Analysing The Temporal And Spatial Distribution Characteristics Of Precipitable Water Vapor Based On Multi-source Data In Japan

Atmospheric precipitable water vapor content is an important part of the Earth's atmosphere,and it plays an important role in the global water vapor cycle,atmospheric radiation,energy balance,the formation of severe weather,and climate changes.Therefore,it is necessary to obtain precipitable water vapor information with high precision and high temporal and spatial resolution.The traditional atmospheric precipitable water vapor observation technology suffers from low space-time resolution and expensive observation equipment.In recent years,due to the rapid development of GPS technology,GPS measurement accuracy can greatly meet the needs of social development.Ground-based GPS has become one of the important means to obtain PWV quickly,efficiently,and accurately.In addition,dense GPS stations provide a large amount of data support for GPS precipitable water vapor research.In the process of using GPS to obtain PWV,it requires accurate pressure P_s,temperature T_s and atmospheric weighted mean temperature T_m information.For GPS stations without weather sensors,how to accurately obtain these two meteorological parameters has become the key issue in GPS precipitable water vapor research.In the study,the ERA-Interim stratified reanalysis data with the spatial resolution of 0.25°×0.25°and the time resolution of 6h from ECMWF is used to interpolate the pressure and temperature of the GPS stations in Japan.Comparing the interpolation results with the radiosondes co-located with GPS stations,it is found that most biases of the interpolated P_sand T_s are less than 1hpa and 3K,respectively,indicating that it is feasible to obtain GPS-PWV by interpolated P_s and T_s.Weighted mean temperature model is important to improve the accuracy of the estimation of the GPS precipitable water vapor.A weighted mean temperature model was established using the ERA-Interim data from 2014 to 2018 for the Japanese region in this study.The relationship between T_m and temperature T_s was analyzed,and a linear weighted mean temperature model(JTm),suitable for the Japanese region was established.Its accuracy is improved by about 16%compared to the Bevis model.Analyze the residuals of the JTm linear model and find that there is seasonal variation in the residual time series.Therefore,the JTm model is seasonally corrected,and a new model(JTm-I)that takes into account the effects of seasonal changes,is established.Its accuracy is improved by approximately 37%and 25%compared to the Bevis model and the JTm linear model.Comparing the Bevis model,the JTm linear model and the JTm-I model with the radiosonde integral Tm,it is found that most biases of the JTm-I model are less than 2K,which is significantly better than the Bevis model(5k)and JTm linear model(4k).Then,the interpolated meteorological data and JTm-I model data were used to obtain the GPS-PWV.Comparing GPS-PWV with radiosonde station RS-PWV,most biases are less than 2mm,which meet the accuracy requirements.Using GPS station meteorological data interpolated by ERA-Interim reanalysis data and the JTm-I model values,combined with the total delay data to estimate the precipitable water vapor content in Japan from 2014 to 2018.The temporal distribution characteristics of precipitable water vapor in Japan was studied.The results show that the annual average PWV content from southwest to northeast in Japan is relatively rich and the difference is not significant.The higher the altitude,the lower the annual average PWV content.The PWV content is the most in summer,which some stations can reach 47.04mm,and it is the least in winter,only 1.92mm.Besides,the percentage variability of the spring-summer is the largest,about 32%-42%,and the winter-spring is the smallest,about 18%-23%.The daily change of PWV is not obvious,and the maximum daily content is around UTC 6h.At the same time,the radiosonde data was used to estimate the stratified precipitable water vapor content in Japan from 2014 to 2018,in order to analyze the spatial distribution characteristics of precipitable water vapor in Japan.The results show that the cut-off height of PWV in Japan is around 10km.The spatial distribution of PWV is affected by the time scale and topography.The content of PWV spatial distribution is more abundant in summer(120mm/km)than in winter(50mm/km).Besides,the spatial distribution of PWV in two adjacent days is very different,and the maximum can reach nearly 60mm/km.The greater the relief,the greater the rate of change of PWV spatial distribution.At the same time,with the increase of latitude,the content of PWV spatial distribution is less.