Assessment of the performance of CMIP6 model for analysis of temperature and precipitation in Urmia Lake basin

In this research, the performance of CMIP6 general circulation models has been evaluated in the simulation of temperature and precipitation in the catchment area of Lake Urmia. The models used are MPI, CMCC, INM and MRI with a spatial resolution of 100 km. The performance of the models was evaluated before and after the bias correction. To correct the bias of the models, the statistical method of the delta factor change and the data of seven synoptic stations at the basin level have been used. For this purpose, the studied period (1990-2013) was divided into two twelve-year periods for calibration and verification. In this way, the accuracy of the corrected data of the models was evaluated using the delta factor change for the period of 2002-2013 compared to the observational data by using R2, RMSE, NRMSE indicators and Taylor and scatter diagrams. The results showed that for the entire basin, there is a strong linear relationship between the corrected data of models and station observations for temperature and a weak linear relationship for precipitation. The RMSE and NRMSE indices indicate the high accuracy of the models in simulating the monthly temperature and relatively weaker accuracy in simulating the monthly rainfall of the basin, and among the models, CMCC had the highest error. The spatial distribution maps of NRMSE for the total monthly averages showed that all four models have simulated temperature with high accuracy, but the simulation of precipitation in some areas of the basin does not have acceptable accuracy. Therefore, the performance of the studied models was evaluated as good in temperature simulation and poor in precipitation simulation, especially in the CMCC model.