Correlation Analysis Between GNSS Water Vapor Time Characteristics And El Nino In Australia

Water vapor,a vital part of the atmospheric system,plays a crucial role in studying the global water cycle,energy budget balance,extreme weather and long-term climate changes.Understanding water vapor's spatial and temporal distribution characteristics are of great significance to studying various climate models.Therefore,gathering high-precision water vapor information becomes a prerequisite.However,conventional water vapor detection technology has limitations in capturing fine variations with low spatiotemporal resolutions.With the rapid technological improvement,ground-based GNSS technology has become a reliable means of collecting water vapor in recent years.Simultaneously,dense GNSS stations have provided a large amount of data support for GNSS precipitable water vapor research.Using GNSS technology to obtain atmospheric water vapor,two key parameters are required for GNSS station pressure(P_s)and atmospheric weighted mean temperature(T_m).While most GNSS monitoring stations are primarily employed for geodetic research,and few GNSS stations are equipped with meteorological instruments.To obtain these two meteorological parameters with high accuracy has become a critical issue in ground-based GNSS water vapor research.Firstly,in this study,the GIPSY software developed by the Jet Propulsion Laboratory was adopted to process the original observation data of 124 stations in Australia from 2015to 2018 provided by Geoscience Australia to collect the tropospheric zenith total delay data of the corresponding stations.For evaluating the quality of the data solution,the data calculated by the GIPSY software was compared with the corresponding data of 46co-located stations in Australia provided by the Nevada Geodetic Laboratory.The variation of the deviation is less than±2mm,and the root mean square error is less than 4.5mm,which manifests that the solution accuracy meets the requirements of GNSS inversion of water vapor.Secondly,the ERA-Interim stratified reanalysis data with the spatial resolution of 0.5°×0.5°and the time resolution of 6h from ECMWF were interpolated to obtain the pressure of the GNSS stations in Australia.Meanwhile,interpolation accuracy was evaluated by the measured pressure at multiple radiosonde stations.The results show that the maximum root mean square error of the interpolated and extrapolated air pressure does not exceed 1.12h Pa,indicating that the interpolated and extrapolated air pressure values can replace the measured pressure values.Thirdly,the atmospheric weighted mean temperature T_mis an essential parameter for calculating the water vapor conversion factor and retrieving precipitable water vapor(PWV).Based on the Global Geodetic Observing System(GGOS)atmosphere T_mgrid data and European Centre for Medium-Range Weather Forecasts(ECMWF)2m T_sgrid data from2007 to 2017,the correlation between grid weighted mean temperature and temperature was discussed in this study.The residuals of the T_m-T_slinear model considering the grid exhibit apparent changes in annual,semi-annual and diurnal cycles.The triangular fitting can be applied to improve the accuracy and practicability of the model.Hence,a weighted mean temperature model(q Tm)suitable for the Australian region was constructed,taking into account the seasonal and diurnal variations of T_mresiduals.Besides,GGOS atmosphere T_mgrid data and radiosonde data in 2015 were adopted to evaluate the model.Compared with the GGOS Atmosphere Tm data,the root mean square error of the q Tm model is only 2.01K,which is 19.4%and 21.8%higher than the Tm accuracy calculated by the GPT2w-1 and GPT2w-5 models,respectively.Compared with the discrete integral Tm data of 28radiosonde data in the Australian region,the root mean square error of the q Tm model is only 2.38K.Compared with the Tm calculated by the GPT2w-1 model and the GPT2w-5model,it has increased by 12.96%and 14.36%,respectively.Meanwhile,to evaluate the accuracy of GNSS inversion of water vapor,compared with the water vapor measured by the nine radiosondes,the water vapor deviation obtained by combining the calculated total zenith delay data with the q Tm model and interpolated meteorological data is within±2mm.The root mean square error is within 3mm,which meets the accuracy requirements of GNSS inversion of water vapor.Moreover,based on the combination of calculated zenith total delay data and the q Tm model and the interpolated GNSS station pressure data,the water vapor information of 124stations in the Australian region from 2015 to 2018 was calculated.Then the time-scale change characteristics of water vapor in the Australian region were studied.The results indicate that the average annual PWV value of the Australian region displays a decreasing trend from the coast to the western inland,and the maximum value appears in the northern coastal area.Affected by the northwest monsoon,the maximum annual amplitude appears in the northwest's coastal areas of low latitudes.The semi-annual amplitude is significantly smaller than the annual amplitude,and the diurnal cycle of water vapor changes little.Except for the savannah climate zone,the diurnal PWV changes in other climate zones are subjected to abnormal changes,with the maximum value occurring at 6-8 generally.Finally,sea surface temperature is one of the essential indicators for monitoring ENSO.The water vapor information calculated by 64 coastal GNSS stations was analyzed to explore the relationship between monthly average PWV and sea surface temperature.The results reveal that the correlation coefficient for tropical sites in Australia is about 85.76%.A1 K increase in SST will lead to an approximately 5.56mm increase in PWV in tropical regions and an approximately 3.30mm increase in the range of 23.26°S?30°S in Australia.A1K increase in SST will result in an approximately 12.25%increase in water vapor levels across 64 GNSS stations.Meanwhile,it will increase the PWV of the tropical zone north of the Tropic of Capricorn in Australia by 15.94%.Therefore,the ENSO event has a considerable impact on PWV.Twelve GNSS stations in low-latitude coastal areas significantly affected by El Ni?o were selected to study the differences of water vapor and total precipitation of intense El Ni?o period and non-El Ni?o period from May 2015 to April2016.Compared with El Ni?o,the results show that the increase in water vapor and total precipitation during the non-El Ni?o period is significant.The increase in water vapor and total precipitation during the non-El Ni?o period at the 12 GNSS stations presents good consistency.