According to the Rig Veda, the river Saraswati originated in the Himalayas, flowed between the Yamuna and the Sutlej and discharged itself in the sea. Many archaeologists believed for decades that the Saraswati fed a portion of the Indus valley civilisation, including the Indus settlements of Banawali and Kalibangan, situated between the present-day courses of the Sutlej and Yamuna rivers. Today, the area of land corresponding to this portion is fed only by ephemeral rivers that flow when it rains, and is otherwise arid.
What happened to the Saraswati? In the absence of conclusive evidence, the river was considered to be ‘lost’, and by some to be the now-feeble Ghaggar river (called Hakra on the Pakistani side). Shortly after coming to power at the Centre in 2014, the Narendra Modi government revived efforts to ‘find’ the Saraswati, considered sacred by many Hindus. The governments of Haryana and Rajasthan also allocated Rs 80 crore together to aid the effort.
A new study by researchers from India and the UK could give them pause. It claims that the portion of the civilisation between the Sutlej and Yamuna was not fed by a perennial river driven by melting Himalayan glaciers, as was previously thought. Instead, the researchers claim that Banawali and Kalibangan were located in a valley that used to be watered by the Sutlej, but that the river had shifted course over 3,000 years before the civilisation was even founded.
In other words, the two settlements had flourished in an abandoned valley.
While the idea is not entirely new, the new study has scored with the amount of data it collected. The researchers obtained sediments at multiple locations in this area and analysed them extensively. They inferred that the river Sutlej had shifted course between 15,000 and 8,000 years ago to its current track, flowing generally west-southwest from its origin near Mt Kailash (see image below). The channel of its previous course is called the Ghaggar-Hakra palaeochannel. And the researchers think the palaeochannel’s banks, and not the Sutlej’s or any other river’s, hosted the cities of Banawali and Kalibangan.
“It is an important paper confirming that the region between the Indus and Ganges watersheds was not fed by Sutlej during Harappan times,” Liviu Giosan, an associate scientist at the Woods Hole Oceanographic Institution, Massachusetts, told The Wire. “This is one more nail into the coffin of the Himalayan Sarasvati (sic) theory!”
The Indus valley civilisation, also known as the Harappan civilisation, spanned the years from around 3,300 BCE to 1,300 BCE, overlapping with the period known as the Bronze Age. Based on excavations since the late 19th century, it was estimated to span a total area of over a million sq. km, larger in extent than the Egyptian and Mesopotamian civilisations of its time. Most of the civilisation was fed by water from the Indus river, which flows from western Tibet through the length of what is now Pakistan before draining into the Arabian Sea in Sindh province.
However, where Banawali and Kalibangan got their water from has eluded consensus. And while many scientists have argued that they were fed by a river, there has also been a debate on whether it was a big Himalayan river or a gentler monsoonal river.
The geological investigation, led by Sanjeev Gupta, a professor at the department of earth science and engineering at Imperial College London, claims to have settled these questions.
First, Gupta and his colleagues obtained images taken by the Landsat Earth-observing satellite of the area in the 1980s, before the advent of large-scale irrigation there, and studied the distribution of soil moisture. “What we discovered was because the palaeochannel was at a topographic low in the landscape, it accumulated fine-grained mud,” Gupta told The Wire. “And mud retains water.”
Next, they went to a site near Kalibangan and drilled up to 45 metres below ground and recovered sand grains. “In the lower part we had coarse sand and then in the upper part, in the top eight metres, the sediment became very, very fine-grained,” Gupta said. The coarser the grains, the more the discharge is of the river that deposited them.
To discover the source of the sediments, the researchers recovered two minerals from the sand – mica and zircon – and measured their ages using radiometric dating. According to Gupta, “What this tells us is not the age of the sediments but where the grains have come from.” They got this answer by comparing the grains and their ages to those collected in the modern-day Sutlej, Yamuna, Ganges and Ghaggar rivers as well as with various rock units in the Himalayas.
“What we found was that the grains we had clearly matched the Sutlej exactly. That’s how we can say that this was the former Sutlej,” Gupta said.
He and his team also subjected the mica to potassium-argon dating. They also employed a third method, called optically stimulated luminescence (OSL). “When mineral grains like quartz and feldspar become buried in a sediment column and aren’t exposed to light, natural radiation causes electrons to become trapped in the mineral’s atomic lattice,” Gupta explained. “When the mineral is re-exposed to sunlight, the electrons are released. What it’s doing is it’s building up charge, if you like, when they’re buried.”
So after the minerals from the sediments were recovered in darkness, Gupta and his colleagues shined infrared light on them. As the electrons are released from their cage of atoms, the minerals luminesce. By measuring the amount of luminescence, you can measure the age of the grain. “And we performed a really detailed analysis using this method and found that their ages ranged from 8,000 years to 70,000 years.”
Both potassium-argon dating and OSL led them to the same conclusion: the people of Banawali and Kalibangan inhabited the valley left behind by the Sutlej river over 8,000 years ago.
For Gupta, what is interesting is that the palaeochannel provided a comfortable landscape, with a “topographic low” – a depression in the land – “where rainfall and monsoon flow would have been concentrated and also protected the cities from the big floods of the Himalayan rivers.”
“If the paper passed peer-review” – which it did – then there is no reason to doubt the paper’s conclusions, according to Maarten Kleinhans, chair in process sedimentology and fluviodeltaic morphodynamics at the faculty of geosciences, University of Utrecht. “I can’t see how they could have done a better job with the means available, although my hands are itching to run some models and experiments.”
But there is room for caution. “There are always pinches of salt to be taken, because this sort of science is like detective work: geologists start with hypotheses, collect the clues as far as possible, available and as far as they recognise them, they build up one or a few possible explanatory stories of how the system developed, and keep checking whether pieces of the puzzle tell a different story,” Kleinhans continued. But because Gupta and team have amassed more evidence than before, “it will require even more data and surprises in that data to overthrow these ideas.”
The summer monsoon’s decline
When a river shifts course, it’s called an avulsion, and “rivers do this”, according to Gupta. Between 15,000 and 8,000 years ago, the Sutlej avulsed from where the Ghaggar-Hakra palaeochannel is today to where the Sutlej itself flows today. The two courses are separated by about 150 km. The avulsion node, the point at which the river changed course, is located close to where the Sutlej exits the Himalayas. While the researchers have not identified the cause of the avulsion, they mention – speculatively – that the Indian summer monsoon might have played a part.
Studies, such as one from 2005 and another from 2015, have suggested that the summer monsoon was declining in strength about 5,000 years ago (with many fluctuations before that as well). The reasons are multifarious, including shifting winds and sea temperatures. So Indus valley sites like Banawali and Kalibangan, dependent as they were on monsoonal sources of water according to Gupta’s study, could have been affected by these shifts. However, Gupta is wary, saying that “archaeological studies haven’t been able to date these events in great detail”.
Around 1,300 BCE, or 3,300 years ago, the Indus valley civilisation began to peter out for reasons that remain unknown. Some speculation among scientists has focused on, among various factors, the summer monsoon (note: the civilisation’s end can’t be attributed to a single cause). To contribute to this debate, a group of archaeologists analysed animal teeth and bone phosphates discovered in a trench in Bhirrana village, Haryana. They used the oxygen isotope levels in the specimens as a proxy for rainfall levels and deduced that the monsoon didn’t directly drive the civilisation’s decline.
Instead, they said in a paper published in May 2016 that the faltering rains forced the Harappans to shift from planting wheat and barley crops to the more drought-resistant millets and rice. As a result, they wrote, “Because these later crops generally have much lower yield, the organised large storage system of the mature Harappan period was abandoned, giving rise to smaller, more individual, household-based crop processing and storage systems, and could act as catalyst for the de-urbanisation of the Harappan civilisation”.
Interest in the wane of the Indus valley civilisation also overlaps with discovering the Saraswati river (should it exist). Its description in the Rig Veda, a collection of Sanskrit hymns written sometime after 1,700 BCE, matches what the investigation led by Gupta has found – in that there was a perennial river flowing in the Ghaggar-Hakra palaeochannel. However, the Sutlej only flowed there till about 8,000 years ago.
Although scientists have figured out how Banawali and Kalibangan could have been watered, other mysteries remain. One is of Rakigarhi, a large city of the civilisation – in the present-day Hissar district of Haryana – that is proximate neither to rivers nor to any known palaeochannels. “It’s very bizarre,” Gupta said. “That’s something that people don’t really understand because there doesn’t seem to be any trace of a palaeochannel there. So that’s work in progress. Rakigarhi is a very big site compared to Kalibangan, so how did they get their water resources?”
Other things in progress will be to confirm what Gupta and his colleagues are implying. For example, their paper states that the conclusions of their study means it “seems improbable that Indus settlements flourished due to ‘perennial’ monsoon-fed river flow”, a suggestion made by Giosan in 2012. However, Giosan said, “Whether the monsoonal rivers feeding the abandoned Sutlej valley were perennial or seasonal during Harappan and Vedic periods remains to be seen. Data presented in the new paper is not diagnostic for either of these interpretations.”
Yet others will be to test Gupta et al’s conclusions as well. According to Kleinhans, this can be done through “studies … that can indicate where the water or sediment came from, or what sort of natural plants grew there at the time to unravel how much water they got. Another alternative could perhaps come from powerful computer modelling of the water and sediment fluxes to test whether their hypotheses do not somehow conflict with the laws of physics.”
“All of this is a lot of work,” he added. In the meantime, we can appreciate a new way to think about rivers. To quote from Gupta’s paper: “It was the departure of the river, rather than its arrival, that triggered the growth of Indus urban settlements”.