Statistical Downscaling of Temperature using Stepwise Cluster Analysis Method - A Case Study in Nur Sultan, Kazakhstan

Abstract

In this study, a stepwise cluster analysis (SCA) method is proposed to downscale three temperature variables (i.e Tmean, Tmax and Tmin) from multiple GCMs (i.e., GFDLESM2M, IPSL-CM5R-LR and NorESM1-M) in the City of Nur Sultan, the capital of Kazakhstan. The future trends of temperature projections in the study city and their changes (2075-2100) relative to the baseline period (1979-2004) have been evaluated. Some findings can be summarized: (i) the projected temperatures of Tmean, Tmax and Tmin are presented in consistent increase trends, indicating that the significant warming are likely occurred in the Nur Sultan throughout this century; (ii) the variability of projected temperature is influenced by GCMs and emission scenarios; (iii) for the monthly temperature changes, the magnitudes of monthly temperature changes in the months from November to March are higher than the months from April to October. The above findings can provide decision supports for climate adaptation strategies in the City of Nur Sultan.