Performance of ERA5 data in retrieving precipitable water vapor over Hong Kong

When accurately retrieving atmospheric precipitable water vapor (PWV) using Global Positioning System (GPS) data, errors mainly originate from the effects of atmospheric weighted mean temperature (Tm) and ground meteorological data (e.g. surface pressure and temperature). This article assessed the accuracy of interpolated pressure and temperature from the European Centre for Medium-Range Weather Forecasts (ECMWF) fifth generation reanalysis dataset (ERA5), using actual meteorological data in the Hong Kong region from 2016 to 2020 as a reference. Radiosonde data were used to establish a linear equation (HKTm) between Tm and surface tem-perature (Ts), and a fitted equation between Tm, Ts, and water vapor pressure (HKTm-e). Then, these equations were evaluated using the Tm estimated by radiosonde with higher accuracy as a reference. On this basis, the PWV obtained from the measured meteorological data and BTm (GPS MET PWV) was compared with the PWV obtained from interpolated ERA5 meteorological data and HKTm-e (GPS ERA5 PWV), and the overall mean Bias and RMSE were -0.08 mm and 0.28 mm, respectively. In addition, PWV obtained from radio-sonde data (RS PWV) was used to evaluate the accuracy of PWV obtained from ERA5 data (ERA5 PWV) and GPS ERA5 PWV. Then, the GPS ERA5 PWV was used to analyze the seasonal PWV variation and the effect of elevation on PWV in Hong Kong. The accuracy of ERA5 meteorological data is affected by the level of tropical cyclones in Hong Kong. Most of the differences are negative and the reason for the increased accuracy in temperature is explained based on the measured temperature and interpolated ERA5 temperature variations during all high-level tropical cyclones from 2016 to 2020. This article assessed the feasibility of using ERA5 data to replace the missing measured meteorological data during tropical cyclones. Based on the spatial variation of PWV as the tropical cyclone approaches Hong Kong, it is found that the spatial variation of PWV is related to the direction and location of the tropical cyclone. (c) 2023 COSPAR. Published by Elsevier B.V. All rights reserved.