Evaluating the Performance of Selected Reanalysis Datasets over Vietnam

Vietnam, a country highly exposed to extreme weather events, still struggles to fill in significant gaps in its historical meteorological records, limiting a comprehensive understanding of past climate dynamics. The emergence of long-term reanalysis datasets, such as 20CR, ERA5, ERA-20C, and CERA-20C, offers a promising alternative for exploring historical climate variability. This study provides the first comprehensive evaluation of these four reanalysis datasets over Vietnam for the period 1961–2010, by comparing them against station-level observations. The study focuses on their ability to reproduce climatological means, spatial patterns, interannual variability, and extremes of temperature and precipitation. Results show that all datasets reasonably capture temperature climatology, with ERA5 consistently outperforming others. For precipitation, ERA5 again demonstrates superior performance, while 20CR and ERA-20C exhibit moderate skill. In contrast, CERA-20C, performs poorly across most aspects, especially in representing precipitation and related extremes. Overall, despite inherent uncertainties, ERA5 stands out as a reliable long-term reanalysis dataset for climate studies in Vietnam. Nonetheless, caution is needed when interpreting results derived solely from any single reanalysis product. A multi-dataset or index-specific approach remains essential, especially in regions with complex topography and limited observational coverage, such as Vietnam.