More subversively, administrative texts, being derived from state archives, tend to
support centralised models of irrigation management such as that of Wittfogel ( 1957 ).
In other words, because they only seem to refer to that part of the system that needed
to be recorded for the purposes of the state, cuneiform records cannot therefore be
regarded as providing a picture of the operation of the entire irrigation system (Rost
2010 : 2 ).
Within Mesopotamia there is, unfortunately, a dearth of detailed archaeological data
on the size and layout of Sumerian irrigation systems as well as for actual water control
features. Nevertheless, in addition to texts, the landscape itself provides information on
ancient irrigation, including the canals themselves, alignments of ancient sites along
former channels, and less frequently, evidence of field systems (Wilkinson, T.J. 2003 :
45 – 46 , fig. 5. 7 ). Techniques employed to record such traces include satellite imagery
and air photographs (Hritz 2010 ), detailed contour maps that distinguish ancient levees
(Coles and Gasche 1998 ), digital terrain models (Hritz and Wilkinson 2006 ) and auger
investigations that provide cross-sections of alluvial sediments (Verhoeven 1998 ). In
addition, a wealth of information occurs in consultants’ reports, usually made for
twentieth-century agricultural development projects. Charred plant remains recovered
from excavated archaeological layers are particularly useful because they supply infor-
mation on irrigated crops, weeds of cultivation, as well as on the relative abundance
of water (Neef 1989 ).
Unfortunately, it is not always easy to link the above classes of information together.
For example, although administrative texts might refer to a canal, it is not always easy
to know which of a ganglion of canals on a satellite image might refer to the feature
discussed in a particular cuneiform tablet. Whereas archaeobotanical remains can be
dated by radiocarbon methods, it will not be clear where they were originally culti-
vated; on the other hand, canals and relict channels, although they can be accurately
located within the landscape, are often difficult to date. A particularly serious problem
with the evidence from field archaeology is that much Mesopotamian archaeology was
conducted before palaeobotanical and landscape analysis became commonplace
techniques, and because at the time of writing, it continues to be difficult to gain secure
safe access to Mesopotamia, it is difficult to undertake the appropriate ‘ground con-
trol’. Nevertheless, judicious use of the range of sources discussed above makes it
possible to attempt plausible reconstructions of ancient irrigation systems.
PHYSICAL CONTEXT OF THE HYDRAULIC LANDSCAPE
Sumer developed within an alluvial depression sandwiched between the Zagros
Mountains to the east and the western desert of Iraq to the west (Pournelle this
volume). The alluvium of this sedimentary basin has been deposited by the combined
waters of the Tigris and Euphrates, as well as in the form of marine deposits within
the head of the Persian Gulf, which at around 4000 BCextended some 200 – 250 km
inland of the present coastline near Fao (Ur this volume: fig. 7. 1 ).
The southern part of the Mesopotamian plains that make up the land of Sumer have
been subdivided into four broad soil zones (Buringh 1960 : 121 ):- The floodplain of the Tigris and Euphrates rivers where the sinuous or anasto-
mosing rivers usually flow on low levees with levee slopes ideal for gravity flow
–– Hydraulic landscapes and irrigation systems ––