the basic findings collated by Valdiya on the influence of tectonic activity on fluvial sedimentation and channel migration,  and on the occurrence of paleochannels and thick and wide bodies of buried channel sands suggesting a vigorous long lasting fluvial regime in the past. Valdiya criticizes Giosan et al. on what he think is the excessive importance given by them to climatic controls on fluvial processes. He thinks tectonic influences may also have resulted in the Indus and other rivers incising, and shaping their landforms. He may have a point there. Fluvial processes over the long term may have both tectonic and climatic drivers and the story of the evolution of the Indus megaridge and other features may well be examined from a tectonic perspective as well. That though does not answer the other important question, which is, was the Ghaggar a glacial river during Harappan times.

As the authors of one of the papers Valdiya refers to (Sinha et al 2012)  say; "the timing and provenance of this system remains to be resolved" [emphasis mine]. The question and controversy always has been whether this vigorous fluvial regime reflects a glacial river or a monsoonal river and the timing of changes in fluvial regime if any. Clift et al's study - undertaken in the lower reaches of the same buried Ghaggar- Hakra paleoriver system that Sinha et al 2012 also identify - proposes a provenance and a chronology and takes a first big step towards answering that question. Their findings suggest that glacial fed rivers did flow in these Ghaggar paleochannels, but they changed course by late Pleistocene to early Holocene, several thousand years before the Harappan civilization. Tectonic forces may well have caused the Sutlej and Yamuna to shift course. It just seems to have occurred earlier than what many geologists have reasoned.

In this rebuttal Validya has not presented any new sediment provenance fingerprinting and chronological data to dispute this and he does not address the work of Clift el al directly. I think his suggestion that the provenance of the Sarawati (Ghaggar) system sediments should be studied is a good one. Clift et al have done just that for the lower reaches of the Ghaggar.  New data from the upper reaches of the paleochannels that Valdiya mentions or a newly discovered channel that shows clearly the presence of a glacial river during mid late Holocene would be a serious challenge to Clift et al and Giosan et al's work. But without such new data, Valdiya has not made a convincing argument that the Ghaggar was glacial during Harappan times.

Related Posts:

1) Geological Update On the River Ghaggar
2) New Geochemical and Sedimentological Work On Ghaggar
3) New Geomorphological Work on Ghaggar


  1. Suvrat Kher, When did you think Sutlej took a 90-degree turn at Ropar? Any studies on this which suport your views or Giosan et al views? I am surprised that you have not commented on the course of Sutlej river which by all-reckoning is a glacial river.


  2. Mr. Kalyanaraman- Giosan et al's article points to references and their own dates of deposits of late Pleistocene age on the banks of the present day Sutlej implying that the Sutlej was flowing along its present course by that time. Clift et al data also imply that the Sutlej stopped flowing into the Ghaggar by late Pleistocene earliest Holocene. This strongly suggests that Sutlej changed course much before the Harappan civilization. I have written about this plenty of times in various posts.

  3. I agree that the data provided till now shows that the Ghaggar was rain-fed perennial river during mature Harappan.
    My question is on two political groups in the argument.

    Does a rain-fed perennial river proves that Harappans were not Indo-Aryans?
    Does a glacial river (Ghaggar) proves that Harappans were Indo-Aryans?

    Isn't it a different question altogether,

    Harappans can be Indo-Aryans if Ghaggar was rain-fed/glacial perennial river.
    Harappans can be Non-Aryans if Ghaggar was rain-fed/glacial perennial river.

  4. Anon- I have already expressed my opinion in previous posts that a glacial or non glacial status of Ghaggar does not prove or disprove the Aryan -Harrappa connection. However, geology has been used by many people (example archaeologist B.B Lal) as a important pillar for the argument that the Rig Ved is much older than 2000 B.C and the Aryans and the Harappans are one and the same.

  5. I think following questions are independent of each other.

    Is Ghaggar the Saraswati mentioned in Vedas?
    When was Rigveda composed?
    What was the language(s) spoken by Harappans?

    Also first two questions may not be able to fully answer the question of Who were writers of Vedas?

    Instead of focus on first two questions, we should look at data and provide theories. and do more research and get more data.

    Recent research done by Witzel & Lubotsky on substratum have shown that Dravidian was not spoken anywhere near Punjab, but still people go on discussing Dravidian Harappans.

    Also recent Research by Fuller have used Archaeological and Archaeobotanical data to show continuity from 7000 BC Mehrgarh to 1500 BC Bengal, with Rice domestication in Gangetic plains.
    Also his theories on Southern Neolithic culture fits in Dravidian language expansions.
    His data fits in Theories of Bellwood in Language expansions with Agriculture.

    Why doesn't someone combine all this data and sees the clear picture?