Agricultural Drought Risk Assessment Model for Kermanshah Province, Using Statistical and Intelligent Methods

The agriculture sector has been affected by severe drought in recent years, making development of a drought warning system for agriculture crucial. Such a system can be a useful tool for policy makers and investors. This research develops a model for agricultural drought risk assessment using statistical and intelligent methods. Kermanshah province, a major rain-fed region of Iran, was selected as the study area. The model is specific to rain-fed wheat and was updated during the different phonological stages of the growing season. The inputs are a combination of the PDSI, Z-index, CMI, SPI and EDI drought indices which were selected using genetic algorithm and artificial neural networking techniques. The results show that the Z-index better predicts possible losses. The general performance of the model increased toward the end of the growing season, especially after the third stage, when the significance level of the relation reaches 1% and the results become more reliable. Furthermore, linkage of the model to the geographical information system makes it more capable of spatial analysis and more suitable for presentation of the final results.