Quantitative qualitative prediction of Khorramrud River discharge due to climate change with Neurosolution model and support vector regression

Awareness of water resources quality is one of the most important requirements in planning and developing water resources and protecting it. Rivers are of particular importance as the main source of drinking, agriculture and industry needs. In this study, in order to investigate the long-term effects of temperature and precipitation from the output of the four climatic models of the fifth report (AR5) IPCC under RCP4.5, RCP6 scenarios for the base period was extracted and compared with the observational climatic data. Daily values of precipitation climatic parameters and temperature of superior models for future periods 2020-2052 and 2085-2053 were produced by LARSWG6 Downscaling method in the study area. In order to predict runoff, artificial neural network model was used in Neurosolution software. After calibration of the model in the base period of 1983-2015, the prediction was made for future periods and then in order to predict the water quality parameters of the river, the support vector regression model was used in the Python programming environment. The results showed that the amount of precipitation in CanESM2 and MIROC-ESM-CHEM models increased and decreased in the rest. Temperature increases in all models and climatic scenarios with the highest value of 3.32 °C in the next second period in the GFDL-CM3 model. The average reduction of models and climatic scenarios resulting from Neurosolution model output in the second period was 9.29% compared to the first period. Changes in the amount of discharge affected the quality parameters of the river, so the values of parameters, TDS, EC, CL, Mg Ca and the sum of anions were predicted according to the number of appropriate observational data.