Earlier this year Greg Laden wrote in his Scientific American blog:
The so called “Hydraulic Hypothesis” is an idea first fully characterized by the historian Karl Wittfogel. His original idea was part of a larger model for the origin of civilization that we see today as having several problematic aspects, but the key idea is still valid. If agriculture is the basis for a society, and it is carried out in a semi-arid region, then the management of water through various forms of irrigation and the centralized control of the agricultural cycle lends itself to centralized despotic leadership. or at least, some kid of cultural and social change allowing for organized effort to predominate over individual self interest. (In fact irrigation based systems have emerged without despotic leadership, and complex society has emerged absent a hydraulic beginning, so this is an oversimplification, just so you know.) But in its simplest form we can correctly say that the emergence of stratified, hierarchic, complexly organized societies was often linked in no small part to the emergence of organizational (and technological) solutions to growing food where there is not enough rain at the right time of year. There is a great advantage to growing food in this manner. The crops become, in essence, invasive species, because human activity provides the crops with a leg up on all the other plants in the region. A plant that in wild form is found primarily in limited microhabitats, out competed everywhere else by more arid-adapted plants, suddenly has a free ride across a vast landscape. Despite the fact that the Hydraulic Hypothesis is an oversimplification, we can appreciate the fact that the beginnings of human “civilization” (as a social and economic system, which we retain today by and large) is linked partially but importantly to managing water to grow food.