The communities of Sumer therefore had the choice of digging either long canals
up to 40 km (Hunt 1988 : 194 ) with extremely gentle gradients, or rather shorter canals
from the rivers down the slope of the levee (Figure 2. 4 a and b). The latter have
significantly steeper gradients which are more flexible for water distribution to the
fields and which are less likely to silt up or be subject to salinisation. Such channels
would provide modest-size irrigation modules capable of being organised by kin groups
or small-scale communities (Fernea 1970 ; Rost 2010 ; Pollock 1999 : 31 ). In fact, many
traditional irrigation systems illustrated on maps of the southern Mesopotamian plains
adopt this configuration (Figure 2. 5 ). Significantly, short canals of 1. 8 – 2. 2 km would
correspond roughly to total levee widths of 3. 6 – 4. 3 km, figures comparable to the levee
widths of 2 – 5 km estimated by Pournelle ( 2007 ). This is also similar to the width of
distributions of clay sickles used by Henry Wright to estimate the agricultural belts
around Ubaid and Ur in the Ubaid and Uruk periods (Wright 1981 : 324 ).
Overall, it is therefore possible to reconstruct the hydraulic landscape of third–early
second millennium BCSumer as forming three basic zones:
- The levee crest where channels irrigated palm gardens, other orchard crops (some
as a storey below the palms), vegetables and grain fields. - The levee slopes where short canals irrigated fields mainly to grow cereals, perhaps
pulses and onions (Hrusˇka 2007 : 58 ). Together, zones one and two probably
extended some 2 – 5 km away from the levee crest or main channel. - A mosaic of steppe, marshes, fallow and alluvial desert steppe beyond the main
levees. Although cultivation could have been extended into this zone, this would
result in rapid salinisation. This zone would have supplied grazing, fishing, game
birds, reeds and other marshland resources (Pournelle 2007 ).
There was therefore a choice between either long canals that followed the grade of the
plain, but required more labour and were prone to siltation, or shorter steeper canals
that serviced a small area, but which were potentially vulnerable to peak floods that
could rip through off-takes. Because the levee width would limit the area of cereal
production, and territorial boundaries between neighbouring towns and cities would
limit the cultivated territories along the river channels, the total productive area would
have been constrained. Nevertheless a 5 – 6 km wide levee with settlements some 10 km
apart would have been capable of sustaining towns of up to 8 , 000 – 12 , 000 people,
perhaps covering between 40 and 120 ha. During times of social cohesion and political
accord between communities, the ease of transport of bulk products by boat would
have enabled shortfalls of crop production to be addressed by transporting grain from
neighbouring communities (Wilkinson et al. 2007 : 188 ). However, if political relations
with city states upstream or downstream soured, the existing site territory may have
been insufficient to supply the towns and there would have been a temptation either
to violate fallow in order to provide crops every year (Gibson 1974 ), or to build longer
canals, thereby increasing the cultivable area. However, if such canals were extended
into the lower flood basins, there would have been the risk of salinisation. In other
words, greater political unity would have enabled more boat traffic to take place
between communities as was the case in the Ur III period, a period when it was also
easier to mobilise grain for the payment of the balatax (Steinkeller 1987 ; Algaze 2008 :
56 – 57 ; Widell 2008 ).
–– Hydraulic landscapes and irrigation systems ––