A comparison of East-Asia landfall tropical cyclone in recent reanalysis datasets--before and after satellite era

Abstract

Based on different reanalysis datasets, reconstructions of East Asia landfall tropical cyclones (TCs) were compared with observations. The 20th-century reanalysis version 3 dataset (20CRv3) received the most approval in this assessment. It performed better in terms of annual frequency. The fifth generation of atmospheric reanalysis dataset (ERA5) and Japanese 55-year reanalysis dataset (JRA55) are also recommended in this study. Nevertheless, an apparent inconsistency in reconstructed TCs before and after 1980 is visible. Temporally, after the satellite era, the underestimation on TC frequency of the National Centres for Environmental Prediction and National Centre for Atmospheric Research (NCEP/NCAR) reanalysis dataset (NCAR) and 20-century reanalysis of European Center for Medium-Range Weather Forecasts (ERA20C) has been greatly improved. The downward trend of landfalling TCs is well captured by ERA5 and ERA20C. Spatially, the underestimation of TC track discrepancy is reduced in the post-satellite era. ERA5 and 20CRv3 showed relatively consistent performance compared to the former reanalysis in pre-and post-satellite time, which might be due to their better TC treatment. Despite the essential need for high resolution, this study stressed the importance of observation and assimilation development for the reanalysis TCs.