Prediction of Hydrological Drought in Semi-arid Regions Using a Novel Hybrid Model

Hydrological drought is one of the most important natural phenomena affecting the various aspects of life on Earth, especially with the increasing impact of climate change around the world. Drought forecasting is one of the important strategies in preparing treatments to reduce the effects of drought on life, especially in water resource management. The process of forecasting hydrological drought is one of the complex issues in hydrology, especially in the semi-arid regions. In the current paper, a hybrid model was developed to predict hydrological drought, using two machine learning models, namely the K-Nearest Neighbor (KNN) and the K-means clusters. The results of the stream drought index for the (3, 6, 9, 12) months-time scale were used to predict the hydrological drought of the study area. The flow data of the Great Zab River in Iraq was used as the case study. The results showed the efficiency of the proposed hybrid model in predicting hydrological drought compared to using the KNN model. The results showed that the results of the hybrid model are better than the results of the KNN model by 85%. Thus, the proposed hybrid model has overcome the difficulty of predicting hydrological drought in semi-arid regions.