Assessing climate change impacts on Taleghan reservoir daily inflow using data fusion method

In this research، Data Fusion (DF) method was applied to simulate the hydrological process of Taleghan reservoir daily inflow. Two different DF algorithms were proposed and assessed based on K-nearest neighbors (KNN) algorithm. Four artificial neural network models and two Hammerstein-Wiener (HW) models were used as the individual simulation models. Comparison of the results between individual models and DF algorithms revealed the superiority of the DF method. The performances of the two DF algorithms were comparable in simulating monthly mean inflow values، but AL1 overestimated the monthly standard deviations. Then، the daily time series of Temperature and Precipitation were generated by a well-tested weather generator model and were used as the inputs to the individual models. The results showed that the individual models can result in different or even inconsistent variations under climate change scenarios. It was also revealed that the performance of the AL2 data fusion algorithm was proved by the best HW model and this algorithm resulted in more logical results. Moreover، regarding considerable diversity among the individual models، the DF method can increase the reliability of the simulations related to the predicted variations of reservoir daily inflow under climate change scenarios.