Optimal water allocation in irrigation networks based on real time climatic data

The main objective of this study is to improve allocation of water using real time climatic data to estimate irrigation requirement. A study was conducted on an irrigation network in Northwest of Iran to compare present water allocation technique, calculated based on traditional practice of using long-term averaged climatic data, and proposed practice of using real time data with the actual water allocation determined based on specified season's climatic data. In this study, neural network techniques were used to estimate reference evapotranspiration (ETo), actual evapotrasipiration (ETc), and water allocation requirements. For predicting actual evapotranspiration in the subsequent 10-day period, ETo data for one, two, three previous 10-day periods were used. The results of two different neural network techniques were analyzed and compared separately with season specified and long-term averaged ETc. In regard to ETc prediction, the results showed that focused time-delay method is more efficient than feed-forward, both in 10-day period and in monthly scales. In addition, better estimation can be obtained if climatic data from three preceding 10-day periods are used. Overall, incorporating new techniques resulted in 10–25 percent savings on water allocation within the network.

► We employed real time climatic data to estimate irrigation requirement.
► We used neural network techniques to estimate reference and actual evapotranspiration.
► New techniques resulted up to 25% savings on water allocation in the network.