Forecasting GNSS Zenith Troposphere Delay by Improving GPT3 Model with Machine Learning in Antarctica

Antarctica has a significant impact on global climate change. However, to draw climate change scenarios, there is a need for meteorological data, such as water vapor content, which is scarce in Antarctica. Global navigation satellite system (GNSS) networks can play a major role in overcoming this problem as the tropospheric delay that can be derived from GNSS measurements is an important data source for monitoring the variation of water vapor content. This work intends to be a contribution for improving the estimation of the zenith tropospheric delay (ZTD) obtained with the latest global pressure–temperature (GPT3) model for Antarctica through the use of long short-term-memory (LSTM) and radial basis function (RBF) neural networks for modifying GPT3_ZTD. The forecasting ZTD model is established based on the GNSS_ZTD observations at 71 GNSS stations from 1 January 2018 to 23 October 2021. According to the autocorrelation of the bias series between GNSS_ZTD and GPT3_ZTD, we predict the LSTM_ZTD for each GNSS station for period from October 2020 to October 2021 using the LSTM day by day. Based on the bias between LSTM_ZTD and GPT3_ZTD of the training stations, the RBF is adopted to estimate the LSTM_RBF_ZTD of the verified station, where the LSTM_ZTD represents the temporal forecasting ZTD at a single station, and the LSTM_RBF_ZTD represents the predicted ZTD obtained from space. Both the daily and yearly RMSE are calculated against the reference (GNSS_ZTD), and the improvement of predicted ZTD is compared with GPT3_ZTD. The results show that the single-station LSTM_ZTD series has a good agreement with the GNSS_ZTD, and most daily RMSE values are within 20 mm. The yearly RMSE of the 65 stations ranges from 6.4 mm to 32.8 mm, with an average of 10.9 mm. The overall accuracy of the LSTM_RBF_ZTD is significantly better than that of the GPT3_ZTD, with the daily RMSE of LSTM_RBF_ZTD significantly less than 30 mm, and the yearly RMSE ranging from 5.6 mm to 50.1 mm for the 65 stations. The average yearly RMSE is 15.7 mm, which is 10.2 mm less than that of the GPT3_ZTD. The LSTM_RBF_ZTD of 62 stations is more accurate than GPT3_ZTD, with the maximum improvement reaching 76.3%. The accuracy of LSTM_RBF_ZTD is slightly inferior to GPT3_ZTD at three stations located in East Antarctica with few GNSS stations. The average improvement across the 65 stations is 39.6%.

GEODESY, CARTOGRAPHY AND AERIAL PHOTOGRAPHY

The aim of this work is to evaluate the accuracy of determining the wet component of zenith tropospheric delay (ZTD) from GNSS-measurements and the accuracy of determining the hydrostatic component according to the Saastamoinen model in comparison with the radio sounding data as well. Zenith tropospheric delay is determined mainly by two methods - traditional, using radio sounding or using atmospheric models, such as the Saastamoinen model, and the method of GNSS measurements. Determination of the hydrostatic component of the zenith tropospheric delay was performed by radio sounding data obtained at the aerological station Praha-Libus in 2011-2013 and in 2018. Data were processed for the middle decades of January and July of each year at 0h o’clock of the Universal Time. The wet component was calculated from GNSS observations. By a significant number of radio soundings at the Praha-Libus aerological station, hydrostatic and wet components of zenith tropospheric delay (ZTD) and the same number of ZTD values derived for the corresponding time intervals from GNSS measurements at the GOPE reference station were determined. The values of the wet component of ZTD were determined and compared with the corresponding data obtained from radio soundings. We found that the error of the hydrostatic component in winter does not exceed 10 mm in absolute value, and in summer it is approximately 1.5 times smaller. This is due to differences in the stratification of the troposphere and lower stratosphere in winter and summer. As for the wet component of ZTD, its errors do not exceed: in winter 15 mm, in summer – 35 mm. The resulting differences in summer have a negative sign, indicating a systematic shift, and in winter – both negative and positive. Today, there are many studies aimed at improving the accuracy of determining zenith tropospheric delay by both Ukrainian and foreign authors, but the problem of the accuracy of the hydrostatic component remains open. The study provides recommendations for further research to improve the accuracy of zenith tropospheric delay.

Joint Correction of Tropospheric and Orbital Errors in SAR Differential Interferograms

Synthetic Aperture Radar (SAR) satellite imagery is a source of data widely employed in the quantification and analysis of an earthquake coseismic displacement. However, due to the signal path along the atmosphere and to other sources, the interferometric phase becomes compromised. In this work, a methodology for the correction of tropospheric and orbital errors in the differential interferogram is presented. This methodology was applied to a couple of Sentinel-1A data. The phenomenon studied was the 11th November 2018 Zeribet el Oued earthquake, Mw. 5.2 (The state of Biskra, South East of Algeria). It was possible to correct both tropospheric and orbital errors, where the dominant one was the tropospheric delay, a displacement error of 4 cm was added to the differential interferogram by this noise source. The correction of orbital error led to a better interpretation of the coseismic displacement.

Atmospheric phase delay correction of PS-InSAR to Monitor Land Subsidence in Surabaya

Atmospheric phase delay is one of the most significant errors limiting the accuracy of Interferometric Synthetic Aperture Radar (InSAR) results. In this research, we used the Generic Atmospheric Correction Online Service for InSAR (GACOS) data to correct the tropospheric delay modeling from the persistent scatterers’ InSAR monitoring. Eighty-one (81) Sentinel-1A images and tropospheric delay maps from GACOS monitored land subsidence in Surabaya city between 2017 and 2019. InSAR processing was carried out using the GMTSAR software, continued with StaMPS and TRAIN, which were used to correct the tropospheric delay of PSInSAR-derived deformation measurements. The results before and after the atmospheric phase delay correction using GACOS were confirmed and analyzed in the main subsidence area. The findings of the experiments reveal that the atmospheric phase affects the mean LOS velocity results to some extent. The average difference between PS-InSAR before and after tropospheric correction is 1.734 mm/year with a standard deviation of 0.550 mm/year. The significance test of the two variables, 95%, showed that the tropospheric correction with GACOS data could affect the PS-InSAR results. Furthermore, GACOS correction may increase the error at some points, which could be due to its turbulence data’s low accuracy.

ESTIMATION OF GPS DISPERSIVE AND NON-DISPERSIVE NETWORK CORRECTION FOR ISKANDARnet NETWORK-BASED REAL-TIME KINEMATIC (N-RTK) POSITIONING SYSTEM

The concept of N-RTK positioning has been extensively developed in order to better model the distance-dependent errors of GPS carrier-phase measurements. These errors can be separated into a frequency-dependent or dispersive component (i.e., the ionospheric delay) and a non-dispersive component (i.e., the tropospheric delay and orbit biases) to express the network correction in order to attain better modelling of GPS distance dependent errors. However, the N-RTK performance may degrades due to severe atmospheric irregularities that would seriously affect the modelling of the GPS distance-dependent errors, thus affecting the quality of network correction generation. The development of integrity monitoring for network correction would be great idea to identify the quality and reliability of network correction data dissemination. Therefore, this paper aims to estimates the trend of GPS dispersive and non-dispersive network correction to supports future development of integrity monitoring for network correction of ISKANDARnet N-RTK positioning system. The first part of this paper is to extract the GPS dispersive and non-dispersive network residual components. This part includes the double-differencing technique, ambiguity resolution and carrier-phased linear combination in the process. The LIM then are applied for user network coefficient value computation purpose in the second part. Finally, the GPS dispersive and non-dispersive network correction can be generated with GF and IF network correction algorithm respectively. The trend of GPS dispersive and non-dispersive network correction is expected to aid the estimation and realization of threshold limit value for development of integrity monitoring for network correction of ISKANDARnet N-RTK positioning system.

Precipitation Water Vapour Variation in the East Java Region from Data CORS Using GIPSY

The Global Navigation Satellite System is being developed as an atmospheric remote sensing system through the calculation of Zenith Total Delay. The development of the Continous Operating Reference Station encourages research investigations into Zenith Tropospheric Delay with continuous data and good spatial resolution. This research studies the characteristics of spatial and temporal variations of the Zenith Wet Delay in East Jawa. The case study in East Jawa Province uses 16 Continous Operating Reference Stations. As a comparison, meteorological data from the Badan Meteorologi, Klimatologi, and Geofisika stations are used.The Zenith Total Delay and Zenith Wet Delay values from the Continous Operating Reference Station data are calculated using GIPSY 6.4 Software. The Zenith Wet Delay values are gridded using the kriging method with the size of the grids being 0,25 x 0,25. The ZWD value comparison from the Continous Operating Reference Station and meteorology data has a strong correlation with a coefficient value of 0,712. The mean of Zenith Wet Delay’s trend is increasing by about 0,712 mm/yr. The characteristics of the spatial and temporal variations of the ZWD value are influenced by the monsoon of Asia-Australian, which causes dry and rainy seasons, global phenomena such as El Nino and La Nina, rainfall, local meteorological conditions such as temperature and humidity, weather, and the topography of the stations.

Effect of broadcast and precise satellite orbits in the estimation of Zenith tropospheric delay and integrated precipitable water vapour from GPS

Global Positioning System (GPS) estimates the total delay in zenith direction by the propagation delay of the neutral atmosphere in presence of water vapour present in the troposphere. This total delay has been treated as a nuisance parameter for many years by the geodesists. The above delay have two parts dry delay and wet delay and known as Zenith Hydrostatic Delay (ZHD) and Zenith Wet Delay (ZWD) respectively. The Integrated Precipitable Water Vapour (IPWV) is estimated through ZWD overlying the receiver at ground-based station. The accuracy of the above said estimates depends on the quality of the predicted satellite orbits, which are not the same for each individual satellite. India Meteorological Department (IMD) is operationally estimating the IPWV on near real time basis at five places and matches fairly well (error ~6.7 mm) with Radisonde (RS) data. This paper examine the effect of International GPS Service (IGS) predicted precise orbits and near real time predicted rapid or broadcast orbits supplied by the Scripps Orbit and Permanent Array Center (SOPAC) on Zenith Total Delay (ZTD) and IPWV estimates by calculating the mean Bias and Root Mean Square Error (RMSE) for ZTD and IPWV in mm for all the five stations. The observed bias for ZTD is almost of the order of less than 1 mm in most cases and RMSE is less than 6 mm. Similarly the bias observed in the case of derived IPWV is almost negligible and RMSE is less than 1 mm.

Mitigating Atmospheric Effects in InSAR Stacking Based on Ensemble Forecasting with a Numerical Weather Prediction Model

The interferometric synthetic aperture radar (InSAR) technique is widely utilized to measure ground-surface displacement. One of the main limitations of the measurements is the atmospheric phase delay effects. For satellites with shorter wavelengths, the atmospheric delay mainly consists of the tropospheric delay influenced by temperature, pressure, and water vapor. Tropospheric delay can be calculated using numerical weather prediction (NWP) model at the same moment as synthetic aperture radar (SAR) acquisition. Scientific researchers mainly use ensemble forecasting to produce better forecasts and analyze the uncertainties caused by physic parameterizations. In this study, we simulated the relevant meteorological parameters using the ensemble scheme of the stochastic physic perturbation tendency (SPPT) based on the weather research forecasting (WRF) model, which is one of the most broadly used NWP models. We selected an area in Foshan, Guangdong Province, in the southeast of China, and calculated the corresponding atmospheric delay. InSAR images were computed through data from the Sentinel-1A satellite and mitigated by the ensemble mean of the WRF-SPPT results. The WRF-SPPT method improves the mitigating effect more than WRF simulation without ensemble forecasting. The atmospherically corrected InSAR phases were used in the stacking process to estimate the linear deformation rate in the experimental area. The root mean square errors (RMSE) of the deformation rate without correction, with WRF-only correction, and with WRF-SPPT correction were calculated, indicating that ensemble forecasting can significantly reduce the atmospheric delay in stacking. In addition, the ensemble forecasting based on a combination of initial uncertainties and stochastic physic perturbation tendencies showed better correction performance compared with the ensemble forecasting generated by a set of perturbed initial conditions without considering the model’s uncertainties.