Statistical Downscaling of Wintertime Temperatures over South Korea

Abstract:

Reanalysis data have global coverage and faithfully render large-scale phenomena. On the other hand, regional and small-scale characteristics of atmospheric variability are poorly resolved. In an attempt to improve reanalysis data for regional use, statistical downscaling method is developed based on CSEOF analysis. Low-resolution data are downscaled into a high-resolution data. The developed algorithm is applied to National Center for Environmental Prediction-National Center for Atmospheric Research (NCEP/NCAR) reanalysis data and European Center of Medium range Weather Forecast (ECMWF) ERA-interim reanalysis data to downscale them into a form of Korea Meteorological Administration (KMA) measurements at 60 stations over the Korean Peninsula. The developed downscaling algorithm is evaluated by predicting winter daily temperatures from Nov 17 – Mar 16 for the period of 34 years (1979-2013). For validation of the method, the Jackknife method is used, in which winter daily temperature is predicted over a one-year period not used for training. This procedure is repeated for the entire data period. The mean and variance of the resulting downscaled dataset match well with those of the KMA measurements. Validation results show that correlation increases and error variance decreases significantly at grid points near the KMA stations with and without the seasonal cycle. We will also address the utility of this technique for downscaling model predictions based on future scenarios.