How to Choose the Most Proper Representative Climate Model Over a Study Region? a Case Study of Precipitation Simulations in Ireland with NEX-GDDP-CMIP6 Data

With the aim of providing a multi-criteria decision-support system to capture the spatio-temporal climatological patterns derived from climate models and identify the best representative climate model over each target area, this study developed a toolbox. This toolbox includes (1) climate data from observations and simulations, (2) a broad range of statistical and categorical metrics to quantify the models' assessment, and (3) a multi-criteria decision-making method of the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) as the central engine, where the climate models are investigated and ranked based on the applied evaluation metrics. To make the concept more tangible, the procedure was utilised for the case of Ireland where the effectiveness of precipitation estimates from the new version of the National Aeronautics and Space Administration Earth Exchange (NEX) Global Daily Downscaled Projections (GDDP) is analysed. The applied archive comprises downscaled hindcast projections based on the outputs from the Phase 6 of the Climate Model Intercomparison Project (CMIP6). Using a set of 13 categorical and statistical metrics, 34 NEX-GDDP-CMIP6 models were compared to the reference data from the Multi-Source Weighted-Ensemble Precipitation (MSWEP) dataset in a 25-year period of 1990-2014. A comprehensive evaluation was done at different temporal scales of daily, monthly and annual. The obtained findings illustrate that the reliability of the estimations varies significantly across time and space. The NEX-GDDP-CMIP6 models, best reproducing the climatological and spatio-temporal features of rainfall data in wetter areas of Ireland, do not perform well in the drier zones and vice versa. Therefore, there is a strong uncertainty in choosing the best representative model. As a result, this framework uses the TOPSIS method, prioritizing the applied 34 climate models based on the employed metrics. This toolbox is easily replicable for other case studies, which can be used as a guideline for policy makers, hydrologists, as well as climate scientists for choosing the best climate model over each target area according to the prediction task: water resources allocation, flood and disaster preparedness, ecosystem conservation, agriculture security, public health, infrastructure planning and risk assessment, hydropower energy, coastal management and climate adaptation and mitigation strategies.