Ionospheric scintillation is a major factor limiting the performance and reliability of Global Navigation Satellite Systems (GNSS), particularly at high latitudes where auroral precipitation and polar cap plasma dynamics generate rapid and spatially irregular electron density structures. During geomagnetic storms, these effects intensify, leading to pronounced carrier-phase fluctuations, signal fading, and an increased risk of loss of lock. Conventional GNSS scintillation indices, such as the amplitude scintillation index and the phase scintillation index, often provide limited insight into the rapid temporal evolution and frequency-dependent characteristics of storm-time high-latitude scintillation.
This paper presents a GNSS-based estimation approach aimed at improving the detection and characterisation of ionospheric scintillation under disturbed geomagnetic conditions. The method combines long coherent integration with adaptive carrier-phase detrending, dual-frequency cross-spectral feature extraction, and Bayesian state-space estimation applied directly to high-rate GNSS observables. Raw measurements, including carrier phase, Doppler frequency, and carrier-to-noise density ratio (C/N₀), are jointly exploited to enhance sensitivity to rapid phase variations while maintaining robustness in weak-signal environments.
The approach is validated using high-rate, multi-constellation GNSS data collected in Northern Norway during a moderate geomagnetic storm on 15 September 2025. The dataset captures a range of scintillation regimes associated with auroral activity and polar cap patches. The results demonstrate improved phase-tracking continuity, enhanced detection of phase scintillation events, and more stable estimation of scintillation intensity compared with conventional index-based techniques. These findings highlight the potential of the proposed approach to support GNSS monitoring and resilience assessment in high-latitude environments and to contribute to the development of reliable positioning, navigation, and timing services under storm-time ionospheric conditions.

Ionospheric scintillation refers to rapid and random fluctuations in amplitude and phase of Global Navigation Satellite System (GNSS) signals caused by small-scale plasma irregularities These fluctuations pose a significant threat to Advanced Receiver Autonomous Integrity Monitoring (ARAIM) in aviation navigation. Scintillation can invalidate standard ARAIM assumptions by increasing measurement noise and inducing intermittent loss of lock, leading to underestimated protection levels or unreliable fault exclusion outcomes. These effects directly degrade continuity and integrity performance, particularly during severe scintillation events when multiple satellites may be affected simultaneously.
ARAIM fundamentally relies on redundancy and statistical consistency checks of pseudorange observations, typically assuming zero-mean, independent, Gaussian errors with a fixed variance. However, the validity of these assumptions may be compromised under scintillation. Rapid carrier phase variations introduce additional measurement errors, potentially causing cycle slips or loss of lock. The elevated and time-varying noise propagates into code measurements when carrier smoothing is applied, inflating pseudorange errors. Therefore, the resulting pseudorange errors inflate the ARAIM global test statistics and distort its nominal chi-square distribution. Consequently, conventional fixed-variance ARAIM thresholds become mismatched to the actual error statistics, raising the risk of false alarms or missed detections in integrity monitoring. Moreover, fault detection and exclusion may fail when the disturbance cannot be attributed to a single satellite fault.
This paper proposes a scintillation-aware integrity monitoring framework that utilizes carrier phase residuals from Precise Point Positioning (PPP) to quantify carrier phase disturbances induced by scintillation. An adaptive measurement noise model suitable for ARAIM is also introduced. The carrier phase residuals treated as a proxy for unmodeled phase errors, are processed through a dedicated filter to suppress non-scintillation components. The phase error sequence caused by scintillation is further estimated. An error-propagation model is then derived to map the extracted phase disturbance into the variance of the Hatch-smoothed pseudorange. This model explicitly captures the dependence on smoothing window length and phase error statistics. By combining the carrier phase residual and signal quality parameters, an adaptive per-satellite measurement covariance model is formulated. This model is injected into a weighted ARAIM global test statistic. The proposed approach aims to restore the statistical consistency of the ARAIM test under scintillation and improve integrity availability without compromising detection capability. Experimental evaluation is conducted using representative scintillation datasets. The proposed method quantifies the propagation of carrier phase errors to the smoothed pseudorange under ionospheric scintillation. The model is also compared against conventional fixed-noise ARAIM in terms of test statistic behavior.

Precise global navigation satellite system (GNSS) positioning is strongly affected by ionospheric disturbances, which introduce significant errors in signal propagation, especially during periods of increased solar and geomagnetic activity. Advanced positioning techniques, such as Precise Point Positioning (PPP) and Real-Time Kinematic (RTK), mitigate these effects through dual-frequency observations and modeling strategies. However, residual ionospheric errors, especially those related to strong spatial gradients, remain a major challenge to achieving high accuracy and reliability. In this context, the Gradient Ionospheric Index (GIX) is a promising tool for monitoring ionospheric conditions and supporting precise GNSS applications.

The GIX is designed to quantify the horizontal structure of the electron distribution in the ionosphere by identifying spatial changes in total electron content (TEC). High GIX values are associated with ionospheric irregularities, gradients, and disturbances that can degrade GNSS positioning performance, increase ambiguity resolution time, and reduce solution reliability. This presentation explores the potential applications of the GIX index in precise GNSS positioning, particularly its role as an indicator of ionospheric quality and supporting parameter in positioning strategies.

One of the primary applications of GIX is real-time ionospheric monitoring for GNSS users. Integrating GIX information into positioning algorithms makes it possible to identify periods and regions of increased ionospheric risk and adapt processing strategies accordingly. GIX-based thresholds can be used to exclude unreliable observations, adjust stochastic models, or trigger warnings in safety-critical applications, such as aviation, autonomous navigation, and precision surveying. GIX can also support the evaluation of RTK and Network RTK performance by identifying areas where strong ionospheric gradients may hinder the effectiveness of interpolation techniques.

Another promising application is enhancing PPP solutions. GIX can serve as an external quality indicator for ionospheric modeling and ambiguity convergence. This allows for adaptive weighting schemes and dynamic model selection under disturbed ionospheric conditions. Additionally, the index can be employed in post-processing analyses to evaluate positioning accuracy and investigate the relationship between ionospheric activity and GNSS performance degradation.

The potential integration of GIX with multi-GNSS and multi-frequency observations is also discussed, and its relevance in the context of modern GNSS constellations and high-rate data processing is highlighted. In summary, GIX is a valuable diagnostic and supporting tool for precise GNSS positioning that offers new possibilities for improving robustness, reliability, and situational awareness under challenging ionospheric conditions.

Accurate forecasting of ionospheric scintillation remains a critical challenge for ensuring the robustness of Global Navigation Satellite System (GNSS)–based positioning, especially under disturbed space weather conditions. In this study, we develop a multimodal deep-learning network for predicting the aggregated amplitude scintillation index using both recent ionospheric total electron content (TEC) maps and local scintillation observations. The proposed model ingests the four most recent global TEC maps (updated every 15 minutes), where the time span between the oldest and newest TEC map is 45 minutes and the latest TEC map is the one nearest to the prediction time, together with a 15-minute time series of historical scintillation measurements. This input design captures both the evolving large-scale ionospheric background and the high-frequency local fluctuations that jointly drive scintillation dynamics.
The architecture is built on a vision–language multimodal backbone (Qwen3-VL) and incorporates three key contributions. First, a Time Modulation Block injects explicit temporal embeddings into the TEC image sequence so the model can discriminate maps by their relative time offsets. Second, a Reprogramming Layer projects patch-embedded time-series features into the LLM latent space, enabling effective cross-modal fusion while keeping the pretrained backbone weights frozen. Third, we use prompt-based statistical conditioning, which feeds input statistics such as minima, maxima, median, trend, and top autocorrelation lags, as textual prompts to the language pathway to incorporate domain priors into forecasting. The second and the third part is the same as the TimesNet.
Fusion is achieved by concatenating prompt embeddings, time-modulated TEC image embeddings, and reprogrammed time-series embeddings; the large language model decoder then generates future latent representations which are projected to yield the aggregated amplitude scintillation index forecast. Preliminary experiments indicate that jointly modelling recent TEC maps and short-term local scintillation series substantially improves forecast skill compared with single-modality baselines. The approach demonstrates the utility of temporal modulation and reprogramming-based cross-modal fusion for scintillation forecasting.

The European navigation satellite system Galileo transmits multiple signals that contain both the ranging information and various navigation message data. While ranging information enables pseudorange determination, navigation data provide the parameters required to compute complete position, velocity and time solutions. Among the navigation data are also the ionospheric model parameters, which are used as input to correction algorithms to assess the effect on the satellite signals during the ionosphere propagation.
Single frequency Galileo users rely on the NeQuick G model to reduce ionospheric errors. Although NeQuick G offers high performance, it's computational requirements may be too high for simple devices. To address this limitation, the alternative model NTCM G was later proposed.
We present the performance monitoring results obtained in the framework of the GEMOP project.
The analysis covers all latitude zones but with more focus on the European region than other regions.
We evaluate TEC (Total Electron Content) accuracy, Precise/Single Point Positioning (PPP/SPP) accuracy, and PPP convergence times using NTCM G model.
We compare and assess these results against the equivalent results when using the NeQuick G and Klobuchar models.
Assessment metrics based on the official NTCM G definition document are presented and evaluated.
In addition, the validity of the broadcast parameters is examined through correlation with the solar index F10.7.
To assess the optimalization of the broadcast parameters, a perturbation analysis is performed.
Our findings demonstrate that NTCM G delivers similar performance to NeQuick G, both outperforming the Klobuchar model.
Seasonal and latitudinal variations are observed across all models, although their phase and magnitude differ, leading to annual fluctuations in relative performance.
Importantly, NTCM G errors remain within the defined limits, and the broadcast parameters are generally near optimal, with further refinements remaining still possible.
The NTCM G model provides end users with performance equivalent to NeQuick G while requiring fewer computational resources.
Our study confirms NTCM G as a suitable alternative in scenarios where computational resources are limited, thereby extending the accessibility of high accuracy Galileo services to a wider range of applications and devices.