Rainfall Downscaling Using Statistical Downscaling Model and Canonical Correlation Analysis: A Case Study

The climate change impacts on different components of water cycle have been revealed in recent years all over the world. Climate change impact assessments require data of spatial and temporal resolutions that are not currently available from the output of General Circulation Models (GCMs). Since GCM outputs are given in low resolution, it is important to downscale them in the desired spatial resolution in real case studies. Different methods are considered for this purpose, varying from simple regression models to complicated physical models. However in practice, it is important to use the methods with the desired accuracy and maximum achievable simplicity. The Canonical Correlation Analysis (CCA) seems to cover both of these requirements based on its application in different fields for developing relationships among dependent (e.g. rainfall) and independent variables (e.g. pressure). In this study, the CCA and other downscaling model namely Statistical Down-Scaling Model (SDSM) are utilized for downscaling GCM outputs in a study area in the central part of Iran. Finally the results of comparing these models show the acceptability of the CCA outputs based on statistical tests. There are no important advantages of SDSM on CCA results but CCA is much simple to use for practical purposes.