Statistical downscaling and projection of future temperature change for Tabriz city, Iran.

Abstract In the 21st century, Climate change has become one of the prominent global challenges which threats the world, and the changes in climate extremes are estimated to have catastrophic consequences on human society and the natural environment. To overcome the spatial-temporal inadequacy of the GCMs, Linking large-scale General Circulation Model (GCM) data with small-scale local climatic data highly comes to the fore. In this paper, two statistical downscaling techniques encompass LARS-WG and SDSM were employed for assessing the fluctuations of temperature predictand in Tabriz city, Iran. In order to choose the well-response GCMs a Multi-GCM ensemble approach was utilized by EC-EARTH, HadCM2, MIROC5, MPI-ESM GCMs from the CMIP5. To study the impact of climate change over the region, the periods of 1961-1990 and 1991-2005 were used as the baseline and validation period, respectively. Results of evaluation criteria disclosed the superior performance of Multi-GCM ensemble for predicting temperature predictand compared to single GCM models. Furthermore, the result of climate projection for the temperature predictand by both models revealed that the city will experience an increasing trend in temperatures for the horizon of 2021-2080. The average temperature will increase by 2.9 and 3.7 (°C) under Representative Concentration Pathways (RCPs) (i.e., RCP4.5 and 8.5), respectively.