This study demonstrates the utility of the Global Navigation Satellite System Reflectometry (GNSS-R) method for remote sensing of ice shelf surface deformation and roughness. The phase coherence of the reflected GNSS carrier signal is related to surface deformation and roughness: smoother, less deformed areas scatter the signal more coherently. More than 900000 tracks of Spire Global Inc. grazing angle GNSS-R data collected over Antarctica between 2021 and 2023 were processed to determine signal carrier phase coherence. Results are presented in the form of coherence maps, where each measurement has a spatial resolution of approximately five kilometers, and the Ross Ice Shelf (RIS) is used as a case study for more detailed examination. Clear visual patterns in signal coherence are shown to correspond with the locations of streaklines and deformation features on the surface. A GNSS-R coherence-based roughness proxy is introduced and shown to correlate with surface roughness derived from the reference elevation model of Antarctica digital elevation model (REMA DEM). This novel application of the GNSS-R method demonstrates the exciting potential to complement existing data sources and improve understanding of ice shelf structure and evolution.

This paper presents a methodology to estimate Global Navigation Satellite System (GNSS) code delay biases between zenith-facing antennas used for precise orbit determination (POD) and side-facing antennas primarily used for radio occultation (RO) and reflectometry on Spire’s low-Earth orbiting (LEO) cubesats. The approach involves connecting POD and RO carrier phase arcs, subtracting them from corresponding code arcs, and computing biases from the resulting residuals. We also derive multipath maps for these spacecraft, which is an important step in our bias estimation approach due to the need to connect long arcs of data. Analysis of 2023 Spire data reveals bias variations across spacecraft, with newer bus versions in sun-synchronous orbits (SSO) showing more consistent results. Accurate bias estimates are crucial for enhancing the use of RO code measurements for absolute total electron content (TEC) estimation, as well as for improving LEO orbit determination.

This paper presents Integrated Water Vapor (IWV) retrieval using spaceborne grazing-angle GNSS reflectometry (GNSS-R) data collected over the equatorial ocean. IWV is an important parameter for weather and climate modeling. There is a lack of observations over the oceans with short revisit times and high spatial resolution to capture the dynamic nature of IWV. This study demonstrates that GNSS signals, coherently reflected over calm ocean surfaces and received by a side-looking antenna at low elevation angles, can produce high precision range measurements from which IWV can be derived. This is achieved by using Spire Global’s grazing-angle GNSS-R signal carrier phase data collected over the Southeast Asia ocean region in 2022, at elevations between 5° and 8°. Detailed methodology of the retrieval process is presented, using coherent reflection tracks lasting more than 30 seconds corresponding to about 140 km along the ground track. The retrieval incorporates the VMF3 mapping functions to provide relative along-track IWV profiles which are compared with two IWV models, VMF3 and ERA5. To preserve the high spatial-temporal resolution of the GNSS-R measurements and ensure a fair comparison with models, we present the GNSS-R based IWV retrievals at the original 50-Hz (~100m spatial sampling) rate and a smoothed version at 5 seconds interval. The results highlight the dynamic nature of GNSS-R derived IWV measurements. Their deviations from the ERA5 model are statistically consistent over time, independent of the model’s fluctuations. Classification analysis shows 15% of GNSS-R derived IWV closely match either VMF3 and/or ERA5, suggesting that they provide an alternative measure of IWV variations. Meanwhile, 50% show no agreement, suggesting that GNSS-R captures unique IWV dynamics not fully represented by the models. This is supported by a comparison with IWV statistics from Sentinel-6 which reveals that GNSS-R retrievals follow the same patterns with slightly more dynamic variability than Sentinel-6. On-going studies are investigating additional direct validation between GNSS-R and other observations as well as application of this method in polar regions using signals reflected over sea ice.

This study utilizes Global Navigation Satellite System Reflectometry (GNSS-R) to map the surface of the Ross Ice Shelf in Antarctica. Spaceborne GNSS-R is a relatively new method for cryosphere applications and utilizes reflected GNSS signals of opportunity collected on a platform in low-Earth orbit, allowing for widespread data coverage with better temporal resolution than conventional remote sensing technologies. Coherency of the reflected signal, computed using phase-based circular statistics, can be used as a proxy for surface roughness. Using 25819 Spire Global grazing angle reflectometry tracks, we generate a high-resolution map of coherency over the Ross Ice Shelf. This technique shows promise for identifying regions with similar surface characteristics and deformation history, as well as for detecting the ice shelf/sea ice border, large rifts, and flow lines.

In this paper, we compare water surface height profiles measured by the Ice, Cloud, and Elevation Satellite-2 (ICESat-2) ATL13 data product on April 25, 2023 that intersect with a GNSS-Reflectomety (GNSS-R) altimetry track collected on April 17, 2023 over the Tonlé Sap Lake in Cambodia. Overall, we find reasonable agreement between the two datasets, with the nearest segment of ICESat-2 and GNSS-R tracks having a correlation of 0.79. However, some deviations exist, with the GNSS-R track measuring 7 cm lower than the ICESat2 track at an intersection point, potentially due to different lake surface conditions at the time of each measurement. This comparison is motivated by the interest in applying GNSS-R altimetry for a broader study of large inland lakes. Our focus on the Tonlé Sap Lake is inspired by the lake’s key role in the region combined with the difficulty in monitoring it by other techniques due to its large size, frequent cloud cover, and surrounding dense mangrove forests. GNSS-R (operating in L-band) can penetrate these environmental factors while accumulating frequent and spatially diverse water surface measurements. In total, we find 53 GNSS-R tracks collected by Spire Global between 2021 and 2023 that pass our altimetry criteria. In contrast, we find only 27 ICESat-2 tracks (each laser counted individually) through 12 total passes, likely limited by cloud cover.

The objective of this study is to investigate the utility of the GNSS-R based stare processing technique in combination with machine learning (ML) for ocean surface wind speed retrieval from CYGNSS delay-Doppler maps (DDM). Building on our previous work using a simulated data set [1], this study presents an analysis of stare processing and machine learning methodologies applied to real CYGNSS data. We train three ML algorithms, neural network, bagging regressor, and gradient boosting regressor, with six feature sets derived from the DDM observable and satellite geometry. We compare stare processing performance to the baseline best-performance result from our previous work in [2], measuring performance with root mean squared error (RMSE) and mean absolute percentage error (MAPE). Stare processing in combination with the bagging regressor algorithm achieves an 8.3% RMSE accuracy improvement over the baseline, with an RMSE of 2.5 $\frac{m}{s}$. By replicating and comparing to the process used to produce the best performance result in [2], we can directly see the effect of the addition of stare processing features. We show that stare processing improves ocean wind speed prediction accuracy, and expect that it could be an advantageous addition to other ML-based wind speed prediction methods in the future.

This paper investigates using Spire’s grazing-angle GNSS-R data to detect the existence of surface water in the lower Mississippi region. The Signal-to-Noise Ratio (SNR) of the reflected L2 GPS signals acquired using the Spire Global constellation are analyzed over several tracks covering the Mississippi river, along with Normalized Difference Vegetation Index (NDVI) from Sentinel-2 Multispectral Imager (MSI) and backscattering coefficients from Sentinel-1 Synthetic Aperture Radar (SAR) C-band acquisitions. The specular points are segregated on the basis of thresholds based on the GNSS data and associated remote sensing data with correctly classifying the points over water with a true positive rate between 55-70% for L2 SNR thresholds above noise floor.

This paper describes a preliminary investigation into the relationship of Global Navigation Satellite System-Reflectometry (GNSS-R) phase-based coherency and signal-to-noise ratio (SNR) with the extent of surface water contained in the reflected the signal's spatial footprint. Our study focuses on 82 phase-based GNSS-R tracks collected by Spire Global grazing angle satellites in 2021 over Lake Okeechobee, FL. For each track, we calculate its spatial footprint (assumed equal to the First Fresnel Zone (FFZ)) and determine the amount of surface water within the FFZ using a lake mask derived from the 2021 Global Surface Water Explorer (GWSE) Seasonality product. The percentage of surface water in the signal footprint (denoted %F) is compared with the reflected signal's SNR and level of coherency, with coherency measured by the phase-rate circular length. Overall, we see a positive relationship between SNR, circular length, and %F. However, circular length appears to be a more robust indicator of water, with more consistent, high magnitudes once the scattering surface is dominantly water (>70%F). Further, circular length appears to maintain a high magnitude at different distances into Lake Okeechobee when the footprint is full of water, whereas SNR gradually decreases as the signal track approaches the center of the lake. This is most likely a result of wind roughening the lake surface farther away from the shore.

This article presents two empirical models, the low wind bagged trees (LWBT) and high wind bagged trees (HWBT) ensemble models to estimate ocean surface wind speed using spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) data. The models are empirically trained using NASA’s Cyclone GNSS (CYGNSS) mission level 1 data (version 2.1). The truth label for the LWBT model is the wind speed product derived from European Centre for Medium-Range Weather Forecasts (ECMWF) ERA-5 and Global Data Assimilation System (GDAS), while the label for the HWBT model is wind speed measurements from stepped frequency microwave radiometer (SFMR). Testing results show that the LWBT and HWBT models achieved global wind speed retrieval root-mean-square-error (RMSE) of$\sim $1.5 and$\sim $1.4 m/s, respectively, corresponding to an improvement of 29% and 65% with respect to the CYGNSS Level 2 standard wind speed product. The maximum bias is reduced by 65% and 60% for LWBT and HWBT over the Level 2 wind speeds, respectively. Two typhoon case studies are presented to corroborate the model performances and their retrieved wind speeds are consistent with reports from World Meteorological Organization (WMO) and with the measurement provided by the Huangmao Zhou (HMZ) weather station.

When GNSS signals reflect off of the surfaces of lakes, rivers, wetlands, and other inland water bodies, the surfaces are often sufficiently smooth to produce coherent reflections. The observable produced from coherent reflections made by GNSS Reflectometry (GNSS-R) instruments exhibits particular features with respect to diffusely scattered signals by rough land and wind-driven oceans allowing detection of such smooth bodies. Several different GNSS-R coherence detection approaches have been reported in the literature and developed among the GNSS-R community over the last several years; however, the merits of each approach are difficult to compare because they are often applied to different scenarios and quantified in different ways, independently of each other. This paper provides a unified comparison of a wide variety of different GNSS-R coherence detection approaches, which is the most extensive published to date. The approaches are applied to a common data set from the NASA CYGNSS satellites that includes both the standard Level-1 DDM science product as well as raw baseband signal recordings. Additionally, simulated observables are generated with varying coherent and non-coherent reflection components to exercise algorithms over a wide range of SNRs and relative powers. Objective measures of accuracy are used to quantify the performance of each approach in the context of relative implementation complexity. Conclusions are presented on the pros/cons of the various methods as they relate to various applications such as real-time in-orbit coherence detection or post-processing on the ground.

Global Navigation Satellite System (GNSS) coherent ocean reflections collected at grazing-angles by Spire Global’s CubeSats are explored to retrieve Sea Level Anomalies (SLA) over the Gulf of Mexico. Dual-frequency GPS carrier phase estimations are derived from the 50-Hz I and Q samples generated onboard the low Earth orbit CubeSats. First order ionospheric phase advancement, troposphere delay, and precise orbit solutions for GPS satellites and CubeSats mitigate the range measurement errors. Comparisons between collocated specular reflection tracks from various Spire CubeSats and GPS satellites configurations demonstrate the self-consistency of the resulting SLA solutions. The Spire retrievals are then validated against conventional altimetry SLA products. There is a Root Mean Square (RMS) difference of ~ 20 cm between the Spire and the gridded model mesoscale SLAs, unless the reflected elevations are very low. To better evaluate the performance of the Spire retrievals they are examined over sharp ocean topography. About 40% of coherent reflections with elevations above 12° detected the signature of the energetic Loop-Current during the 2020-2021 period. The Spire retrievals were also compared with collocated Sentinel-3 tracks. One track with elevations above 12° demonstrated high correlation and agreement with the Sentinel-3 retrievals. The other two low-elevation tracks showed large magnitude discrepancies, but similar trends once appropriate scaling factors are applied. The study demonstrates the potential for GNSS reflectometry’s application to SLA mapping while also highlighting where significant improvements in error mitigations are needed. A particularly large error source is the troposphere error for signals with lower elevations.

CYGNSS satellites record the power of GNSS reflection signals in delay and Doppler shift bins. Myriad delay Doppler maps (DDM) have been collected since 2016. Research has shown that reflection range delays can be derived from CYGNSS DDMs and their metadata for satellite altimetry. However, the performance of these DDMs in estimating inland water levels has not been thoroughly evaluated. This study leverages the coherent reflection-dominated DDMs from 2020 to 2022 to estimate water levels at five lakes. Compared to radar altimetric (RA) observations, the CYGNSS results have an overall bias of ~2.0 m, a root mean square difference (RMSD) of ~3.1 m, and an unbiased RMSD (ubRMSD) of ~2.4 m, when combining measurements from all study sites over the three-year period. During 2020–2021, the bias is lower at 1.1 m, the RMSD is ~1.9 m, and the ubRMSD is reduced to ~1.6 m. Smaller bias, RMSD, and ubRMSD values are also observed in 2022 nighttime data. However, the daytime data from 2022 show higher water level estimations compared to RA observations with a bias of 4.7 m, a RMSD of 6.1 m, and an ubRMSD of ~3.9 m. The CYGNSS overestimations are linked to an increase in ionospheric total electron content. This study demonstrates the feasibility of spaceborne GNSS-R for measuring inland water levels and highlights the usefulness of the current version of CYGNSS DDMs in monitoring reservoirs and inundation events characterized by substantial water level changes.

Spaceborne global navigation satellite system-reflectometry (GNSS-R) has the potential to provide a new data source for inland water surface observations. In this paper, we present several case studies for the Mississippi River surface gradient observations using the GNSS-R data from Spire's CubeSats. The retrieved river surface slopes are in a range from 2.2 to 16.6 cm/km, and the lower river sections tend to be more flat. Higher phase noise is observed from Spire's GNSS-R data compared previous studies using CY GNSS data. It is likely due to the Spire CubeSat antenna design and the low elevation angles of the reflection signals. A lower elevation angle corresponds to a larger footprint of up to several kilometers and thus may introduce strong multipath interference in the measurements. The presented results also confirmed the GNSS-R observation of superelevation phenomenon that was first discussed in [1]. The ionosphere effects and L2 signal's competitive contributions to the retrieval noise and ionospheric correction are also discussed.

The objective of this study is to investigate the utility of the GNSS-R based stare processing technique in combination with machine learning for ocean surface wind speed retrieval. Using a simulated data set produced with real CYGNSS satellite geometries, we train a multi-hidden layer neural network with six feature sets derived from the delay-Doppler map (DDM) observable. We compare stare processing performance to the baseline best-performance result from our previous work in [1], measuring performance with root mean squared error (RMSE). We find that in the noise-free case, stare processing performs similarly to the baseline. When thermal and speckle noise are added to the simulated DDMs, stare processing significantly outperforms the baseline result. Stare processing improves our ability to accurately predict ocean wind speed with machine learning.

S IGNALS from Global Navigation Satellite Systems (GNSS) that were initially developed for positioning, navigation, and timing applications are now operationally being used to sense the Earth surface using GNSS reflectometry (GNSS-R) techniques. With over 140 current GNSS satellites transmitting open, high-quality signals, LEO-based GNSSR can achieve global measurement coverage using passive, low-power receivers. For example, the delay-Doppler maps produced by the Cyclone Global Navigation Satellite System (CYGNSS) constellation is operationally used to retrieve ocean surface wind speeds associated with hurricanes [1]. Additionally, coherent signal reflections from CYGNSS and other low Earth-orbiting (LEO) satellites have also been used to achieve centimeter-level surface altimetry over calm ocean, inland water bodies, wetlands, or sea ice [2], [3], [4]. Compared to terrestrial-based GNSS-R, reflectometry using spaceborne platforms benefits from nearly global measurement ground-track coverage.

This paper presents an updated evaluation of the coherency of grazing-angle ocean reflected GPS signals received on low-cost Spire CubeSat receivers. It is based on almost a year of dual-frequency carrier-phase observations sampled at 50 Hz from 2020. This dataset contains more than 2000 ocean reflections every week in 2020. As expected, coherency increases for low grazing angles: by a factor 20 from grazing angles above 20° to below 10° in the Indonesian Seas. Coherency also increases for smooth ocean surface conditions: by a factor 80 for Significant Wave Height (SWH) decreasing from 1.5 m to 0.5 m using grazing angles below 10°. The results provide clear relationships between coherency and various ocean surface parameters. Overall coherency is more sensitive to SWH with an almost exponential trend, followed by mean wave period, and less by wind speed which shows a relatively constant behavior for wind speeds < 4 m/s. The mean wave direction relative to the specular point along-track direction has no effect on coherency for wave periods less than 5s.

This paper presents a quantitative analysis of spaceborne GNSS-R signal coherence dependence on satellite elevation angles at specular reflection points over sea ice and ice sheets in Greenland and Antarctica using TechDemoSat-1 delay Doppler maps. Over sea ice and ice sheets, the probability of coherent reflections is high at low elevation angles and decreases with increasing elevation angles, as expected. For sea ice, the coherence rate is ~96% at elevation angles of$20-43^{\circ}$, then drops to a minimum of ~25% near nadir. Over the Greenland ice sheet, the maximum coherence rate is 86% at the elevation angle of$47^{\circ}$, and the minimum is 21% at$73^{\circ}$. In Antarctica, the coherence rate reaches a maximum of 90% at$48^{\mathrm{o}}$and a minimum of 36% at$72^{\circ}$. The findings provide a quantitative characterization of the GNSS-R coherency over sea ice and ice sheets and are useful for future GNSS-R mission designs.

Coherent reflections of global navigation satellite system (GNSS) signals have a measurable carrier phase, enabling higher precision for certain GNSS-based Earth remote sensing applications. In this letter, we explore the dependence of coherence on three land surface characteristics: surface water, topography, and soil moisture (SM). Carrier phase measurements are obtained by tracking raw intermediate frequency data collected by the cyclone GNSS (CYGNSS) mission. In total, several hundred data collections between 2017 and 2019 are analyzed. The phase coherence, quantified using statistics of the tracked carrier phase, is compared to the corresponding land characteristics on a per-track basis and across the entire dataset. On a per-track basis, we find that the level of coherence can often be explained by the presence of surface water, with no obvious dependence on topography or SM. However, by analyzing the entire dataset, we show that topography and SM have a weak but noticeable impact on the coherence.

This letter demonstrates the potential of estimating river surface slope using space-borne global navigation satellite system-reflectometry (GNSS-R) carrier phase measurements. A case study is presented where the slopes of several segments of the Orinoco river are retrieved based on the processing of raw intermediate-frequency (IF) data recorded by the cyclone global navigation satellite system (CYGNSS) mission. The retrieved river slopes mostly vary between 3.9 and 5.1 cm/km and are in agreement with the surveyed mean slope of 4.5 cm/km for the lower Orinoco River. The correction and calibration of several systematic errors, i.e., CYGNSS satellite orbit errors, tropospheric and ionospheric effects, and river surface reference height error, are discussed in this letter. For the Orinoco River case study, the river slope retrieval is calibrated using the reflection signal over a nearby lake to mitigate the mis-modeled and unmodeled errors. Analysis of the specular point (SP) tracks of GPS reflection signal indicates that CYGNSS has the potential to provide river slope observations for rivers that have sufficient width with high temporal and spatial resolutions.

Ionospheric total electron content (TEC) maps with high spatial and temporal resolutions are essential for depicting the state of the ionosphere and for performing ionospheric delay corrections associated with satellite navigation applications. Low Earth orbit (LEO) CubeSat-based global navigation satellite system (GNSS) reflectometry (GNSS-R) measurements provide a promising opportunity for retrieval of ionospheric TEC over sea ice and calm waters, which offers a potential new data source to fill the gaps of ground-based GNSS networks. However, the GNSS-R slant TEC (sTEC) estimations include contributions from the incident and reflection ray paths, whose ionospheric piercing points (IPPs) can be separated by hundreds of kilometers. This article presents an algorithm that integrates sTEC measurements from the GNSS-R CubeSats and available ground-based GNSS receivers to derive Arctic vertical TEC (vTEC) maps. A simulation study using the model ionosphere constructed from the NeQuick-2 is conducted to assess the performance of the algorithm. Varying levels of temporal resolutions and solar activities, and the number of CubeSats and the number of maximum simultaneously tracked reflection signals by a CubeSat are implemented in the simulation. The results show that the inclusion of coherent GNSS-R measurements improves the accuracy of the vTEC maps under all levels of solar activities. The RMSE improvement percentage is most obvious when the update interval is the shortest. Increasing the number of CubeSats further improves the accuracy. However, no significant improvement in vTEC map accuracy is observed when the number of maximum simultaneously tracked GNSS-R satellites is higher than 4. Quantitative measures and analyses of the algorithm performances are presented in this article.

Signal-phase measurements from global navigation satellite systems (GNSSs) have become an important tool for various remote sensing applications, including measuring ionosphere plasma content, atmospheric radio occultation, and water and ice reflectometry. In these types of scenarios, GNSS signals often experience harsh propagation conditions, such as low signal-to-noise ratios, multipath, and semicoherent scattering. These conditions, in turn, lead to the frequent occurrence of cycle slips, which manifests as persistent discrete changes in the bias of the carrier phase measurement. In order to effectively use the precise GNSS phase measurements under such conditions, we argue that a window of high-rate measurements must be used. In addition, we suggest that enforcing sparsity in the occurrence of detected cycle slips can aid in detection. We, therefore, develop a batch cycle-slip detection and estimation method that is effective and computationally tractable under harsh signal conditions. This work focuses in particular on strong ionosphere scintillation, which is among the most difficult scenarios for estimating cycle slips. We demonstrate the effectiveness of our method on both simulated and real GNSS scintillation datasets, showing around a 90% reduction of slips.

The rate of total electron contents (ROT) calculated from global navigation satellite system (GNSS) observables has mainly been used as a quantity that can be averaged into an ROT index (ROTI), which is unable to detect small-scale, fast-moving disturbances. In this study, we demonstrate that a unique signature on ROT time series can be used to estimate the horizontal velocity and dimension of the irregularities when GNSS signals propagate through ionospheric irregularities (IIs). This was achieved by estimating the time lags and durations of the signatures observed by an array of closely spaced GNSS receivers. We applied this ROT-based estimation method to a three-element GNSS receiver array yielding high-rate carrier phase measurements in Poker Flat, Alaska. The velocities were estimated with a precision of 10 m/s and 2° in speed and direction, respectively. In summary, this method can be used to better describe spatial and temporal distributions of IIs. Furthermore, such distributions can be used to project or predict ionospheric fluctuations for receivers located in the directions of the ionospheric disturbance drift.

This article assesses the coherency of Global Navigation Satellite System (GNSS) signals reflected off the oceans and sea ice under grazing angle geometries and received aboard low Earth orbit (LEO) CubeSats for precision altimetry applications. The coherency is characterized as a function of ocean surface conditions and reflected signal parameters based on Spire Global CubeSat data collected from January to April 2019. The data contain 50-Hz GPS L1 and L2 carrier phase estimations obtained by open-loop tracking. Indicators based on the circular statistics of the excess-phase noise are developed to identify coherent and semicoherent reflections. Based on these indicators, we found that ~1% and 44% of GPS reflections over the ocean and sea ice, respectively, have potential for precision altimetry. The coherent and semicoherent reflection rates reach 23% in areas less than 200 km from the coastline and under calm sea conditions. Over young sea ice over the Arctic, this rate can be as high as 70%. There is a strong relationship between coherency and signal strength, and the coherency occurrence rate improves as the grazing angle decreases. The quality of the L1 and L2 coherent reflections is similar over sea ice, while, for reflections over the ocean, L1 signals are predominantly noisier and less coherent than the L2 signals. Using a postprocessing filtering method, the semicoherent reflections can achieve a similar level of altimetry precision as that of the coherent ones, thereby increasing the along-track length of the retrieved altimetry profile.

In this article, we investigate coherent global navigation satellite system reflectometry (GNSS-R) measurements obtained at the low earth orbits (LEOs) as a potential new data source for ionospheric total electron content (TEC) and ionospheric disturbance observations. Current global ionospheric TEC maps (GIMs) have limited spatial and temporal resolutions and accuracy due to lack of GNSS observations over oceans, polar caps, and inaccessible terrains. Our analysis of Spire Global’s CubeSat data indicates that coherent GNSS signals reflected off sea ice, inland water bodies, and calm ocean surface can be processed to achieve cm-level precision carrier phase estimations. Signal coherency is especially prevalent over sea ice where 41.7% reflections are coherent, compared to 4.3% in the overall dataset. This article presents the methodology to estimate slant TEC along the reflection signal ray path using coherent dual-frequency GNSS-R pseudoranage and carrier phase estimations obtained from low-cost CubeSats. The methodology is applied to Spire Global’s CubeSat data. The results show that the slant TEC retrieved from GNSS-R measurements and from GIM follow a similar trend. Moreover, the GNSS-R-based TEC time series offer a nearly “frozen in time” observation of the ionospheric structures due to the rapid scan velocity of GNSS-R rays across the ionosphere. The study demonstrates the potential of GNSS-R observations to fill the data gaps over the polar regions where space weather activities and TEC disturbances are most frequent and intense. Potential error sources and mitigation techniques are also discussed.

Coherent GNSS reflections over land predominantly occur over surface water bodies. This study presents a method to jointly use carrier phases and signal strengths of reflected signals to identify coherent reflections and applies it to the 50-Hz GNSS-R measurements from Spire Global Cubesats and CYGNSS microsatellites to map inland water bodies. A coherence detector was first developed using the circular statistics of carrier phase noises, identifying the input samples as coherent, semi-coherent, or incoherent. For any given track of data, we used this coherence detector to iteratively assess the coherency levels of the samples by a moving time window, then derived the coherency levels with the highest confidence. The circular statistics-based semi-coherent reflections with signal strengths above the prescribed threshold were regarded as coherent. The specular reflection points of the coherent reflections represent the locations of surface water. This method was applied to the Spire data to obtain the surface water extents for 1951 lakes and the CYGNSS data for 113 lakes. Compared to Global Surface Water Explorer observations, around 90% of the disagreements of the Spire data-based surface water boundaries are less than 0.73 km with a mean of 0.28 km and a standard deviation of 0.24 km. As for CYGNSS, ~90% of the disagreements are less than 0.43 km with a mean value of 0.18 km and a standard deviation of 0.16 km. The possible error sources are mainly fractional surface water, nearly flat and saturated ground surface, background land cover, and GNSS-R geometry.

We introduce a new method for retrieving ocean wind direction from GNSS-reflectometry measurements. Using a technique called stare processing, where information from a sequence of consecutive delay-Doppler maps is combined, values related to the wind-dependent surface slope statistics are extracted. These values are used to build a feature vector (the input to the machine learning model) that isolates dependence on wind speed and direction. Simulated sequences of 15 noise-free delay-Doppler maps are used to explore the feature engineering approach and to train and test the machine learning retrieval. The retrieval RMS error is 9.0 degrees. The algorithm is configured so that it can be applied, at least in principle, to standard CYGNSS level 1b data.

This paper presents an evaluation of the relative sea ice surface heights retrieved using carrier phase measurements of coherent GNSS reflection signals recorded by Spire Global Inc.'s LEMUR-2 low-Earth orbit (LEO) cubeSats between January to April 2019. Coherent GNSS signal reflection usually occurs at relatively low elevation angles, and sea ice is a good reflector for GNSS signals. As analyzed in [1], most of the uncertainty for GNSS-R relative sea surface height (SSH) retrievals is from the tropospheric delay models, which usually have several centimeters error in the zenith total error (ZTD), and the effects will be significantly amplified at low elevation angles. In this paper, the retrieved sea surface height anomalies (SSHA) using LEMUR-2 GNSS-R data are compared with the monthly averaged ice freeboard measurements from ICESat-2. Overall, the comparison shows good consistency. However, for some reflection events/segments, adjustments of the zenith troposphere delay model of up to 11 cm is required in order to fit the ICESat-2 ice freeboard measurements. Therefore, the tropospehere delay correction may be a challenge for high-precision GNSS-R sea ice altimetry.

The carrier phase measurements of coherent Global Navigation Satellite Systems Reflectometry (GNSS-R)signals have demonstrated the potential for high-accuracy altimetry applications. However, the coherent components in the reflected signal are often accompanied by noncoherent scattered signals, which may cause numerous cycle slips in the conventional open-loop (OL) tracking and postprocessing approach. These cycle slips can lead to meter-level errors in the altimetry retrieval. This article presents a simultaneous cycle slip and noise filtering (SCANF) method for dual-frequency GNSS-R phase measurements generated by OL tracking. The performance of this method is demonstrated using real GNSS-R phase measurements recorded by receivers on board Spire’s low-Earth orbit (LEO) satellites. The retrieved sea surface height anomaly (SSHA) has an average root mean square (rms) of 7.3 cm for sea and sea ice surface reflections relative to the mean sea surface (MSS). The technique presented in this article is also applicable to GNSS carrier phase measurements obtained under other challenging conditions, such as signals propagating through ionosphere plasma structures, radio occultation (RO) signals traversing lower troposphere, and multipath-rich environments.

This article presents an adaptive hybrid-tracking (AHT) algorithm designed to process GNSS-R signals with a sufficient coherent component. Coherent GNSS-R signals have the potential to enable high-precision and high-resolution carrier-phase measurements for altimetry, sea-level monitoring, soil-moisture monitoring, flood mapping, snow-water equivalent measurements, and so on. The AHT algorithm incorporates the model inputs typically used in the master-slave open-loop (MS-OL) architecture into a closed-phase lock loop. Raw IF data recorded by the CYGNSS satellites over in-land water, land, and open-ocean surface are used to demonstrate the performance of the AHT. The results show that the AHT algorithm achieves comparable robustness with the MS-OL implementation while maintaining centimeter-level accuracy and excellent carrier-phase continuity that can be achieved with a fine-tuned Kalman filter (KF)-based adaptive closed-loop (ACL) system. Moreover, the AHT is suitable for real-time implementation and is applicable to other radio signals-of-opportunity.

Precise carrier phase measurements from coherent reflections of Global Navigation Satellite Systems (GNSS) signals are poised to enable new capabilities in Earth remote sensing. However, there is still much to learn about the reflection conditions that produce coherence. For land reflections especially, the degree to which various surface characteristics impact coherence is unclear. In this paper, we explore the dependence of coherence on four surface characteristics: maximum water extent, surface topography, soil moisture, and vegetation water content. To quantify the level of coherence, the phase is tracked from two sets of CYGNSS raw IF data and the circular length of the changes in phase is calculated. A strong link is found between coherence and the presence of surface water; the impact of the other surface characteristics on coherence is still under examination.

This paper presents two approaches to achieve high quality GNSS carrier phase estimations that can be implemented in bi-static receivers onboard low Earth orbiting satellites for precision altimetry applications. These two approaches are an adaptive hybrid open- and closed-loop tracking for single frequency GNSS receivers and an adaptive multicarrier tracking algorithm that optimally combines the outputs generated by multiple carrier tracking loop outputs. As a case study, raw GPS L1 intermediate frequency data collected by CYGNSS satellites are used to assess the adaptive hybrid carrier tracking algorithm performances. The results show that robust, accurate carrier tracking at cm-level precision is possible using a hybrid carrier tracking architecture. Analysis of the tracking results confirmed that low signal grazing angle, high antenna gain, and low wind speed over the ocean are necessary conditions for successfully carrier tracking.

This paper presents assessment on the coherency of reflected GPS signals received on a LEO satellite for high accuracy ocean altimetry applications. We analyze the coherent signal levels dependence on sea surface conditions and signal parameters based on GPS reflection data collected by Spire cubeSats between January and April 2019. The data contains carrier phase estimations obtained from an open-loop tracking process sampled at 50Hz for GPS carrier frequency L1 and L2. A conservative coherency quality test based on two circular statistics of the carrier phase measurements, circular-length and circular-kurtosis, is applied to the data. The results show that ~ 1% and ~40% of the events contain sufficient coherent energy for high precision altimetry applications over ocean and sea ice respectively.

This paper presents a support vector machine (SVM) -based detection method to identify coherent reflections in GNSS-R measurements. The data was collected by Spire low-Earth orbit (LEO) satellites during January to April, 2019, and it contains the In-phase (I) and Quadrature (Q) correlation outputs at 50 Hz from open-loop (OL) tracking process. The coherence detection is investigated based on 1100 labelled 1-second data segments. The detection results show that the SVM-based method achieves a detection accuracy of 98.66% and outperforms several other algorithms.

In this paper, we provide an analysis of radio frequency interference for global navigation satellite system reflectometry (GNSS-R) using multi-frequency global positioning system (GPS) data collected on Haleakala, Hawaii during a mountain-top radio occultation and reflection experiment conducted on May 6, 2017. The results show that the interferences on GPS L1 and L2 are continuous wave signals and occur at a 2 Hz rate. The interference on GPS L5 is pulsed signals originating from two aeronautical navigation systems. Moreover, the analysis reveals that the interference on GPS L1 can overwhelm the GNSS-R delay-Doppler map (DDM) if 1 ms coherent integration time is applied. Interference mitigation techniques are suggested for different types of interference on the three GPS frequency bands.

Long-term evolution (LTE) signals are potential signals-of-opportunity for position and navigation, especially in challenging urban and indoor environments. A major challenge is that the LTE signal time-of-arrival (TOA) estimations are susceptible to the multipath propagation effects. In this paper, the multipath estimating delay lock loop (MEDLL), which is originally designed for global positioning system receivers, is applied to LTE signal TOA estimation in multipath environments. We derive the analytical expression of the correlation function for LTE signals and present the procedure for estimating parameters of the detected multipath components. Two initialization methods without and with super-resolution algorithm (SRA) are developed for the MEDLL. Our analyses show that the MEDLL with SRA-based initialization can achieve better multipath resolution, while the one without SRA has less complexity. Extensive simulations involving static multipath scenarios are conducted to examine the statistical TOA estimation performance of the proposed MEDLL with LTE cell-specific reference signal. The simulation results and computational complexity analysis indicate that the proposed MEDLL outperforms the conventional delay lock loop and SRA in term of multipath mitigation performance and computational complexity. Experimental results using real collected LTE signals in urban environments are also provided to demonstrate the effectiveness of the proposed technique for realistic scenarios.

In this paper, we propose a machine learning framework to conduct GNSS-R wind speed retrieval. While the conventional method tries to retrieve wind speed using a single scalar value, the proposed framework is capable of incorporating and employing more features such as DDM and incidence angle. The results show that the proposed framework outperforms the conventional retrieval method with a notable margin.

This paper presents a Kalman filter-based closed-loop carrier phase tracking algorithm and results of coherent GNSS-R signal from ocean and land surfaces using CYGNSS raw IF measurements. The carrier phase measurement from coherent GNSS-R signal allows centimeter level carrier phase altimetry. Closed-loop tracking of coherent reflection GNSS-R signal can provide higher resolution and accuracy estimates of signal parameters for other remote sensing applications, e.g., SNR measurements for land surface monitoring, than non-coherent measurements.

Global Navigation Satellite System (GNSS) Radio Occultation (RO) is an atmospheric remote sensing technique that improves global weather forecasting, climate monitoring, and ionospheric studies. Planetary boundary layer height (PBLH) is a crucial parameter in modeling the troposphere. Space-based GNSS RO has been used in detecting the PBLH with receivers onboard low earth orbit satellites. This paper presents a method of PBLH detection using GNSS signal amplitude measured by a mountaintop-based RO (MRO) system on the summit of Haleakala, Hawaii. The estimated PBLHs are comparable with those derived from space-based RO measurements, space-borne lidar, and local radiosonde profiles. With advantages such as having dense temporal and spatial coverage, low-cost, and an easy-to-implement algorithm, the MRO-based signal amplitude method can be a useful addition to existing methods and could contribute to regional weather study.

This paper applies a machine learning (ML) algorithm based on the multi-hidden layer neural network (MHL-NN) for ocean surface wind speed estimation using global navigation satellite system (GNSS) reflection measurements. Unlike conventional wind speed retrieval methods that often depend on limited scalar delay-Doppler map (DDM) observables, the proposed MHL-NN makes use of information captured by the entire DDM. Both simulated and real data sets are used to train and evaluate the performance of the MHL-NN and compare it to a conventional wind speed retrieval method and other prevailing ML algorithms. The results show that the MHL-NN algorithm outperforms the other methods in terms of the root mean square error (RMSE) and mean absolute percentage error (MAPE) of the wind speed estimation.

In recent years, campaigns have been launched to perform GNSS-reflectometry on a global scale using receivers on low-Earth orbiting satellite constellations. As the specular points of reflection track out paths on the surface of Earth, it is possible that the glistening zones of two different transmitter-receiver pairs will overlap, giving two independent measurements of the common surface. These types of measurements could potentially help improve measurements of geophysical parameters, validate these measurements, resolve measurement ambiguities, or calibrate reflectometry receivers. This paper proposes a definition of the common surface scenario and describes the results of a simulation study to calculate the probability of its occurrence for different receiver constellations.

Throughout the past several decades, reflected GNSS signals have been utilized to measure properties of the ocean surface. Although spaceborne receivers are often used to study surface properties, they are limited in their ability to identify and quantify error sources due to size, weight and power limitations. In this paper, we describe a mountaintop ocean reflectometry experiment performed in May 2017 on the summit of Haleakala in Maui, Hawaii. A high-gain dish antenna is steered to follow rising and setting GNSS satellites and collect their direct and reflected signals. Delay-Doppler maps from GPS L1 and L5 signals are compared and altitude estimations obtained from L1 and L5 delay waveforms are presented. By combining L1 and L5 altitude estimations, the results are close to the true receiver altitude. These results demonstrate the value of the mountaintop experiment for GNSS-reflectometry research and dual-frequency processing techniques.

The conventional spaced-receiver approach uses amplitude scintillations to estimate equatorial ionospheric irregularity drift velocities. This approach is less applicable at high latitudes where there is a lack of substantial amplitude scintillations. This paper presents a method to estimate ionosphere irregularity horizontal drift velocities based on GPS signal carrier phase measurements. Joint time-frequency analysis of the carrier phase measurements using an adaptive periodogram technique generates time-varying spectrograms of ionospheric irregularity-induced phase fluctuations. Cross correlation of the spectrograms between antenna pairs provides time lag information on propagating radio signals through the same ionospheric structure. The time lag information is combined with known positions of the receiver array, satellite orbits, and assumed irregularity altitude to infer ionospheric irregularity horizontal drift velocity. This paper presents the methodology and demonstrates its feasibility using data collected by a GPS receiver array at Gakona, Alaska. The potential error sources of this method are also analyzed.

In this letter, we apply energy detection based time- of-arrival (ToA) estimation to multi-band orthogonal frequency-division multiplexing signals. We analyze the mistiming performance of the ToA estimator in the Nakagami-m channel. Analysis shows that the slope of the probability of mistiming curve increases with the number of subbands and the Nakagami-m parameter. This is known in communications as diversity. Simulations are carried out in various channels to corroborate our theoretical analysis.