Analysis and prediction of extremes temperature in the cities of central Iran Using the model Artificial neural network

The aim of the current research is to study and predict hazardous extreme temperatures in some cities of central Iran, for this purpose the minimum and maximum temperature data of fifteen meteorological stations (Cities: Isfahan, Shahreza, Natanz, Nain, Ardestan, Semnan, Shahroud, Garmsar, Damghan, Yazd, Bafaq, Gariz, Meibod, Qom and Salafchagan) in the study area in the time period (1999 - 2019) using the innovative method of hybrid artificial neural network and ANFIS adaptive neural network models were used. And Finally, Topsis and Saw multi-variable decision-making models were used to prioritize more exposed areas of temperature increase. The results of this study showed that according to ANFIS modelling model for predicting station temperatures, the lowest mean educational error and the average error of validation for the minimum temperature, with a value of 0.010 for the station Yazd and 1.66% for Damghan station. And the lowest mean educational error and the mean error of validation for the maximum temperature curve were obtained for 0.016 for Garmsar station and 9.39% for Shahroud station, respectively. The maximum temperature fringe based on the Topsis model of two stations of Garmsar and Bafgh with a percentage of 1 and 0/96, will be in higher priority with increasing temperature. Based on the Saw model, Garmsar and Salafchegan stations with the highest percentages 1 and 0.98, respectively, were exposed to higher temperatures.