Besides, in 1983, M.B. McKean studied pollen and sediments in the region of Balakot (northwest of Karachi) and concluded: ‘There is nothing in the Balakot pollen data, which might suggest that the climate during the protohistoric period in Las Bela was decidedly wetter than at present.’15 And in 1999, Y. Enzel and eight colleagues analysed sediments of the now mostly dry lake of Lunkaransar and found that it held water in 8000 BCE, began to decline around 4000 BCE, and dried up by 3500 BCE.16
Fig. 8.1 graphically captures the above seven studies, all of which conclude that the Mature Harappan phase developed in an arid environment.
OR, WAS IT IN A WETTER PHASE?
That is not the end of the matter, however. Against the above seven studies, we can array seven others (Fig. 8.2), delineated below, that tend towards an apparently opposite view:
In 1983, R.J. Wasson and six Indian colleagues probed sediments in the Didwana lake of Rajasthan, which still holds some salty water (Fig. 8.3). They found that ‘freshwater, high lake level conditions prevailed’ between 4000 and 2000 BCE.17 This precisely includes the Mature Harappan phase.
In 1996, P.D. Naidu, studying planktonic foraminifers from the Arabian Sea, found that the upwelling, and therefore the southwest monsoon, was at its lowest from about 1500 BCE to 800 CE. The preceding period, therefore, appears to have had greater monsoon intensity.18
In 1999, Ulrich von Rad and five colleagues studied sediments in the Arabian Sea off Karachi and concluded that ‘precipitation decreased in southern Pakistan after 40003500 yr BP’,19 that is, after 2000 BCE, which agrees with the preceding study.
A year later, palynologist Netajirao Phadtare examined pollen and peat in the Garhwal Himalayas (west of the Gangotri glacier) and found evidence of ‘a warm, humid climate, with highest monsoon intensity’ from about 4000 to 2500 BCE; after 2000 BCE, there was ‘a sharp decrease in temperature and rainfall’, reaching a minimum about 1500 BCE. Phadtare cited five other independent studies (not part of our list here) from other regions that support ‘a decrease in the strength of the Southwest monsoon about 4000 cal yr BP’, that is, about 2000 BCE.20
In 2003, M. Staubwasser and three colleagues analysed planktonic oxygen isotope ratios off the Indus delta. Their findings revealed climatic changes during the last 6000 years, ‘with the most prominent change recorded at 4.2 ka BP’,† that is, 2200 BCE, along with ‘a reduction in Indus river discharge’. They observed, ‘The 4.2 ka event is coherent with the termination of urban Harappan civilization in the Indus valley. Thus, drought may have initiated southeastward habitat tracking within the Harappan cultural domain.’21
In 2006, Anil K. Gupta and three colleagues synthesized research on the monsoon and other climatic inputs from many sources including their own. ‘It appears to us,’ they concluded, ‘that the arid phase in the Indian subcontinent started ca 5000-4000 cal yrs BP [calibrated years before present] coinciding with a stepwise weakening of the SW monsoon . . . The arid phase might have intensified ca 4000-3500 cal yrs BP as has been in the Himalayas, western peninsula and northwestern India, and ended ca 1700 cal yrs BP, when the SW monsoon was the driest.’22 Here again, the arid phase starts around 2000 BCE.
In 2008, the archaeologist Rita Wright and two colleagues used models of archaeoclimatology to plot the intensity of the monsoon and river flow in the Harappa region. They found that ‘around 3500 BC the volume of water in the rivers increases, and the rivers flood’, until ‘from around 2100 BC the river flow [in the Beas] begins to fall’. Around Harappa, ‘a 600-year period of reduced rainfall [sets in] after 2100 BC’, leading to ‘an unexpected agricultural crisis’.23 Those two dates roughly bracket the Early and much of the Mature phases.
A CLIMATIC CLIMAX?
More studies could easily be quoted on either side.24 Dealing as they do with different areas of the Indo-Gangetic plains and using different inputs and methods, it is hardly surprising that they should reach differing conclusions. This shows, if at all it were necessary, that palaeoclimatology is a complex field. After highlighting technical problems with sample selection and processing (especially in the case of pollen) and the possibility of human interference in some of the changes noted (especially as regards vegetation), Dorian Fuller and Marco Madella, British and Spanish archaeologists respectively, caution us against drawing hasty conclusions:
Changes in vegetation and hydrology, if present at a given lake, should not be generalized into climatic changes for the whole of Rajasthan, let alone the entire Harappan region . . . There is growing discomfort with simplistic environmentally determined understanding of change.25
More recently, they repeated much the same warning, but acknowledged at the same time the indirect effects of a climatic event from 2200 BCE onward:
Harappan urbanism emerged on the face of a prolonged trend towards declining rainfall. No climatic event can be blamed for a precipitous end of this civilisation, although strategic local shifts in agriculture that may have begun in response to prolonged droughts at ca 2200 BC . . . A climatic event cannot be blamed simplistically for [Harappan] collapse and de-urbanisation, but Quaternary science data make it clear that we cannot accept a view of climatic and environmental stability since the mid-Holocene in the region (as promoted by Possehl . . .).26
Fuller and Madella clearly seek to harmonize the two opposite conclusions outlined above. But what are these ‘prolonged droughts’ taking place around 2200 BCE? They are, in fact, a very widespread phenomenon that affected Egypt and Turkey,27 Mesopotamia (bringing about the end of the Akkadian empire28), large parts of Africa,29 China30 and, even, North America.31 If attributing the end of the Indus civilization to this single event would indeed be ‘simplistic’, to ignore its impact altogether is certainly unreasonable. There may be a grain of truth in the Mahābhārata’s mention of ‘a great drought of twelve years’ or the disappearance of thousands of lakes.
ENVIRONMENTAL ISSUES
Climatic and environmental conditions are two distinct things. Even if we accept that the Harappan climate was moving towards aridity, it does not follow that the ecosystem was as degraded as it is today.
Early archaeologists such as John Marshall had argued that the widespread use of baked bricks and the depiction on Indus seals of animals such as the elephant, the rhinoceros, the water buffalo and the gaur (often misnamed as the ‘Indian bison’) pointed to a moister and greener environment. In reply, it has been observed that those animals were still to be seen in parts of the Indus Valley till recent decades or centuries, and therefore, except for denser gallery forests along the rivers, the environment need not have been markedly different in Harappan times.
Nevertheless, if that were the case, it would be hard to explain the presence at Kalibangan of bone remains of the elephant, the one-horned rhinoceros, the water buffalo, several deer species and the river turtle. For archaeozoologist Bhola Nath, ‘the remains of these animals show that the climate at that time was more humid than the arid climate of present day’.32 To his colleagues S. Banerjee and S. Chakraborty, the occurrence at Kalibangan of the rhinoceros in particular ‘strengthens the geological evidence that the desert conditions of this area are of recent origin’.33 Moving to Gujarat, P.K. Thomas observes that the same animal ‘is identified from a large number of Harappan and Chalcolithic sites . . . [and] inhabited a major part of the Gujarat plains in the protohistoric period . . . The identification of large herbivores like rhinoceros, wild buffalo and probably wild cattle at many of the Gujarat Harappan sites suggests that the ecological conditions were more congenial for animal life during the protohistoric period in Gujarat’.34
Taking all viewpoints into consideration, I propose to strike a middle path and accept Thomas’s more ‘congenial’, lusher conditions in Harappan times together with a wetter climate, gradually moving towards aridity, and culminating in a prolonged drought around 2200 BCE.35
This takes us to the question of human interference: did the Harappans contribute to an ecological degradation
of their environment? Initially proposed by Wheeler, this theory is known as the ‘wearing out of the landscape’ and argues that the Harappans’ industrial activities must have accelerated deforestation: baking bricks or pots, working copper and a host of other activities from the making of faience to plain cooking demand fuel—that is, wood. Mohenjo-daro alone consumed many millions of baked bricks and thousands of tons of timber; construction apart, its population of at least 40,000 souls must have put a considerable strain on the local environment in terms of firewood collection and agriculture.
Nowadays, archaeologists generally disagree that this could have hastened the city’s demise,36 arguing that the rich Indus alluvium would have soon regenerated the forests on its banks: in 1961, R.L. Raikes and R.H.J. Dyson37 calculated that ‘400 acres of gallery forest would have been sufficient for the building of Mohenjo-daro at intervals of about 140 years’.38 But such calculations do not take into account the daily consumption of fuel wood for bronze and pottery industries and for cooking. Also, Walter Fairservis, while endorsing the calculations of Raikes and Dyson, added his own concerning the amount of fodder consumed by the cattle used in Mohenjo-daro, both as a source of food (dairy products and meat) and for ploughing. His conclusion was that
three-quarters [of Mohenjo-daro’s fodder needs] had to be obtained by foraging in the surrounding forests and grass lands. This formidable assault on the indigenous flora most certainly affected the ecology and had an adverse effect on the land and aided the spread of the active floodplain.39
Grazing, and possibly overgrazing, should be added to this picture. Man-made damage could also have been compounded by natural factors: in Mohenjo-daro’s case, deforestation and land degradation would have made annual floods more violent and the Indus more prone to shifts. Indeed, there is evidence that ‘within at least 500 years of existence of the city, the river must have changed its course several times’.40
Also, in a general trend towards aridity, even small shifts in land or water use can tip the scales towards desertification; recent and tragic examples of this have been witnessed in regions as different as the Sahel and Inner Mongolia. Ecological damage inflicted by the intense industrial activity and population concentration of Indus cities could have set off such a vicious circle, or rather a spiral. With all their ingenuity, the Harappans might have taken the land’s bounty for granted—just as we do today.
MOHENJO-DARO AND THE INDUS
Environmental changes apart, rivers have their own whims and fancies.
A thesis first propounded by M.R. Sahni in the 1950s, and expanded a decade later by Robert Raikes and George Dales, proposed evidence of a tectonic uplift that might have dammed the course of the lower Indus, provoking destructive floods that would have engulfed Mohenjo-daro. It rested its case on one such event observed as recently as in 1819, when an earthquake raised a huge embankment called ‘Allah’s bund’ in the northern part of the Rann of Kachchh: it was 3 to 8 m high, over 100 km long and 25 to 30 km wide! This natural dam impounded eastern courses of the Indus, submerging an area of some 5000 km2 and swallowing numerous villages, until it was breached a few years later by the sheer pressure of the waters. Something of the sort could have happened at Mohenjo-daro, they reasoned. And several excavators have reported ‘the presence of massive disruptive floods throughout the history of the city’,41 in the words of George Dales.
But that thesis was challenged in the 1960s, notably by the archaeologist H.T. Lambrick, who found the evidence unconvincing. He proposed that, quite on the contrary, the Indus shifted away from Mohenjo-daro (in a process known as ‘avulsion’‡): ‘The surrounding country, starved of water, immediately began to deteriorate.’42 The deterioration would have stemmed not just from the loss of the river, but also from that of its yearly floods that watered the soil in time for the winter crop and fertilized it with rich alluvium. Worse, the river-based communication network that Mohenjo-daro vitally depended on, in Michael Jansen’s opinion,43 would have been completely disrupted. Louis Flam, a US archaeologist who has done considerable research in Sind, recently elaborated that ‘a major change . . . in the main river channel would have brought widespread abandonment of many sites and a movement of population out of the Lower Indus basin into adjacent and more “stable” areas’.44
These, then, are the two contending theories on Mohenjo-daro and Sind, both unproven, although the second one is now more favoured. But unless we go by the obsolete view of an emperor sitting enthroned at Mohenjo-daro, its abandonment alone need not have brought about the end of the whole civilization; the Sarasvatī basin and Gujarat could have absorbed at least some of the distant impact.
THE DEATH OF A RIVER
If the evidence is inconclusive in the case of the Indus, in the Sarasvatī’s it is far more convincing, and of two kinds.
First, the abandonment of Kalibangan has been dated to around 1900 BCE, although with some imprecision in the radiocarbon samples.45 This town was clearly of importance, and must have depended on the Sarasvatī for water as well as communication; the river’s disappearance would have understandably brought about its end.
At this point, it is sometimes objected that sites away from the Sarasvatī were also abandoned more or less during the same epoch, and therefore some other cause must have been at play. Our second type of evidence disposes of the objection: the radical changes in site distribution occurring between the Mature and the Late phases cannot but be a reflection of hydrological changes. We saw in Figs 6.8 and 6.9 how the section of the Hakra bed close to today’s international border between India and Pakistan was deserted: Cholistan has 174 Mature sites, but just 50 Late ones (a drop of 71 per cent), all of them clustered around faraway Derawar Fort; on the Indian side of the border, northern Rajasthan, with 31 Mature Harappan sites, has none at all of the Late phase, while further upstream in Haryana, over a thousand Late sites mushroom (Table 6.1). The only possible conclusion from the complete absence of Late Harappan sites on either side of the border is that the central portion of the Sarasvatī had stopped flowing.
Not only would this have greatly affected the hundreds of Harappan sites in the region, it would also have had repercussions in Sind: let us recall that the loss of the Sarasvatī was caused partly by the eastward capture of the Yamuna, and partly by the progressive desertion by the Sutlej, which ended up joining the Beas. The latter desertion, envisaged by many scholars since the nineteenth century, is confirmed by archaeological evidence as argued in Chapter 7. But the Beas is a tributary of the Indus, which suddenly found itself swollen by the very waters that the Sarasvatī had lost! ‘As a result,’ write the Allchins, ‘the Indus floods would have become greater in volume and more erratic.’46 Kenoyer makes the same point, linking the Indus’s eastward ‘swing’ to the capture of part of the Sarasvatī’s waters by the Indus system.47 Flam is more specific: ‘The Sutlej River has the highest average annual discharge of all the main Indus tributaries of the Punjab as they exit their mountain catchments and enter the plains’, and therefore ‘an increase in water and sediment discharge of that magnitude [provoked by the westward shift of the Sutlej] would have had dramatic effects downstream in the Lower Indus Basin.’48
This domino effect provides a coherent explanation covering both core regions: the Sutlej’s shift causes the Sarasvatī’s final desiccation and aggravates the Indus floods, which wash away some sites and cut others off or bury them under sediment. This would also explain why so few Late Harappan sites have been found in Sind: only six of them have been identified.49
Whether things happened exactly in that way we will only know after a great deal more exploration. In the meantime, let us stress the consensus among archaeologists on the radical consequences of the Sarasvatī’s disappearance. To those quoted above—Mughal, Possehl, the Allchins, Kenoyer and Flam—we can add many more, among them the following:
B.B. Lal: ‘The obvious result [of the diversion of the Sarasvatī’s waters into the Yamuna system] was the migration
of the [Harappan] people towards the north-east where some water was still available in the uppermost reaches of the Sarasvatī and Ghaggar and further east in the upper plains of the Gangā-Yamunā Doāb.’50
Dilip Chakrabarti : ‘The Sutlej which was the main supplier of water volume through the Ghaggar-Sarasvati-Hakra channel, shifted and joined the Indus River drainage. The Yamuna was likely to have played a [similar] role in the fate of the Drishadvati system51 . . . To a considerable extent the process [of weakening of the political fabric of the Indus civilization] must have been linked to the hydrographic changes in the Sarasvati-Drishadvati system.’52
Jane McIntosh: ‘[The desertion of the Drishadvati and the Sutlej] is typical of the instability of the river courses in the Indus plains—but in the case of the Saraswati, the effect was not localized but devastating on a major scale. Cities, towns, and villages were abandoned, their inhabitants drifting to other regions of the Indus realms and eastward towards the Ganges, pushing back the centuries-old eastern boundaries of Indus culture and venturing into uncharted territory.’53
The Lost River: On The Trail of Saraswati Page 15