Indus civilization may not have depended on rivers

[Duncan Head]

An important former river-valley, home to several Indus sites, may have dried up before the settlements were founded:

http://www.bbc.co.uk/news/science-environment-42157402

https://www.nature.com/articles/s41467-017-01643-9

[Patrick Waterson]

Or as the authors say:

Here we analyse the sedimentary architecture, chronology and provenance of a major palaeochannel associated with many of these settlements. We show that the palaeochannel is a former course of the Sutlej River, the third largest of the present-day Himalayan rivers. Using optically stimulated luminescence dating of sand grains, we demonstrate that flow of the Sutlej in this course terminated considerably earlier than Indus occupation, with diversion to its present course complete shortly after ~8 ka. Indus urban settlements thus developed along an abandoned river valley rather than an active Himalayan river. Confinement of the Sutlej to its present incised course after ~8 ka likely reduced its propensity to re-route frequently thus enabling long-term stability for Indus settlements sited along the relict palaeochannel.

This of course depends upon one major assumption:

Here we reconstruct the chronology of a major late Quaternary avulsion in the Himalayan foreland and evaluate its role in urban settlement patterns of the Bronze-age Indus Civilisation (~4.6–3.9 ka B.P.).

i.e. that the civilisation was younger than the 'avulsion' of the river.  I am not sure that the dating methods employed are sufficiently reliable, or perhaps more accurately, compatible, to allow that conclusion.