The Rig Veda's Greatest River Has No Water. ISRO Says It Found It Under the Sand — Here Is Exactly What the Satellites Saw

Jodhpur, November 2014: a report signed on a desert campus

The Regional Remote Sensing Centre (West) sits inside the Central Arid Zone Research Institute campus at Jodhpur, on the edge of the Thar Desert. In that month its scientists released Technical Report No. RCJ: 04.9/2014/Saraswati-Integ, River Saraswati: An Integrated Study Based on Remote Sensing & GIS Techniques with Ground Information. Four names are on it: J. R. Sharma, B. K. Bhadra, A. K. Gupta and G. Sreenivasan. The project had begun on 1 April 2004.

The preface does not hedge. "The existence of an extinct mighty river 'Saraswati' in the Thar Desert region is well known to the people of northwest India," it begins, and ends: "It is hoped that this scientific findings will give an end to the debate on the existence of 'River Saraswati' in the past and the reasons for its disappearance."

What the report describes is a decade of looking at satellite pictures. Indian satellites had photographed Punjab, Haryana, Rajasthan and Gujarat in February 2004 with four cameras, from one that sees a 188-metre square as a single dot to one that sees 5.8 metres. The scientists stretched the contrast on patches ten kilometres square until faint dark ribbons appeared in the sand. They drew the ribbons. Then they went to see whether anything real lay underneath.

That is the question this Investigation follows: not whether the Sarasvati is sacred, but whether a river of the size the hymns describe ran where the satellites drew their lines, and when.

At a glance

• Seventy-five verses of the Rig Veda mention Sarasvati, "spread over all the ten Mandals (books) except the fourth"; "no other river has received so much importance", the report says. (ISRO/NRSC Report RCJ: 04.9/2014, p. 9)
• "Best Mother, best of Rivers, best of Goddesses, Sarasvatī" — Rig Veda 2.41.16 in Ralph Griffith's 1896 translation. (Griffith, Hymns of the Rigveda, Book II, Hymn XLI, v. 16)
• 6 to 8 km wide, "as much as 10 km in some reaches" — the dry channel of the Ghaggar–Hakra, as described by the Government of India's 2016 expert committee. (Palaeochannels of North West India, 2016, p. 65)
• Fourteen tube wells drilled on satellite-mapped palaeochannels in Jaisalmer district hit coarse river sand at 40–125 m depth in nine of them; the water is drinkable where water nearby is salt. (ISRO/NRSC 2014, pp. 37–41)
• 1,340 to 18,880 years old — the radiocarbon ages of seventeen groundwater samples from those channels, measured by the Bhabha Atomic Research Centre. (ISRO/NRSC 2014, p. 43)
• "Only monsoonal-fed rivers were active there during the Holocene" — Giosan and colleagues in PNAS, 2012, against the glacier-fed river. (Giosan et al., PNAS 109(26), 2012, Abstract)

First, the words a beginner needs

Palaeochannel. The bed of a river that no longer flows there; "palaeo" means old. It may be a visible dry valley or a ribbon of river sand buried under later soil and dunes. River sand holds water better than clay, so a palaeochannel is usually also an underground reservoir.

Ghaggar–Hakra. One river with two names. It rises in the Siwalik hills near Chandigarh as the Ghaggar, crosses Haryana and north Rajasthan, and dies in the sand near Anupgarh. In Pakistan's Cholistan the same dry bed is the Hakra; further south, the Nara. Below Sirsa it carries no water most of the year.

Sarsuti / Saraswati Nadi. A small seasonal stream in Haryana from Adi Badri past Kurukshetra and Pehowa to the Ghaggar near Rasula. Its name is the local form of Sarasvati and appears on Survey of India maps of 1969–70. The ISRO report treats it as distinct from the Vedic river.

Rig Veda and the Nadistuti. The oldest collection of Sanskrit hymns, 1,028 in ten books ("mandalas"). Hymn 10.75, the Nadistuti or "praise of the rivers", lists the rivers of the north-west in order: the oldest geography of India we have.

IAST. The standard way of writing Sanskrit in Roman letters with dots and long marks, so sarasvatī and śutudrī can be read exactly. The verses that carry the argument are given once in IAST, then in plain English.

Harappan (Indus) civilisation. The Bronze Age cities of the Indus plain, roughly 2600 to 1900 BCE in their urban phase. Kalibangan, Banawali, Bhirrana and Rakhigarhi are Harappan sites on the Ghaggar system in India; Ganweriwala is in Cholistan.

IRS, WiFS, AWiFS, LISS, SAR. India's Earth-observation satellites and their cameras: WiFS sees a 188-metre square per pixel, AWiFS 56 m, LISS-III 23.5 m, LISS-IV 5.8 m. Radarsat SAR is a Canadian radar satellite that senses moisture under thin dry sand.

OSL dating, Holocene, BP. Optically Stimulated Luminescence measures when a grain of quartz sand last saw sunlight, that is, when it was buried. The Holocene is the last 11,700 years, since the ice age ended; "BP" means years before 1950.

Monsoon-fed versus glacier-fed. A monsoon-fed river rises in low hills and swells only with the rains; in a dry year it fails. A glacier-fed river rises above the snowline and flows all year. The Sutlej and Yamuna are glacier-fed. The Ghaggar is not.

The claim we are testing

The popular version: "The Sarasvati of the Rig Veda was a mighty glacier-fed river, fed by the Sutlej and the Yamuna, along which the Harappan cities grew; it dried up around 1900 BCE when its tributaries were captured, and ISRO's satellites have now found it under the Thar." Layer by layer:

1. Buried river channels, some very wide, run from the Siwaliks into Cholistan and the Rann, and the Harappan sites line them. True. Satellite mapping since 1979, hundreds of well logs and the 2016 Government of India review all agree.
2. Those channels once carried water from the Higher Himalaya. True for the deep past, disputed for the Harappan and Vedic periods. Himalayan sand is in the channel; when it last arrived is put at about 5,000 years ago by the Valdiya committee and before 10,000 years ago by Giosan and Clift.
3. The Rig Vedic hymns describe this channel at its full glacier-fed size. Unsupported. No document we opened dates a Himalayan-fed flow in the Ghaggar to the period when the hymns were composed. The ISRO report asserts the link; it does not demonstrate it.
4. The river "disappeared around 3000 BC" and the Rig Veda was composed "more than 8000 years ago". Myth layer. These are the ISRO report's own figures, resting on a dating of the Rig Veda the report does not argue for and standard scholarship does not share.
5. A single lost river ran from Lake Mansarovar in Tibet to Dwarka. Myth layer. The report says the linkage is "established"; the evidence is a line joining the modern Sutlej, a 75 km palaeochannel, the Ghaggar–Hakra and the Rann.

Document one: what the hymns say, in the order they say it

The Rig Veda has one hymn that reads like a map. In Book 10, hymn 75, verse 5, Griffith translates:

"Favour ye this my laud, O Gaṅgā, Yamunā, O Sutudri, Paruṣṇī and Sarasvatī: With Asikni, Vitasta, O Marudvrdha, O Ārjīkīya with Susoma hear my call."

Something is wrong with that English, and it matters. The Sanskrit runs imáṃ me gaṅge yamune sarasvati śutudri stómaṃ sacatā páruṣṇy ā́: Gaṅgā, Yamunā, Sarasvatī, Śutudrī, Paruṣṇī. Griffith's English moves Sarasvatī to the end of the line. In the original the poet walks east to west: Ganga, Yamuna, then Sarasvati, then the Sutlej (Śutudrī), the Ravi (Paruṣṇī), the Chenab (Asiknī), the Jhelum (Vitastā). The Sarasvati sits between the Yamuna and the Sutlej. That is exactly where the Ghaggar runs today, and nobody disputes it.

The praise is elsewhere, and it is extravagant. Book 2, hymn 41, verse 16 — ámbitame nádītame dévitame sárasvati, apraśastā́ iva smasi práśastim amba nas kṛdhi — Griffith gives as: "Best Mother, best of Rivers, best of Goddesses, Sarasvatī, We are, as 'twere, of no repute and dear Mother, give thou us renown."

Book 7, hymn 95, verse 1 has her "surpassing in majesty and might all other waters". Verse 2, the line every modern argument quotes — ékācetat sárasvatī nadī́nāṃ śúcir yatī́ giríbhya ā́ samudrā́t — "Pure in her course from mountains to the ocean, alone of streams Sarasvatī hath listened." Book 6, hymn 61, verse 2: "She with her might, like one who digs for lotus-stems, hath burst with her strong waves the ridges of the hills."

Read together, the hymns describe a river that comes from mountains, breaks through hills, floods, reaches a samudra, and outranks all other rivers. If that is a literal description of one river in the Punjab, it is not today's Ghaggar. One caution from the text itself: samudra in Griffith is "ocean", but the word also means "a gathering of waters", and Book 3 uses it of the Vipāś and Śutudrī (3.33.2: "ye move as 'twere on chariots to the ocean"), which reach no ocean on their own.

Document two: the Mahabharata already knows the river is dead

By the time the Mahabharata was composed, the Sarasvati was a river that stopped. The epic says so in the voice of pilgrims.

In the Vana Parva (Book 3) the sage Pulastya, reciting the pilgrim's itinerary to Bhishma, directs the traveller "to Vinasana, where Saraswati disappearing on the breast of Meru, re-appeareth at Chamasa, Shivodbheda and Nagadbheda" (Section LXXXII, Ganguli). Vinaśana means "the vanishing". The fullest account is Balarama's pilgrimage up the Sarasvati from the sea, in the Shalya Parva (Book 9), Section 37: "Then Valadeva, O king, proceeded to Vinasana where the Sarasvati hath become invisible in consequence of her contempt for Sudras and Abhiras. And since the Sarasvati, in consequence of such contempt, is lost at that spot, the Rishis, for that reason, O chief of the Bharatas, always name the place as Vinasana."

The river ends in the ground, not at the sea; the place has a name, and pilgrims visit it; and the explanation is social, not geological. The epic is describing, in its own idiom, a monsoon river that sinks into desert sand, which is what the Ghaggar does at Anupgarh. The ISRO report reads it the same way: "the river vanished in the desert sand at certain place along its regular course i.e. at Vinasana" (p. 10).

So between the Rig Veda and the Mahabharata, in the texts alone, the Sarasvati goes from "surpassing in majesty and might all other waters" to a stream that dies in the sand. Either the river changed, or the poets did, or both. That is the gap the geology has to fill.

Document three: what the satellites saw

The method is simple to state. Satellite cameras record the brightness of the ground in several colours, including infrared. Damp sand and the plants on it look different from dry dunes. A buried channel of river sand holds moisture; after enough image-stretching it shows as a darker, winding ribbon, a "serpentine drainage course with high moisture content (dark tone)" (p. 20).

The data: IRS-P3 WiFS and IRS-P6 AWiFS, LISS-III and LISS-IV images of February 2004, plus Landsat ETM and Radarsat SAR radar of December 2002 for Punjab (pp. 7, 23). Canals, seasonal streams and waterlogged fields were excluded by hand (p. 20).

In Rajasthan, a network of ribbons from the Ghaggar's end near Anupgarh south-west through Jaisalmer towards the Rann (pp. 22–23).

In Punjab, "the delineated palaeochannel between Ropar and Patiala is named as Sutlej palaeochannel which is an N–S trending palaeochannel extending for a length of about 75 km with having width from 1 to 6 km. The Sutlej palaeochannel is connecting the present day Sutlej River near Ropar and the Ghaggar River in the south of Patiala" (p. 23). This is the key link in the whole argument: a channel joining the glacier-fed Sutlej, where it turns sharply west at Ropar, to the Ghaggar. The radar "could able to pick up moist zone (torquous channel with dark tone), possibly indicating a subsurface palaeochannel" (p. 23). "Possibly" is the report's word.

In Haryana, the Sarsuti, Markanda and Chautang courses, with the Harappan sites plotted on top (pp. 23–24, 35). In Gujarat, a "Bird's Foot type" delta pattern in the Great Rann, read as the Sarasvati's mouth (p. 25); after the Bhuj earthquake of 26 January 2001, channel patterns appeared there on satellite images that were absent weeks before, as shaking forced groundwater up through buried sand (p. 27).

Put end to end, the ribbons run from the Siwalik foothills to the sea. The report's Figure 27 draws them as one system. That drawing is what "ISRO found the Sarasvati" means.

Document four: what the drills found

A satellite ribbon could be a buried canal, a fossil dune slack or a trick of the light. The report's real strength is that it went underground.

Jaisalmer. Fourteen tube wells were drilled on the mapped channels along the Tanot–Kishangarh–Longewala–Ghotaru line. "Coarser sediments are noticed at a depth ranging from 40–125m (general depth of 55–85 m) and have been encountered in 9 of the 14 tube wells drilled. Channels are about 35 to 80m thick with depth of water level ranging from 35–60m" (p. 37). Jaisalmer gets less than 150 mm of rain a year. On the channels the water met Rajasthan's drinking norm; away from them it was "saline (non potable)" (p. 41), and after 35–40 years of pumping there was "no indication of any decline" (p. 41). A river you cannot see is still supplying villages.

Haryana. Well logs from nine northern districts show "medium to coarse grained sands and associated gravel and pebbles at a depth between 10 to 100m" near the mapped channels (p. 40). Resistivity soundings along the Ghaggar, reported in 2013, found "a thick and extensive sand body" whose dimensions "imply that these are the deposits of a large river system" (p. 39).

Groundwater age. Seventeen water samples from the Jaisalmer channels were dated by the Bhabha Atomic Research Centre using tritium, oxygen-18 and carbon-14: "from 1340 to 18880 BP at different localities from NE to SW", the oldest at Sadewala (p. 43). And one sentence the report qualifies in the next breath: "The ground water do not have isotopic signature of present day Himalayan Rivers. However, the deep groundwater indicates towards origin from Himalayan source" (p. 43).

Archaeology. From a 1984 compilation by Joshi and colleagues the report tabulates Early Harappan, Harappan and Late Harappan sites: Haryana 103, 44 and 297; Punjab 23, 32 and 102; Rajasthan 8, 28 and none (Table 1, p. 33). None of the Early or Mature sites lie on the present Sutlej or Yamuna, and no Late Harappan sites exist in Rajasthan along the Ghaggar bed, "which must have dried up during that period" (p. 33).

The 2016 expert committee, chaired by the geologist K. S. Valdiya, put the mapped palaeochannels of Haryana, Rajasthan and Gujarat at "more than 2200 km" and the settlements on the Ghaggar–Hakra system at "nearly 2378" (Palaeochannels of North West India, 2016, Summary and p. 67). Something big ran here. The question is when it last ran big.

Document five: the dating that splits the room

Everything turns on a column of sand in western Haryana, and on who reads it.

The ISRO report leans on an OSL study by H. S. Saini and colleagues (Current Science, 2009), which reconstructed the buried Ghaggar floodplain between Tohana and Ellenabad from well logs and OSL-dated its sand layers. The report's summary (p. 42):

• 26,000–28,000 years ago: a "Himalayan-fed channel network", humid climate.
• 21,000 to 15,000 years ago: weak flow, then dunes; arid.
• 5,900–2,900 years ago: a channel, sub-humid climate, which the report labels "Vedic Saraswati channel".
• 4,300 to 2,900 years ago: "Desiccation of the channel".
• After 2,900 years ago: a "weak, piedmont-fed Ghaggar system".

Notice what the labels do. The genuinely Himalayan-fed channel is 26,000 years old, long before any Veda. The channel of 5,900–2,900 years ago is called "Vedic Saraswati", but an OSL date says only that sand moved then, not where the water came from. The report's conclusion stretches across both: the river "may be as old as 28,000 years and flourished during 8000 to 5000 years ago" (p. 55).

The Valdiya committee took the stronger line. Sand under Kalibangan, dated by OSL to between 9,700 and 5,000 years ago, carries strontium and neodymium ratios "implying derivation of the sediments from the Great (Higher) Himalayan rocks, the river depositing its load until about 5000 years ago" (2016, p. 75). The committee concluded that the ancient Sutlej flowed down the Ghaggar as a western branch and the ancient Tons–Yamuna down the Sarsuti–Markanda as an eastern branch, meeting at Shatrana, "25 km south of Patiala where the Ghaggar channel abruptly becomes 6 to 8 km wide" (pp. 72, 87). On that reading a Himalayan river ran here until about 5,000 years ago, then went sluggish between 4,300 and 2,900 years ago, in the Harappan decline.

Document six: PNAS, 2012 — "only monsoonal-fed rivers"

In 2012 (online 29 May, print issue 26 June) the Proceedings of the National Academy of Sciences published "Fluvial landscapes of the Harappan civilization" by Liviu Giosan of Woods Hole, Peter Clift, Mark Macklin, Dorian Fuller and eleven co-authors, built on radar elevation data, cores in Pakistan, and radiocarbon and OSL dates. The abstract is direct: "Contrary to earlier assumptions that a large glacier-fed Himalayan river, identified by some with the mythical Sarasvati, watered the Harappan heartland on the interfluve between the Indus and Ganges basins, we show that only monsoonal-fed rivers were active there during the Holocene."

The argument is about shape. Glacier-fed rivers in the Punjab, with steady year-round flow, cut themselves deep valleys in the early Holocene as the monsoon weakened and they carried less sand. The Ghaggar–Hakra tract shows no such cutting: "the lack of large-scale incision on the interfluve demonstrates that large, glacier-fed rivers did not flow across the Ghaggar-Hakra region during the Holocene." Sand chemistry "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."

That is not the same as saying the Ghaggar was always a trickle. "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, and recent work on the upper Ghaggar-Hakra interfluve in India also documented Holocene channel sands that are approximately 4,300 y old." The explanation: "reliable monsoon rains were able to sustain perennial rivers earlier during the Holocene", and as the monsoon weakened after about 5,000 years ago, "monsoonal rivers gradually dried or became seasonal".

So the two documents agree on more than the headlines suggest. Both say a real, perennial river ran down the Ghaggar–Hakra in Harappan times. Both say it failed between roughly 4,300 and 2,900 years ago, and the cities emptied as it failed. They disagree on one thing: whether the water came from Himalayan glaciers through the Sutlej and Yamuna until about 5,000 years ago (ISRO, Valdiya), or from a strong monsoon through the Siwalik hills, the glacier link having ended before the Holocene began (Giosan, Clift).

Document seven: the state takes the river back

The ISRO report ends with a plan. Section 7.2, "Suggestions for Reviving Saraswati River": clear silt from the Sarsuti, stop sewage entering it, cut short channels to the Somb and the Markanda, and "as a permanent solution" a 6–7 km feeder canal from the Western Yamuna Canal "to make it perennial" (pp. 56–57). The Sarsuti today, it says, "has turned into a sewerage drain" (p. 31).

Eleven months later the Government of Haryana constituted the Haryana Saraswati Heritage Development Board by notification no. 1/13-2015-1PP, dated 12 October 2015. Its purposes include "construction of DAM, Reservoir, and Barrage for continuation of flow of water in the palaeochannel of Sarasvati". The Board puts the alignment from Adi Badri to the Ghaggar at about 200 km. The Chief Minister chairs it.

In March 2016 the Union Ministry of Water Resources constituted its own expert committee under Valdiya. Its report of 15 October 2016 recommended, in order, a databank of palaeochannel maps, aquifer delineation, and "Mission Palaeochannel Recharge" — using the buried channels to store floodwater, with an initial estimate that about 1,058 million cubic metres could be recharged across Haryana, Rajasthan and Gujarat (2016, pp. 63, 87–88).

Whatever the river was called, its sand is where the water is.

Read the actual document

From the ISRO/NRSC report, Chapter 6, section 6.2, "Linkage of Saraswati Nadi with Yamuna River", pp. 50–51:

"However, the catchment area of Saraswati Nadi is confined to the southern slope of the Siwalik Hills and the river receives only seasonal flow due to low precipitation. Hence, no linkage of Saraswati Nadi with any major perennial river systems like Yamuna could be found out based on the available data and evidences. It remains doubt that Saraswati Nadi could have formed a powerful drainage system in the past and served as the main channel of Vedic Saraswati River."

Five pages later the Conclusions state that "a linkage of the entire drainage network has been established from Mansarovar (Tibetan Himalaya) to Dwarka" (p. 56). The body says "no linkage could be found"; the conclusion says the linkage is established. A reader is entitled to notice.

And from Giosan et al. 2012, the sentence that answers the ISRO map directly:

"However, the lack of large-scale incision on the interfluve demonstrates that large, glacier-fed rivers did not flow across the Ghaggar-Hakra region during the Holocene."

In plain words: a big glacier river, running for thousands of years, digs itself a trench. The Sutlej and the Yamuna have trenches. The Ghaggar does not.

The strongest case for the other side

The case that the Vedic Sarasvati was a Himalayan river, and that the ISRO map is right in substance, deserves its best form.

First, the hymns. 7.95.2 says she runs "from mountains to the ocean". 6.61.2 says she "burst with her strong waves the ridges of the hills". 10.75 ranks her among the Sutlej, Ravi, Chenab and Jhelum, every one glacier-fed, and puts her where the Ghaggar is.

Second, the channel. A valley 6 to 8 km wide does not come from the Siwalik hills. At Shatrana the Ghaggar "abruptly becomes 6 to 8 km wide" exactly where the Sutlej palaeochannel from Ropar would join it (2016, p. 72). The sand at Kalibangan is chemically Higher Himalayan and was deposited until about 5,000 years ago; Clift's zircon dating puts Yamuna and Sutlej sand in the Hakra until about 4,500 years ago (p. 75). The disagreement is about centuries, not about whether the Himalaya ever reached the Ghaggar.

Third, the cities. Kalibangan, Banawali, Bhirrana and Karanpura have no wells inside the settlement, which "implies that the inhabitants were still depending upon the water from the river directly" (2016, p. 48). People do not build a 150-hectare city on a stream that fails in a dry year.

The weakness of this case is the date it needs. For the hymns to describe the Himalayan river, that river must have been flowing when the hymns were composed. The committee's own dates end the Himalayan supply at about 5,000 years ago and the river's vigour at 4,300. The conventional dating of the Rig Veda is more than a thousand years later. To close the gap, the ISRO report moves the Rig Veda to "more than 8000 years ago" (p. 1) and the Mahabharata war to 1792 BCE by "astronomical dating" (Table 7, p. 45). Those are the load-bearing numbers, and the report does not argue for either; it cites them.

What we could not prove

• That any Rig Vedic hymn was composed while a glacier-fed river ran down the Ghaggar. The two dating columns do not overlap unless the Rig Veda is moved thousands of years earlier than its conventional date, and we did not open a primary source that justifies the move.
• The Sutlej palaeochannel's age. The ISRO report maps a 75 km channel from Ropar to the Ghaggar. We found no OSL or radiocarbon date for it in the report; its age, which decides when the Sutlej last fed the Ghaggar, is inferred, not measured.
• Clift et al. 2012, Saini et al. 2009 and the Oldhams (1886, 1893) at first hand. All are quoted through the ISRO and Valdiya reports' summaries.
• The Jamison–Brereton translation. Not open-access; not consulted. All Rig Veda English is Griffith (1896), checked against the Sanskrit on Sanskrit Wikisource for 10.75.5, 2.41.16, 7.95.1–2 and 6.61.2. We did not open a lexicon to settle the meaning of samudra.
• The archaeology counts. Table 1 of ISRO 2014 and "nearly 2378" in the 2016 committee come from compilations by other authors; we did not verify them against the original surveys.
• The report's own date. The cover says November 2014; the document control sheet says "Date of Publication 28th November, 2015". We use the cover date.

Why it still matters today

The ISRO report is not a hoax and not a press release. It is 76 pages of named satellites, listed wells, measured depths and tabulated ages, most of it corroborated by independent work before and since. Where it overreaches is at the seams: when a dated sand layer is given a name from the Veda, when a "possibly" in the body becomes "established" in the conclusions, when a line on a map joins Mansarovar to Dwarka.

And there is water in it. The palaeochannels hold drinkable water in a desert where the water nearby is salt. The 2016 committee's first three recommendations have nothing to do with the Veda: map the aquifers, measure them, recharge them. The Haryana board's plan to run canal water down the Sarsuti is, whatever else it is, a water project on a channel that was found by looking for a goddess. Both can be true.

Verdict

The documents prove that a very large river once ran down the Ghaggar–Hakra from the Siwaliks to the Rann of Kachchh; that its bed, 6 to 8 km wide, is still there under the fields and dunes; that the Harappan cities stood on it and emptied as it failed between about 4,300 and 2,900 years ago; and that its buried sand still holds water thousands of years old. ISRO's satellites did not invent that river. They mapped it, and the drills confirmed it.

The documents do not prove that the river the Rig Veda praises is that river at that size. The Himalayan water in the channel is dated, by the people most convinced of it, to before about 5,000 years ago; the hymns are conventionally dated more than a millennium later; and the one paper that tested the landscape for a glacier-fed river in the Holocene found none. The ISRO report closes the gap by moving the Veda, not by dating the river.

What the Rig Veda's poets saw may have been a monsoon river in a wetter age, running all year and flooding every summer, the "best of rivers" in the only country they knew. What the Mahabharata's pilgrims saw was the same river dying in the sand at Vinasana. What the satellites saw was its grave. All three are real. The claim that they are the same river at the same size is the one thing nobody has yet shown.