Atiyeh Bozorgi; Abbas Roozbahani; Seied Mehdy Hashemy shahdany
Abstract
Agriculture is one of the most important economic sectors of Iran and the largest consumer of surface and underground water resources. Since various risks threaten agricultural water supply systems, developing a risk analysis framework is inevitable in order to identify and assess the impact of the aforementioned ...
Read More
Agriculture is one of the most important economic sectors of Iran and the largest consumer of surface and underground water resources. Since various risks threaten agricultural water supply systems, developing a risk analysis framework is inevitable in order to identify and assess the impact of the aforementioned systems on promoting sustainable agricultural development. Therefore, for the first time, the present study has attempted to develop a drought risk analysis model for agricultural water supply system with the Bayesian network. The structure of this model, which consists of nodes and their interactions, is designed by using river discharge, water released from the upstream water diversion system, groundwater and irrigation network demand to provide system risk. In order to investigate the capability of the developed model, the risk analysis of the modern Roodasht Irrigation System in Isfahan was investigated. The risk values of the Roodasht agricultural water supply system were estimated in the range of zero to 58.11% and an average of 39.82%. The results showed that this model has good accuracy and performance in both training and test sections, with the root mean squares error of 0.09 and 0.1, coefficient of determination of 0.85 and 0.75, and overall index of model performance of 0.82 and 0.74. The results of this research and the proposed model helps operators and decision makers to better plan the allocation of irrigation water based on the risks predicted in hydrological drought conditions.