ESTIMATING WATER DEMAND AT IRRIGATION SCHEME SCALES USING VARIOUS LEVELS OF KNOWLEDGE: APPLICATIONS IN NORTHERN TUNISIA

Abstract

ABSTRACTThe precise estimation of water demand at large‐scale irrigation perimeters is a key requirement for water management. Most methods for estimating water demand use biophysical models and cropping patterns but do not account for the wide range of actual farming practices. As a result, these methods only allow estimation of theoretical water demand. The objective of this study was to analyze different methods for evaluating irrigation water demands using several levels of knowledge about irrigation perimeters such as crop water requirements, cropping pattern, irrigation techniques, and water application depths (WAD). Most of this knowledge was related to farming practices, which can vary. Farm typology is an appropriate way to account for the variability in farming practices. These methods were applied in an irrigation scheme in northern Tunisia. The perimeter was subdivided into two sectors: the first sector was designed for surface irrigation and the second for sprinkler irrigation. The first method estimated water demand at the scheme scale based on actual cropping patterns and assuming that farmers used the planned irrigation techniques and satisfied plant water requirements (L1). However, the farmers did not use the planned irrigation techniques. Instead, they combined surface, sprinkler, and drip irrigation techniques depending on the crops they were growing.The second method examined current irrigation techniques using the results of farm typology that provided the areas under each crop irrigation system (L2). The two methods provided a theoretical water demand model at the scheme scale. However, water demand also needed to account for actual WADs (L3). Field surveys were performed to evaluate the WAD of each crop irrigation technique. Methods L1 and L2 showed similar estimations of water demand at the scale of the perimeter, and the savings in water obtained by converting to drip irrigation were balanced by a reduction in sprinkler irrigation due to conversion to surface irrigation. In contrast, the L3 method estimated 50% lower water demand than that provided by L1 or L2, since the supply of water was insufficient. The irrigation network was designed 50 years ago and no longer corresponds to the needs of current farmers. Farm typology was also shown to be a method that accounted for farm diversity; water demand estimated based on typology differed by only 5% from that obtained by summing the needs of individual farms. Copyright © 2011 John Wiley & Sons, Ltd.