Evaluating the added values of regional climate modeling over China at different resolutions Academic Article uri icon

  •  
  • Overview
  •  
  • Research
  •  
  • Identity
  •  
  • Additional Document Info
  •  
  • View All
  •  

abstract

  • Previous studies have suggested that dynamical downscaling to global climate models can produce improved climate simulations at regional and local scales. However, the expensive computational requirements of dynamical downscaling inevitably add a limit to the spatial resolution of the resulting regional climate simulations. In order to find a balance between computational requirements and simulation improvements, it is extremely important to investigate how the spatial resolution of regional climate simulation affects the added values of dynamical downscaling; yet, it is still not well understood. Therefore, in this study, we conduct long-term climate simulations for the entire country of China with the PRECIS regional climate model at two different spatial resolutions (i.e., 25 and 50 km). The purpose is to evaluate whether a fine-resolution model simulation, given its considerable requirements for computational resources, would add more valuable information for understanding regional climatology than a coarse-resolution model simulation. Our results show that the PRECIS can reasonably reproduce the spatial distribution of seasonal and monthly mean temperature and precipitation over the most of regions in China. However, in the process of downscaling, RCM with higher resolution cannot always produce more accurate output. In regard to precipitation simulations, compared with the host GCM, it is difficult to determine exactly a homogeneous improvement of performance in downscaling, both in terms of spatial patterns as well as magnitude of errors. For interannual variability, variations in temperature are closer to observation than precipitation and the high-resolution R25 has better skills over the northwest than R50. Moreover, except for the west, it is shown that PRECIS is able to better reproduce the probability distribution function of precipitation and some impact-relevant indices such as the number of consecutive wet days and simple precipitation intensity index in spatial distribution.

authors

  • Guo, Junhong
  • Huang, Gordon
  • Wang, Xiuquan
  • Wu, Yinghui
  • Li, Yongping
  • Zheng, Rubing
  • Song, Limin

publication date

  • May 2020