The performance of AR4 models in simulating climate parameters of temperature and precipitation with artificial neural network (Case study: Qara-Su watershed)

The increase in the concentration of greenhouse gases in the atmosphere due to human activities such as changes in usage and use of fossil fuels has led to global warming and global energy imbalance. This increase in greenhouse gases has caused a phenomenon called climate change. In this research, the performance of 4 GCM models named HADCM3, CGCM3T63, 5.CSIROMK3, NCARCCSM3 (from the set of AR4 models) under scenario A2 in simulating the climate parameters of average temperature and precipitation in the Qara-Su basin using artificial neural network. ANN) were evaluated. The forward perceptron model was used to train the artificial neural network. According to the evaluation of the performance of the models by using the coefficients of the maximum absolute error, the average absolute value of the error, the root mean square and the coefficient of explanation, among the set of AR4 models, on average, the NCARCCSM3 model has the best performance in simulating the climatic parameters of the temperature of the Qara-Su area. This model together with CGCM3T63 has the least difference with the observed temperature climate parameter, while the CGCM3T63 model has the least difference with the observed precipitation climate parameter. Also, the results showed that the CSIROMK3.5 and NCARCCSM3 models have the biggest differences with the climatic parameters of temperature and observations, respectively. According to the results of the neural network, the coefficient of explanation for the two climatic parameters of temperature and precipitation are on average 0.97 and 73. 0 was obtained for the entire domain, which indicates the accuracy of the neural network in simulating this parameter.