Evaluation of <scp>ERA5</scp> Precipitable Water Vapour Data in China With Radiosonde Data

ABSTRACT Precipitable Water Vapour (PWV) plays a critical role in the hydrological cycle and is closely linked to global climate change. Owing to its high temporal and spatial resolution and wide availability, ERA5 PWV data have been extensively utilised in climatology, ecology, and weather forecasting. However, influenced by geographic and climatic factors, the accuracy of ERA5 PWV data varies considerably across regions, necessitating thorough evaluation prior to rigorous application. Based on the observed data from 91 radiosonde stations in China for the period 1979–2019, we evaluated the accuracy of ERA5 PWV data across China using a combination of methods. The results indicate that although ERA5 data generally agree with the magnitude and spatial distribution of radiosonde PWV, regional discrepancies remain evident. Specifically, ERA5 tends to underestimate PWV across China, with more pronounced underestimation in the southeastern regions and the Qinghai‐Tibet Plateau. Trend analysis using the Mann–Kendall test reveals a significant improvement in the agreement between ERA5 and radiosonde data after 2002. Over the study period, the PWV trend increased at a rate of 0.178 mm/a for radiosonde data and 0.095 mm/a for ERA5 data. The growth rate was notably higher across more than 80% of northwest China. We further proposed three sub‐regions of China based on five accuracy metrics. The analysis showed that although ERA5 data accuracy improved in most regions after 2002, the southeastern Qinghai‐Tibet Plateau continued to exhibit significant discrepancies, highlighting the need for further improvement. Our findings underscore the importance of region‐ and period‐specific calibration of ERA5 PWV data and will help to ensure the accuracy and reliability when it's used in climate‐related studies.