LAUSR

The accurate simulation of precipitation extremes in climate models is important for the understanding of future climate change and its impacts. Due to the poor coverage and the lack of uniformity of observational precipitation extreme data in Asia, it is difficult to evaluate whether existing climate models can accurately reproduce the patterns of and changes in precipitation extremes. The present study presents a comparison between 3 observational data sets and CMIP5 models and shows that improving the data coverage in the future is of great significance for understanding precipitation extreme changes in Asia. The 3 observational data sets show good consistency with the models in the midand high-latitude regions of Asia, where the data generally have good coverage. Meanwhile, the differences between the observations are obvious in the lowlatitude region, and the differences between the models are also large. Although these discrepancies may be related to different operational orders and interpolation methods, the spatial coverage of data can constitute an important factor leading to such differences at low latitudes. The results for the mid-latitude region show little change in the new Asian observation data set (ADEX) compared with in HadEX2, but the trend of precipitation extreme intensity indices in ADEX are closer to the simulated trend than in HadEX2 for the high-latitude region. Furthermore, the differences in the intensity indices between the observations and models are larger than those in the frequency indices, indicating that the simulated and projected values of the intensity indices are higher than those of the frequency indices. In summary, there are differences between the 3 datasets and CMIP5 models, especially for low-latitudes and the intensity indices.