Space weather describes varying conditions between the Sun and Earth that can degrade Global Navigation Satellite Systems (GNSS) operations. Thus, these effects should be precisely and timely corrected for accurate and reliable GNSS applications. That can be modeled with the Vertical Total Electron Content (VTEC) in the Earth’s ionosphere. This study investigates different learning algorithms to approximate nonlinear space weather processes and forecast VTEC for 1 h and 24 h in the future for low-, mid- and high-latitude ionospheric grid points along the same longitude. VTEC models are developed using learning algorithms of Decision Tree and ensemble learning of Random Forest, Adaptive Boosting (AdaBoost), and eXtreme Gradient Boosting (XGBoost). Furthermore, ensemble models are combined into a single meta-model Voting Regressor. Models were trained, optimized, and validated with the time series cross-validation technique. Moreover, the relative importance of input variables to the VTEC forecast is estimated. The results show that the developed models perform well in both quiet and storm conditions, where multi-tree ensemble learning outperforms the single Decision Tree. In particular, the meta-estimator Voting Regressor provides mostly the lowest RMSE and the highest correlation coefficients as it averages predictions from different well-performing models. Furthermore, expanding the input dataset with time derivatives, moving averages, and daily differences, as well as modifying data, such as differencing, enhances the learning of space weather features, especially over a longer forecast horizon.

The sudden increase of X-radiation and EUV emission following solar flares causes additional ionization and increased absorption of electromagnetic (EM) waves in the Earth’s atmosphere. The solar flare impact on the ionosphere above Europe on 05 and 06 December 2006 was investigated using ground-based (ionosonde and VLF) and satellite-based data (Vertical Total Electron Content (VTEC) derived from GNSS observations and VLF measurements from DEMETER satellite). Based on the Kp and Dst indices, 05 December 2006 was a quiet day, while there was a geomagnetic storm on 06 December 2006. The total fade-out of the EM waves emitted by the ionosondes was experienced at all investigated stations during an X9 class flare on 05 December 2006. The variation of the fmin parameter (first echo trace observed on ionograms, it is a rough measure of the “non-deviative” absorption) and its difference between the quiet period and during the flares have been analyzed. A latitude dependent enhancement of fmin (2–9 MHz) and Δfmin (relative change of about 150%–300%) was observed at every station at the time of the X9 (on 05 December) and M6 (on 06 December) flares. Furthermore, we analyzed VTEC changes during and after the flare events with respect to the mean VTEC values of reference quiet days. During the X9 solar flare, VTEC increased depending on the latitude (2–3 TECU and 5%–20%). On 06 December 2006, the geomagnetic storm increased ionization (5–10 TECU) representing a “positive” ionospheric storm. However, an additional peak in VTEC related to the M6 flare could not be detected. We have also observed a quantifiable change in transionospheric VLF absorption of signals from ground transmitters detected in low Earth orbit associated with the X9 and M6 flare events on 05 and 06 December in the DEMETER data. Moreover, amplitude and phase of ground-based, subionospherically propagating VLF signals were measured simultaneously during the investigated flares to analyze ionosphere reaction and to evaluate the electron density profile versus altitude. For the X9 and M6 flare events we have also calculated the ionospheric parameters (sharpness, reflection height) important for the description and modelling of this medium under forced additional ionization.

Machine learning (ML) has been increasingly applied to space weather and ionosphere problems in recent years, with the goal of improving modeling and forecasting capabilities through a data‐driven modeling approach of nonlinear relationships. However, little work has been done to quantify the uncertainty of the results, lacking an indication of how confident and reliable the results of an ML system are. In this paper, we implement and analyze several uncertainty quantification approaches for an ML‐based model to forecast Vertical Total Electron Content (VTEC) 1‐day ahead and corresponding uncertainties with 95% confidence intervals (CI): (a) Super‐Ensemble of ML‐based VTEC models (SE), (b) Gradient Tree Boosting with quantile loss function (Quantile Gradient Boosting, QGB), (c) Bayesian neural network (BNN), and (d) BNN including data uncertainty (BNN + D). Techniques that consider only model parameter uncertainties (a and c) predict narrow CI and over‐optimistic results, whereas accounting for both model parameter and data uncertainties with the BNN + D approach leads to a wider CI and the most realistic uncertainties quantification of VTEC forecast. However, the BNN + D approach suffers from a high computational burden, while the QGB approach is the most computationally efficient solution with slightly less realistic uncertainties. The QGB CI are determined to a large extent from space weather indices, as revealed by the feature analysis. They exhibit variations related to daytime/nightime, solar irradiance, geomagnetic activity, and post‐sunset low‐latitude ionosphere enhancement.

The ionospheric refraction of GNSS signals can have an impact on positioning accuracy, especially in cases of single-frequency observations. Ionosphere models that are broadcasted by the satellite systems (e.g., Klobuchar, NeQuick-G) do not include enough details to permit them to correct single-frequency observations with sufficient accuracy. To address this issue, regional ionosphere models (RIMs) have been developed in several countries in the western Balkans based on dense Continuous Operating Reference Stations (CORS) observations. Subsequently, a RIM for the western Balkans was built using an artificial neural network that combined regional ionosphere parameters estimated from the CORS data with spatiotemporal (latitude, longitude, hour of day), solar (F10.7) and geomagnetic (Kp, Dst) parameters. The RIMs were tested at the solar maximum (March 2014), a geomagnetic storm (March 2015), and the solar minimum (March 2018). The new RIMs mimic the integrated electron density much more effectively than the Klobuchar model. Furthermore, RIMs significantly reduce the ionospheric effects on single-frequency positioning, indicating their necessity for use in positioning applications.

Activities on the Sun’s surface can produce dynamic conditions in the Earth’s outer space environment, which can affect the Earth, space-borne and ground-based technologies, including Global Navigation Satellite Systems (GNSS). Delay of GNSS signal can occur during its propagation through the upper Earth’s atmosphere—the ionosphere, representing the major limitation in GNSS positioning applications. In this paper, high level of solar activity and intense bursts of radiation from the release of magnetic energy on the Sun, known as solar flares, are studied. The investigation covers the detection of events on the Sun’s surface, conditions in near-Earth’s space environment, geomagnetic field, ionosphere and GNSS positioning estimates. In October 2014, more than 200 solar flares were detected and about a quarter of total amount belonged to solar flares of M and X class. Impact on ionospheric layers is studied: D layer with SuperSID (sudden ionospheric disturbances) monitor and electron density to F2 layer with GNSS-derived total electron content. Used GNSS stations belong to EUREF Permanent Network (EPN) in Bosnia and Herzegovina and Croatia. Precise Point Positioning is performed in the Bernese GNSS Software. Solar radio emissions were high in the second half of the month, when more M and X solar flares occurred. Ionospheric electron density was enhanced, reaching its peak during the high level of solar activity and the period of strongest solar flares occurrence, while position estimates show higher deviations from the EPN weekly solution in Up component (at least for two times). Higher-order ionospheric terms remained after applying the L3 ionosphere-free solution, which should be taken into account in precise positioning during increased level of solar activity.