Dynamical downscaling is one of the major approaches to obtain finer-scale weather and climate information. Dynamical downscaling has a broad application prospect and needs in meteorology, hydrology, agriculture, ecosystem, wind power, and atmospheric environment studies and projections. The traditional dynamical downscaling (TDD) of future climate employs a continuous integration of regional climate model (RCM) where GCM data are used to provide initial conditions, lateral boundary conditions. It is known that GCMs are not perfect and all simulations suffer from systematic biases to a certain extent. The TDD approach certainly brings GCM biases into RCMs through the lateral boundary of the RCMs and degrades the downscaled simulation.

Given this reason, Dr. XU Zhongfeng and YANG Zong-liang from the Institute of Atmospheric Physics, Chinese Academy of sciences developed an improved dynamical downscaling (IDD) method. The GCM climatological means and the amplitudes of interannual variations are adjusted based on the NCEP–NCAR global reanalysis products (NNRP) before using them to drive RCM. The comparison of four 30-year hindcast downscaling simulations suggested that the IDD (WRF_CAMbc_std) greatly improves the downscaled climate in both climatological means and extreme events relative to the TDD approach (Fig. 1).

Like GCMs, RCMs can also contain significant systematic biases. XU and YANG further improved the IDD approach by introducing spectral nudging techniques during RCM integration. Spectral nudging introduces the effects of GCM bias corrections throughout the RCM domain rather than just limiting them to the initial and lateral boundary conditions, thereby minimizing climate drifts resulting from both the GCM and RCM biases. The comparison of eight 30-year hindcast downscaling simulations suggested that both the GCM biases and RCM biases need to be constrained during the dynamical downscaling simulations. The new dynamical downscaling approach (WRF_CAMbc.Ng or WRF_CAMbc.Nglow1 in Fig.2) developed by XU and YANG (2015) shows the best performance among all GCM-driven downscaling simulations. In contrast, the traditional dynamical downscaling approach (WRF_CAM) did not employ any technique to constrain GCM and RCM biases, thereby generally showing the largest errors (Fig.2). 

Fig.2 Annual mean RMSEs of downscaled (a) air temperature, (b) geopotential height, (c) wind vector, and (d) specific humidity. The RMSEs are computed between the dynamical downscaling simulations and NARR over the validation region over 1981–2010. WRF and CAM are the RCM and GCM used in this study, respectively. bc represents the GCM biases was corrected. Ng indicates spectral nudging was used during RCM integration. low1 and low2 denotes that the nudging strength was reduced by one and two order of magnitude relative to the default nudging strength. NNRP represents the NCEP-NCAR reanalysis data.  (Image by IAP) 

The studies have been published in J. Climate and J. Geophys. Res. Atmos.. 

References:
Zhongfeng Xu and Zong-Liang Yang, 2012: An Improved Dynamical Downscaling Method with GCM Bias Corrections and Its Validation with 30 Years of Climate Simulations. J. Climate, 25, 6271–6286.
Zhongfeng Xu and Zong-Liang Yang, 2015: A new dynamical downscaling approach with GCM bias corrections and spectral nudging. J. Geophys. Res. Atmos., doi:10.1002/2014JD022958.

Contact:
Dr. XU Zhongfeng
E-mail: xuzhf@tea.ac.cn