论文部分内容阅读
Water vapor is the most active component of the atmosphere, playing a significant role in formation of severe weather such as thunderstorms and hail. Therefore, accurate water vapor information is of great significance for studying its temporal and spatial variation and regional weather forecasting. Obtaining atmospheric PWV from GNSS observations requires two key meteorological parameters weighted mean temperature and surface pressure. Meteorological parameters are typically measured by sensors deployed at the GNSS station. However, a large number of GNSS stations are mainly used for positioning, and are not equipped with meteorological sensors, so that the water vapor cannot be accurately estimated. In this paper, the GPT2w model and the method of meteorological data interpolation are used to simulate the meteorological data observation. It is found that the GPT2w model and the meteorological data interpolation can accurately retrieve the water vapor content, but the method of interpolating meteorological data is more accurate than the GPT2w model. In addition, the paper also analyzes the interpolation accuracy from four aspects: the number of interpolated stations, the distance between the interpolated stations, the height difference between the interpolated stations, and the weather conditions. It is found that when the height difference between the interpolated and target stations is small, the meteorological data obtained by interpolating from a short distance station is more accurate. In terms of weather conditions, the interpolated value of sunny day is more accurate; the interpolation accuracy is not directly related to the number of interpolated stations used.