At the end of the third millennium the strong flow from the Sarasvatī dried up... This [the drying up of the Sarasvatī towards the end of the third m… - Gregory Possehl

It seems that during the Indus Age the Sarasvati was a large river and that water that now flows in the Yamuna and/or Sutlej Rivers made it so. Over time these waters were withdrawn and the Sarasvati became smaller, eventually dry. The agency for these changes was the tectonic reshaping of the doab [interfluve] separating the Yamuna from the rivers of the Punjab.

Over the course of the third and second millennia, the Sarasvati dried up.

Of relevance here is Possehl's observation (1977) of the "extraordinary 'empty spaces' between the Harappan settlement clusters," as well as "the isolated context for a number of individual sites" (546). He proposes that "pastoral nomads, or other highly mobile (itinerant) occupational specialists filled in the interstices," since such spaces are un- likely to have been unoccupied. He goes so far as to suggest that "pastoralists formed the bulk of the population during Harappan times since there do not seem to be any settled village farming communities there" (547). Pastoralists and farmers coexisted "not . . . as isolated from one another, but as complementary subsystems: two aspects of an integrated whole. One relied on the intensive exploitation of plants and arable land, the other on the extensive exploitation of animals and pastures" (547). Moreover, "the presence of pastoralists makes very good sense if we see them as the mobile population which bridged the gap between settlements as the carriers of information, as the transporters of goods, as the population through which the Harappan Civilization achieved its remark- able degree of integration" (548).