G5.2

The capability of Deep Learning (DL) for operational wind speed retrieval from the measured Delay-Doppler Maps (DDMs) is recently characterized. It is shown that such techniques can lead to a significant improvement in the derived atmospheric data products. A global ocean dataset is developed processing the measurements of NASA Cyclone GNSS (CYGNSS). The model is based on convolutional layers for direct feature extraction from bistatic radar cross-section (BRCS) DDMs and fully connected layers for processing ancillary technical and higher-level input parameters. This model leads to an RMSE of 1.36 m/s and a significant improvement of 28% in comparison to the officially operational retrieval algorithm.

From the theoretical knowledge, several error sources are known, the modeling and correction of which is not easy due to their highly nonlinear interaction with other and the dependent parameters. DL is potentially able to learn such trends and correct the associated biases. For instance, rain splash on the ocean surface and swell waves alter the surface roughness, and consequently, the GNSS scattering patterns, which appear as a considerable bias in GNSS-R wind products. The magnitude of such biases is nonlinearly dependent on several technical and environmental parameters including the reflection geometry, and ocean surface state. After a brief introduction to the known physical mechanisms, we discuss how a DL-based fusion with data on bias-causing parameters, can improve the wind speed predictions.

How to cite: Xiao, T., Asgarimehr, M., Arnold, C., Zhao, D., Weigel, T., Mou, L., and Wickert, J.: Deep learning in spaceborne GNSS-R: Recent methodologies and atmospheric products, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-8226, https://doi.org/10.5194/egusphere-egu22-8226, 2022.

Sea level rise and sea state variability due to climate change and global warming are major research topics in the scientific community. Wind speed (WS) and significant wave height (SWH) are usable parameters to monitor the sea state threats and the impact of the ocean weather conditions in coastal areas. GNSS reflectometry (GNSS-R) has shown considerable promise as a remote sensing technique for ocean parameters estimation. Multiple studies have been successfully conducted in the recent two decades by using GNSS-R ground-based, airborne and spaceborne data to retrieve geophysical properties of the sea surface.

The focus of this study is to investigate the Doppler shift of the reflected signal as observable to estimate the Doppler spread (DS) and determine its correlation with sea state changes, making use of GNSS-R airborne data in coastal areas. An additional aim is to study the possibility of using the Doppler spread as a metric for coherent GNSS reflectometry for applications such as precise altimetry and precise total electron content (TEC) estimates. An experiment was conducted from the 12th to the 19th of July 2019 along Opal Coast, between the cities of Calais and Boulogne-sur-Mer, France. The experiment consisted of multiple flights at an altitude of ~780m (a.m.s.l). The direct and reflected signals were received by dual-polarized (Right-Handed and Left-Handed Circular Polarizations) antenna mounted on a gyrocopter.

A software receiver is used to process the direct and reflected signals from the right-hand channel. The resulting in-phase (I) and quadrature (Q) components (at 50 Hz rate) of the reflected signals are analyzed in the spectral domain every ten seconds to obtain the relative Doppler shift and power estimates. The coherence is established by analyzing the phase observations obtained from I and Q. The sensitivity of the reflected signal estimates and the sea state is determined by the correlation between the Doppler Spread with wind speed and significant wave height. The latter two were obtained from the atmospheric, land and oceanic climate model, ERA5, provided by the European Centre for Medium-Range Weather Forecasts (ECMWF).

Initial results have shown promising performance at a calm sea (WS: 2.9 m/s and SWH: 0.26 m) and grazing angles. Satellites with low elevations (E < 10°) present a Doppler Spread of 0.3 Hz and its Pearson correlations with respect to WS and SHW are 0.89 and 0.75, respectively. The performance is relatively poor for high elevation events (E > 30°). The DS increases up to 2.1 Hz and the correlation decrease to 0.55 and 0.42 respectively. Coherence conditions are still under study; however, preliminary phase analysis reveals coherent observations at events with elevations below 15° and sea state with a significant wave height of 0.26 m.

How to cite: Moreno, M., Semmling, M., Stienne, G., Dalil, W., Hoque, M., Wickert, J., and Reboul, S.: Sea state dependent Doppler spread as a limit of coherent GNSS reflectometry from an airborne platform, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10832, https://doi.org/10.5194/egusphere-egu22-10832, 2022.

Integrated Water Vapour (IWV) measurements from similar or different techniques are often inter-compared for calibration and validation purposes. Results are traditionally interpreted is terms of bias (difference of the means), standard deviation of the differences, and slope and offset parameters of a linear regression between the IWV measurements of the tested instrument with respect to the reference instrument. When the two instruments are located at different elevations, a correction must be applied to account for the contribution of atmosphere between the sites. Therefore, empirical formulations are often used. In this work it is shown that the widely-used model based on a standard, exponential, profile for water vapour density cannot properly correct the contribution of the atmospheric layer on the bias, slope, and offset parameters simultaneously. For example, correcting the bias degrades the slope and offset parameters, and vice-versa, with this model. An alternative method is proposed to derive an empirical model from real profiles observed by radiosondes. The method is developed for the special case of a tropical mountainous area with high IWV contents and strong diurnal and seasonal variations. Its application is illustrated with two examples, i) GPS to GPS comparisons and ii) GPS to satellite microwave radiometer comparisons.

How to cite: Bock, O.: An improved vertical correction method for the inter-comparison of Integrated Water Vapour measurements, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2772, https://doi.org/10.5194/egusphere-egu22-2772, 2022.

The reduction of geophysically induced high-frequency harmonic variations inherent in space geodetic measurements is carried out on the basis of physically-driven empirical models that may vary spatially, under the implicit assumption that they do not vary in time. Motivated by the fact that the parameters driving processes such as the atmospheric tides (largely induced by the absorbtion of ultraviolet and infrared radiation by Ozone and water vapor) that in turn partly excite the oceanic tides are affected by climate change, in this contribution we put the time-invariance modelling assumption to the test with a focus on harmonic deformation and atmospheric delay. To study temporal variations in atmospheric tides, we have analyzed hourly series from ECMWF’s latest reanalysis, the ERA5, including its back-extension. To validate the harmonic estimates and the temporal evolution thereof, we have resorted to two largely independent reanalyses: the MERRA2 and the JRA55. Recovering the evolution of harmonic coefficients has been carried out employing a square-root information filter (SRIF) and a Dyer-McReynolds smoother. Subsequently, the pressure harmonic fields are convolved with load Green’s kernels to simulate crustal displacements and harmonic pressure, temperature, and humidity fields are used to evaluate the refractivity integrals via ray-tracing. We found that the most important high-frequency waves, that is, the solar diurnal and semi-diurnal exhibit marked secular and seasonal variations. We find that harmonic S2 estimates for sites in the tropics from an hourly five-year long time series segment vary by up to 25% depending on the temporal boundaries thereof and the annual amplitudes of the S2 amplitude series from the SRIF can exceed 10 Pa.

How to cite: Balidakis, K., Zus, F., and Dobslaw, H.: On the Temporal Variability of Tidal Atmospheric Signals, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-1789, https://doi.org/10.5194/egusphere-egu22-1789, 2022.

In recent years, the significant growth of positioning applications has come with the development of low-cost dual frequency Global Navigation Satellite Systems (GNSS) receivers. The accuracy of these receivers in terms of positioning has been proved. Various studies have also highlighted the ability of these receivers to precisely monitor the atmospheric water vapour. The low cost of such receivers enables large deployment, thus presenting an advantage for many geoscience applications.

In this context, we have developed a hydrographic buoy prototype, equipped with low-cost GNSS receiver and antenna. This buoy was first aimed to be used for the monitoring in delayed time of offshore tides and currents, by a precise point positioning analysis of the GNSS raw data. In addition, the ability of this low-cost GNSS buoy for the water vapour monitoring was also investigated through the assessment of Zenith Tropospheric Delay (ZTD) estimates from the post-processing of the raw data. The comparisons with ZTD estimates from a nearby ground-based GNSS geodetic antenna and the ECMWF fifth ReAnalysis (ERA5), provide pretty good results with RMS differences lower than 10 mm. 

These conclusive results highlight the opportunities for the use of such low-cost systems for meteorology and climatology applications over the Oceans.

How to cite: Bosser, P., Bennini, V., Bouasria, M., Grit, Y., and Panetier, A.: A low-cost GNSS buoy for water vapour monitoring over the Oceans, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-1811, https://doi.org/10.5194/egusphere-egu22-1811, 2022.

How to cite: Panetier, A., Bosser, P., and Khenchaf, A.: Investigation of shipborne GNSS ZTD retrieval processing parameters by simulation, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5237, https://doi.org/10.5194/egusphere-egu22-5237, 2022.

Global Navigation Satellite System (GNSS) receivers are very versatile sensors, which have not only revolutionized positioning and navigation applications, but also provide numerous opportunities for environmental monitoring and remote sensing. Beside the monitoring of long-term ground movements and geodynamics, typical applications include the provision of water vapor estimates for numerical weather prediction (NWP) and climate studies as well as real-time applications such as seismic and geohazard monitoring. The rising number and quality of low-cost GNSS equipment, coupled with innovative telecommunication approaches (Internet of Things), allow for an increased and more cost-effective usage of such devices for those monitoring purposes, and thus foster a fast decision-making process.

An especially beneficial approach is the collocation of GNSS sensors at already existing meteorological or seismic stations. By using available infrastructure for power supply and communication, it can provide a sustainable and energy-effective extension of existing monitoring capabilities. The different parameters collected on-site can be used for cross-validation or provision of corrections for GNSS positioning. Furthermore, through (close to) real-time availability of observations, such collocated stations can aid early-warning systems for many different types of natural hazards (from extreme weather events to landslides and earthquakes). At the Institute of Geodesy and Photogrammetry at ETH Zürich we develop GNSS instrumentation to equip meteorological stations from the SwissMetNet (SMN). The work is carried out in the course of a pilot study in cooperation with MeteoSwiss.

This contribution introduces initial results of the SMN station Zürich-Affoltern, where the first prototype GNSS payload has been installed. We highlight key capabilities of the low-cost GNSS equipment used for these high-precision geomonitoring purposes. Moreover, we discuss concrete ideas for the build-up of a dedicated collocation network within the DACH (Germany-Austria-Switzerland) border area and the opportunities arising from it. These opportunities include the sustainable enhancement of infrastructure for climate change monitoring in the Alpine region as well as the build-up of early-warning systems for multiple types of geohazards. The latter might be achieved through the combination of parameters collected on-site with complementary data (e.g. satellite observations or NWP output) using innovative, data-driven approaches. Finally, we showcase examples and the potential of recent and ongoing works using these data-driven approaches.

How to cite: Aichinger-Rosenberger, M., Moeller, G., Hohensinn, R., and Rothacher, M.: MPG-S-NET: A multi-purpose low-cost GNSS collocation station network, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6873, https://doi.org/10.5194/egusphere-egu22-6873, 2022.

Water vapor is of great importance to the atmosphere and weather research. Airborne GNSS-based tropospheric delay estimation can reveal the atmosphere profile information, which is of more importance than site-based products and acts as an independent observation for meteorological application. On the other hand, progress in the meteorological community such as numerical weather models (NWMs) and forecast operations have the potential to augment GNSS. Many studies have investigated methods for applying NWMs in GNSS, mainly considering NWMs as a priori information. However, most methods implemented are merely suitable for static ground stations. They may not be optimal for dynamic platforms like unmanned aerial vehicles (UAVs) since the troposphere condition changes dramatically with vertical velocity and height. Under this background, we propose an NWM/GNSS tightly coupled method to take the best advantage of NWMs into GNSS data processing. The technique utilizes NWMs as a priori and considers the vertical distribution of the atmosphere to adaptively adjust the stochastic model for tropospheric delay estimation depending on the actual circumstance. The proposed method has been evaluated by an experiment using UAV and Global Forecast System (GFS) and found an improvement of precision and stability of tropospheric delay estimates.

How to cite: Zhang, Z., Zhang, W., Lou, Y., Zhou, Y., Bai, J., and Zhang, Z.: NWM/GNSS tightly coupled tropospheric delay estimation and application on an unmanned aerial vehicle (UAV) platform, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4444, https://doi.org/10.5194/egusphere-egu22-4444, 2022.

Atmospheric interaction with microwaves causes refraction as well as delays of electromagnetic signals. Although the effects are the same for all microwave signals including GNSS and radar, different spatio-temporal sampling of the tropospheric delays, different frequencies resulting in different phase sensitivity to path delays changes, as well as different processing strategies, assumptions and algorithms may lead to differences in the quantified tropospheric estimates. In case of GNSS, the most typical troposphere-related product is the zenith total delay - quantified after the mapping of all GNSS slant delays in the zenith direction. On the other side, double-difference tropospheric slant delays in persistent scatterer interferometry can be obtained in an iterative manner by isolating and subsequently adding the unwrapped low spatial frequency components of the phase residuals to update the estimate of the tropospheric phase. The updated tropospheric phase is then subtracted before the next iteration of point-wise regression-based estimations of topographic corrections and surface displacements. This iterative process is repeated for all scatterers with acceptable standard deviation of the phase residuals until convergence is reached. For comparison of the different tropospheric delays, the spatio-temporal characteristics of GNSS and InSAR observations must be considered. In this work, we use statistical interpolation methods to collocate the GNSS ZTDs with the InSAR measurements. Moreover, as another external (independent) observation we consider 3D fields of numerical weather models, which are integrated in the slant direction to produce (relative) tropospheric delay maps.

As a case study, an alpine region in the Valais area, Switzerland has been selected, which is an interesting scenario due to the high variability of the refractive index over complex terrain. A relatively dense GNSS network, as well as an interferometric time series of Synthetic Aperture Radar (SAR) images are available for the time span of 2008-2013. After introducing the available observations into the collocation approach, we perform the comparison and evaluation of the different tropospheric delays. In addition, we address the following two questions: How should the correct signal part be considered when modeling tropospheric delays using collocation? What is the effect of the GNSS network in terms of size and resolution? This work is an effort in understanding the different estimated/modeled delays, and it aims to set a baseline and a framework for the fusion of GNSS and InSAR tropospheric delays for the monitoring of the atmospheric state over complex terrain.

How to cite: Shehaj, E., Frey, O., Moeller, G., Aichinger-Rosenberger, M., Geiger, A., and Rothacher, M.: Relative tropospheric delay fields by GNSS, InSAR and NWP models in an Alpine Region, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4471, https://doi.org/10.5194/egusphere-egu22-4471, 2022.

Abstract: Usually, one can use the line-of-sight observations of GNSS (Global Navigation Satellite System) in not only positioning and navigation, but also in describing the medium that they pass through, such as ionosphere and troposphere. Particularly, for a specific station, one can further use obtained tropospheric wet delay in zenith direction to calculate integrated water vapor (IWV) or precipitable water vapor (PWV) with a transfer coefficient related to in-situ meteorological elements. Further, with IWVs or PWVs at a number of stations in a region, one can additionally discover the spatio-temporal distribution of water vapor by using the so-called tomography technology, noted as TWV hereafter. In this work, both retrieved PWVs and TWVs are analyzed and used in monitoring a rainfall process in Hong Kong ranging from 113.86°E to 114.36°E and 22.05°N to 22.40°N. A self-generated water vapor tomography package named GWATOS (GNSS Water vapor Tomography Software) is employed, and the Kalman filtering is used in the package to try to include more information that is valuable. The in-situ GNSS observations with interval of 5s and meteorological observations with interval of 60s at 18 stations in SatRef (Satellite positioning Reference station network) from 1^st to 31^st of May, 2016 are processed with BERNESE software by using the precise point positioning mode to retrieve the tropospheric delays. The temporal resolution of resulted PWVs is 30 minutes, and spatial resolution of TWVs is 0.05° in longitude, 0.07° in latitude and 15 unequal layers in height. In the experiment, the radiosonde profile data sets with temporal resolution of 12 hours at a station named Kings Park provided by University of Wyoming are used to externally assess GNSS retrieved water vapor. The retrieved PWVs show good consistency with results from radiosonde. The PWVs indicate obvious periods of high values, e.g., at 10^th and 21^st of May, and frequent variations, e.g., at 27^th and 28^th of May. As an example, the regional PWVs itself and its variation both in time and in space are analyzed to monitor the earlier-start, ongoing and end stage of rainfall process in 10^th of May, where both Red and Amber rainstorm warning are given. The results depict that abnormal period of PWVs are in good agreement with recorded rainfall period by Hong Kong Observatory. The results of retrieved TWVs show good internal and external agreements, the statistics are about 2.0mm and 1.7mm, respectively. The water vapor and its spatio-temporal variations in layers lower than 400 m and between 400m and 800m are investigated emphatically. The results show that both retrieved PWVs and TWVs with high spatio-temporal resolution could reflect rainfall process to some extent, especially the earlier-start and end stage of rainfall. That is to say, GNSS retrieved water vapor could be used to monitor the rainfall process in an auxiliary way. More importantly, no matter PWVs or TWVs could be used to trace the movement of water vapor or its structure both in time and in space, which can be further used in meteorology study for more detail information.

How to cite: Wang, M.: Spatio-temporal distributed water vapor retrieved from GNSS observations and its usage in monitoring a rainfall process in Hong Kong, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-1872, https://doi.org/10.5194/egusphere-egu22-1872, 2022.

Polar precipitable water vapor (PWV) is expected to increase under a warming climate. However, the conventional approach cannot provide sufficient long-term PWV records due to the high maintenance costs. Fortunately, the exponential explosion in the number of geodetic-quality Global Navigation Satellite System (GNSS) stations has broken this deadlock. Utilizing 20 radiosonde (RS) and 105 GNSS station data over two decades (1994-2020), we analyzed and evaluated the spatial and temporal variability characteristics of PWV in Antarctica and Greenland. The multi-year mean PWV values for Antarctica and Greenland were 5.63 ± 1.67 mms and 7.63 ± 1.35 mms, respectively, with annual standard deviations (STD) of PWV of 1.60 ± 0.77 mms and 3.44 ± 0.92 mms, respectively. In both Antarctica and Greenland, the PWV annual STD shows a gradual increase from the land center to the edge; while the PWV mean decreases with increasing latitude in Greenland, there is no significant latitudinal correlation in Antarctica. There is no significant regional difference in PWV trends, and from the statistical results, both Antarctica and Greenland show an increasing trend from year to year. The PWV trends in Antarctica and Greenland were 0.29 ± 0.77 mm/decade and 0.27 ± 0.64 mm/decade, respectively, with relative PWV trends of 5.98 ± 12.93%/decade and 3.87 ± 8.45%/decade, respectively.

How to cite: Ding, J., Chen, J., and Tang, W.: Increasing trend of Precipitable Water Vapor in Antarctica and Greenland, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6891, https://doi.org/10.5194/egusphere-egu22-6891, 2022.

The atmospheric humidity in the Polar Regions is an important factor for the global budget of water vapour, which is a significant indicator of Earth’s climate state and evolution. The Global Navigation Satellite System (GNSS) can make a valuable contribution in the calculation of the amount of Precipitable Water Vapour (PW). We focus on Polar Regions, especially Antarctica. 20-year GPS observations, acquired by more than 40 GNSS geodetic stations, were processed with the purpose of ensuring the utmost accuracy of the PW retrieval, adopting homogeneous, consistent, and up-to-date processing strategies. We also estimated PW from radio-sounding stations (RS), which operate Vaisala radiosondes, co-located with GNSS stations. The PW values from global atmospheric reanalysis model were used for validation and comparison, very high correlation coefficients between times series, have been highlighted both in the Arctic and Antarctica. A small dry bias of RS vs. GPS values was found in the Arctic, while no clear behaviour is present in Antarctica. The PW_GPS and PW_RS seasonal variations are consistent, as also confirmed by scatter plots.

After validation, long-term trends, both for Arctic and Antarctic regions, were estimated using Hector scientific software, which allows the estimation of trends from time series with temporal correlated noise. We applied a function to estimate the linear trend plus the annual/semiannual signals, and autoregressive noise model AR(1) which best fits the residuals of all investigated PW time series. We investigated also on the choice of the most suitable noise model, this study was useful in determining the residuals of the time series, once the trend and seasonal signals were subtracted. Positive PW_GPS trends dominate at Arctic sites near the borders of the Atlantic Ocean. Sites located at higher latitudes show no significant values. Negative PW_GPS trends were observed in the Arctic region of Greenland and North America. A similar behaviour was found in the Arctic for PW_RS trends. The stations in the West Antarctic sector show a general positive PW_GPS trend, while the sites on the coastal area of East Antarctica exhibit some significant negative PW_GPS trends, while in most cases, no significant PW_RS trends were found. The present work confirms also that GPS is also able to provide reliable estimates of water vapour content in regions where data are sparse and not easy to collect as the Arctic and Antarctic regions are.

How to cite: Negusini, M., Petkov, B., Tornatore, V., Barindelli, S., Martelli, L., Sarti, P., and Tomasi, C.: Water Vapour assessment using GNSS and Radiosondes and long-term trends estimation over Polar Regions, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13470, https://doi.org/10.5194/egusphere-egu22-13470, 2022.