Selection of Suitable Precipitation CMIP-5 Sets of GCMs for Iraq Using a Symmetrical Uncertainty Filter
Thair Sharif Khayyun, Imzahim A. Alwan, Ali Mohsen Hayder
Abstract
Abstract Prediction of future water resources in Iraq for the period 2020 to 2100 based on different scenarios of climatic change models by harnessing many calibrated Arc-SWAT models needs to select future precipitation data from suitable sources. Therefore, the selection of the appropriate source of the possible future precipitation time series data sets was studied by examining 20 models of the fifth phase of the Coupled Model Inter-comparison Project (CMIP-5.0) for General Circulation Models (GCMs). The Symmetrical Uncertainty (SU) approach was used to determine the performance of the 20 CMIP-5.0 sets of GCMs against a monthly scale of historical precipitation datasets at each one of 35 rain gauge stations spread throughout Iraq, and an appropriate ensemble of GCMs selected. The examined models were ranked as 1st, 2nd, 3rd, etc. based on the SU values at each station and the final ranking of the models was carried out using a multi-criteria decision-making (MCDM) method. The results showed that the HadGEM2-AO and HadGEM2-ES were the best (1st ranked) models at 31 stations, while the MIROC-5 and CSIRO-Mk-3.6 were the best models at Zakho and Duhok stations, respectively, and the BCC.CSM1.1.m and FIO.ESM were the best models at Samarra and Hilla stations, respectively. There was a variation in simulation preference regionally between the two models HadGEM2-AO and HadGEM2-ES in the second ranking except at Tel-Afer and Najaf stations. No predominant models were found in other ranks throughout Iraq. The results of the final ranking of these 20 CMIP-5 sets by the MCDM method thus showed that there are only four suitable GCMs, HadGEM2-AO, HadGEM2-ES, CSIRO.Mk3.6, and MIROC5 for data projections studying scenarios involving future water resources in Iraq.