Investigation of ionospheric response to a moderate geomagnetic storm over the mid-latitude of Saudi Arabia

1 Introduction

The ionosphere, an ionized layer of Earth’s upper atmosphere, extends from approximately 60 to over 1,000 km in altitude, enveloping our planet. This region is primarily formed through the photoionization of neutral molecules by extreme ultraviolet (EUV) and soft X-ray radiation from the sun. The ionosphere is stratified into four distinct sublayers based on vertical electron density profiles: the D-region (60–100 km), the E-region (100–150 km), the F1-region (150–250 km), and the F2-region ( > 250 km). The electron density increases with altitude, reaching a peak in the F2 region before gradually decreasing (Schunk and Nagy 2009). The ionosphere plays a crucial role in space weather due to its complex interaction with the magnetosphere and its vulnerability to various storm characteristics (Pulkkinen 2007). These characteristics significantly impact radio wave communication, as active auroras can entirely block radio signals, especially at higher frequencies (Lanzerotti 2001). The global positioning system (GPS) depends on signals that pass through the ionosphere, subjecting them to refraction and deceleration, particularly in regions with intense auroral currents. This signal delay, proportional to the electron density along the signal path, provides a unique opportunity for estimating total electron content (TEC). TEC, measured between satellites and receivers, is a fundamental ionospheric parameter that significantly influences the performance of communication and navigation systems. GPS-TEC measurements offer a precise tool for continuously monitoring ionospheric morphology (Akala et al. 2013). Based on geomagnetic latitude (GLAT), Earth’s ionosphere can be divided into three zones: low-latitude (GLAT < 3 0 ∘ ), mid-latitude ( 3 0 ∘ > GLAT < 6 0 ∘ ), and high-latitude ( GLAT > 6 0 ∘ ) (Hunsucker and Hargreaves 2007). Diurnal and seasonal variations in ionospheric electron density at all latitudes are closely related to thermospheric circulation, neutral temperature, ion production rates, and neutral wind patterns (Rishbeth et al. 2000). The equatorial and low-latitude regions of the ionosphere exhibit significant temporal and spatial variations. Above 150 km in the F-region, three distinct dynamical processes emerge: the equatorial ionization anomaly (EIA), post-sunset plasma enhancement (PS-EIA), and evening plasma bubbles (Takahashi et al. 2014). Ionospheric crests arise from upward plasma drift over the magnetic equator in the E and F regions, influenced by the ionospheric fountain effect. Neutral winds at E-region altitudes regulate these crests’ strength and latitudinal extent (Bailey and Balan 1996). Post-sunset, the pre-reversal increase in the zonal electric field elevates the F-layer, intensifying the plasma over the EIA crest (Wang et al. 2012). Geomagnetic storms in mid- and low-latitude regions yield positive and negative ionospheric effects. However, during the initial and main phases of geomagnetic storms, positive effects dominate the ionospheric response (Wang et al. 2010). This positive phase in the middle- and low-latitude ionosphere results from diverse factors, including E → × B → drift-induced F2-layer elevation, plasma fluxes from the plasmasphere, and neutral atmospheric gases downwelling (Danilov and Konstantinova 2013). Various irregularities, such as TIDs, equatorial plasma bubbles (EPBs), and F-region plasma density enhancements, manifest as responses to geomagnetic storms (Saito et al. 2007, Tsugawa et al. 2004). Research employing TEC derived from the GPS and ground-based instruments such as airglow imagers has significantly advanced our understanding of the propagation characteristics of TIDs and their influence on radio wave propagation during geomagnetic storms (Moral et al. 2019).

For instance, Habarulema et al. (2015) utilized GPS-TEC to investigate TIDs moving poleward during geomagnetic conditions over the African sector, illustrating the global propagation of TIDs and their energy and momentum transfer during geomagnetic storms. Jonah et al. (2018) examined ionospheric perturbations associated with TIDs over North America during a geomagnetic storm, identifying TIDs propagating poleward linked to enhanced geomagnetic activity. Numerical simulations, as demonstrated by Horvath and Lovell (2010), have shown that high-latitude auroral events can heat the thermosphere and induce TID generation moving equatorward with varying characteristics. The impacts of TIDs on radio wave propagation and space weather forecasting were thoroughly examined in studies by Hernández-Pajares et al. (2006). In this study, we investigate the effects of moderate geomagnetic storms on the ionosphere and the potential propagation of TIDs across the Saudi Arabian sector using GPS-TEC measurements obtained from three ground-based sites in the Northern Arabian Peninsula. Several studies have concentrated on the ionospheric properties and TEC enhancements in response to geomagnetic activity over the Kingdom of Saudi Arabia (KSA). One study conducted by Alothman et al. (2011) utilizing a regional network of 16 sites encompassing KSA and its surroundings explored short-term ionospheric variations. This study observed peak VTEC values within the 17–26 TECU range. A further investigation utilized GPS data collected at the low-latitude RASH station in Magna, KSA (coordinates: 2 8 ∘ 17 ′ N , 3 4 ∘ 47 ′ E ), to scrutinize temporal variations in ionospheric TEC from 2015 to 2017 (Sharma et al. 2020b). This analysis established a robust correlation between GPS-VTEC observations and key solar indices, including sunspot number (SSN), F10.7, and extreme ultraviolet (EUV) – GPS-derived ionospheric TEC variability. To enhance the monitoring and prediction of the ionosphere over KSA, the Northern Border University set up three ground-based GPS ionospheric monitoring stations in the northern region of the Arabian Peninsula, near 3 0 ∘ N latitude. Utilizing data from the recently established ionospheric monitoring network, we conducted an in-depth analysis of the ionospheric response to a moderate storm event on May 14, 2019. According to a recent study conducted by Chernogor et al. (2020), it was observed that this particular storm led to significant enhancements in VTEC over China during a positive ionospheric storm.

2 Data and methodology

In the period leading up to the geomagnetic storm under study, a series of explosions near sunspot AR2741 on May 13, 2019, instigated the release of three minor coronal mass ejections (CMEs) associated with solar flares directed toward Earth. The most significant eruption transpired on May 12, when an unstable magnetic filament surrounding a sunspot ruptured, resulting in a faint series of CMEs. Activities on May 11 and 13 also contributed to this sequence of events. Although these impending CMEs were relatively minor in scale compared to the bright and massive ones observed during the solar maximum, NOAA forecasters initially predicted a 55–60% probability of G1-class geomagnetic storms on May 15 and 16, 2019, with the potential for isolated G2-class storms. Contrary to the forecast, a potent geomagnetic storm of G3-class was observed on May 14, 2019, as reported by Tassev et al. (2019). Figure 1 displays various geomagnetic parameters with universal time to visually represent the prevailing space weather conditions. During May 13 and 14, 2019, the solar wind bulk speed (Vsw) gradually escalated, peaking at 574 km/s at 08:00 UT on May 14, 2019, before oscillating around 400 km/s. Simultaneously, the southward component of the interplanetary magnetic field (Bz-GSM) fluctuated between − 14 nT and +9 nT.

Figure 1

Southward interplanetary magnetic field (Bz), disturbance storm time (Dst) index, interplanetary solar wind speed (Vsw), interplanetary electric field component (Ey), and geomagnetic three-hourly Kp index were arranged from top to bottom on the panel from 13 to 15 May 2019.

On May 14, 2019, the interplanetary electric field component (Ey) suddenly increased, reaching its maximum of 6.76 mV/m at 05:00 UT. The geomagnetic Kp index reached a value of 6. Notably, the main phase of the storm, which occurred between 03:00 UT and 07:00 UT, exhibited a swift decline in the Dst index, transitioning from 1 nT to − 70  nT. The recovery phase persisted from 07:00 UT on May 14, 2019, until 07:00 UT on May 15, 2019. The data regarding the Vsw, Bz, Ey components, the geomagnetic Kp index, and the Dst index were procured from the OMNI Web-NASA server. In this study, GPS observations were collected from three ground stations located in the Northern Arabian Peninsula, with the local time zone of UTC+3, approximately at latitudes 2 9 ∘ , 3 1 ∘ , and 3 2 ∘ , as illustrated in Figure 2. At these three stations, we utilized the connected autonomous space environment sensor (CASES), a state-of-the-art dual-frequency GPS receiver developed by ASTRA, to capture essential ionospheric information. Data Processing Software: The data processing in this study was performed using Calcula_roti_v411, a software developed by R. Fleury, available on the GIRGEA website. The VTEC was derived based on phase measurements using the following methodology:

• Phase jump detection: Phase jumps were identified when a difference exceeding 5 TECU occurred between consecutive 30 s points. Slant total electron content (STEC) data underwent polynomial regression to generate smoother curves and pinpoint these non-physical jumps.

• Bias incorporation: Adjustments were made for satellite and receiver biases by referencing data from the University of Bern and using VTEC/CODG data in IONEX format. This process involved the application of the secant law, given by the following equation:

  (1) STEC = VTEC 1 − a a + h m cos ( β ) 2 ,

  where a denotes the Earth’s radius (6370 km) and β represents the elevation angle.

• Elevation angle calculation: Using weekly almanac files, GPS satellite positions were determined to calculate elevation angles β , considering positions at both ends. Only data points with elevations above 20° degrees were used to compute receiver bias.

• VTEC calibration and mapping: The STEC resulting from the phase was calibrated, and the secant law was employed to compute the VTEC at each data point. CODG files were manually downloaded from the CDDIS server. Additionally, YUMA almanacs were obtained from the designated site.

where VTEC fit denotes the fitted VTEC at time t , with i = 1 , 2 , 3 representing the receiver stations, and j = 1 , 2 , … , 32 corresponding to the satellites. The term ε represents the residual error from the fitting process, and the coefficients a , b , c , and d are determined using the least-squares method (Habarulema et al. 2016). High-frequency changes in VTEC due to irregular ionospheric plasma were detected by finding the difference between the actual VTEC data and the modeled data, as described by the following equation:

where Δ VTEC indicates the deviation of VTEC, with VTEC being the observed VTEC data at time t , and VTEC fit represents the modeled VTEC data. This approach effectively visualizes the ionospheric response during the storm period, capturing spatial and temporal variations despite using a limited number of stations. To evaluate the impact of CMEs and solar flares on the ionosphere, we calculated the percentage deviation of Δ VTEC between storm days VTEC storm and quiet days VTEC AV 5QD using the following equation:

In addition to the GPS-TEC measurements, VTEC data from the International Reference Ionosphere (IRI) were utilized for comparison. The IRI model offers predictions of VTEC by incorporating advanced models for the F2 peak height hmF2 and improved representations of topside ion densities at various solar activities (Bilitza et al. 2017). We utilized specific configurations for VTEC predictions based on IRI2016. The “NeQuick” model determined the topside electron density (Ne), employing the “Gul-1987ABT-2009” option for the lower atmosphere and the “URSI” option for the Ne-F peak.

Figure 2

Location of the three GPS stations, RFHA ( 2 9 ∘ 38 ′ 1 7 ″ N, 4 3 ∘ 31 ′ 3 4 ″ E), ARAR ( 3 0 ∘ 54 ′ 5 8 ″ N, 4 1 ∘ 04 ′ 4 2 ″ E), and TRIF ( 3 1 ∘ 39 ′ 4 2 ″ N, 3 8 ∘ 43 ′ 4 1 ″ E).

3 Results

Our investigation focused on the ionospheric response within the Northern Arabian Peninsula region during geomagnetic activity, specifically examining the effects of a moderate storm that occurred on May 14, 2019. Figure 3 presents 24 h VTEC plots for all pseudorandom noise numbers (PRNs), illustrating the mean VTEC and providing a snapshot of the daily VTEC variation at the RFHA station on May 14, 2019. Notably, VTEC values gradually increased starting at 06:00 UT, peaking at approximately 28 TECU around 10:00 UT and persisting around noon UT. As evening approached, VTEC values decreased, with a slight post-sunset enhancement – a consistent pattern observed across all stations.

Figure 3

24 h plot of VTEC for all the PRNs (colored lines) and the mean VTEC (black line) on 14 May 2019, RFHA station.

Figure 4 illustrates the DTEC variation over time for all PRNs, with an average DTEC curve superimposed for the RFHA station. This analysis involved fitting VTEC data to a fourth-order polynomial to remove diurnal variability, producing DTEC time series. This method aligns with (Valladares et al. 2009) approach, which reflects variations in integrated density due to large-scale atmospheric gravity waves.

Figure 4

DTEC variation with time from all PRNs in the colored curves, and the average Δ TEC in the black curve over RFHA station on 14 May 2019.

To understand the storm’s impact on VTEC, we considered variations in VTEC from May 12, 2019 (before the storm) to May 16, 2019 (the end of the storm’s recovery phase). Additionally, we independently calculated the average VTEC for five quiet days VTEC AV 5QD for each station. We incorporated the predicted VTEC from the IRI-2016 model (IRI-VTEC) for the same day and location. As depicted in Figure 5, our findings show a nearly consistent pattern in IRI-VTEC values across the three stations. On quiet days, the IRI model tends to overestimate VTEC values, which can be attributed to inaccuracies in representing the topside ionosphere and meridional winds in this region (Tariku 2019). During Geomagnetic activity conditions, however, underestimated VTEC at the start and end of the day and overestimated VTEC between 04:00 and 16:00 UT (Reddybattula and Panda 2019). Underscoring the limitations in the IRI-2016 model’s ability to accurately predict ionospheric behavior during stormy conditions. Additionally, it is important to highlight a significant observation regarding GPS-VTEC values on May 14, notably surpassing those of May 13 and 15. This output underscores the transient nature of ionospheric conditions during storm events. Furthermore, we noted some discrepancies in VTEC AV 5QD values among the stations, likely due to variations in internal receiver bias as well as differences in the latitudinal locations of the stations. The ( σ ) of the VTEC AV 5QD was plotted alongside the storm-time VTEC, revealing peaks during dawn and dusk.

Figure 5

Daily variation of GPS-TEC and IRI-TEC on May 14, 2019, for RFHA, ARAR, and TRIF stations. Red, black, blue, and green lines denote GPS-TEC, VTEC AV5QD , IRI-VTEC, and STD of VTEC AV5QD , respectively.

We calculated the change in TEC ( Δ TEC), which allowed us to estimate disturbances in the ionosphere. Figure 6 illustrates the percentage deviation of Δ TEC from GPS measurements at RFHA, ARAR, and TRIF stations during the moderate geomagnetic storm on May 14, 2019. Following the arrival of a shock wave associated with a weak CME on May 14, we observed a positive increment in Δ TEC during the main phase around 10:00 UT at all three stations. The recorded Δ TEC values were 75, 95, and 120% for RFHA, ARAR, and TRIF, respectively.

Figure 6

Percentage deviation of Δ TEC from the GPS measurements at the sits, RFHA, ARAR, and TRIF arranged from up to down, during the moderate geomagnetic storm on 14 May 2019.

Finally, for Figure 7, the detrended TEC over within the longitude sector of 3 5 ∘ – 4 5 ∘ E and latitude range 2 2 ∘ S– 3 6 ∘ N against universal time (UT) was constructed at a temporal resolution of 30 min, using the DTEC measurements from the three stations. To assess the characteristics of TIDs, we applied the Morlet continuous wavelet transform (CWT), as depicted in Figure 8. Our analysis reveals the persistent existence of poleward medium-scale TIDs throughout the entire day. The power spectrum of the wavelet transform over the Arabian sector was computed from different latitudes. The ( Δ TEC) was computed by subtracting TEC from values of fourth-order fitted polynomial functions and applying a moving average of 30 min for 2 5 ∘ ± 2 ∘ and 3 5 ∘ ± 2 ∘ latitude. From the averaged Δ TEC, the CWT was applied for each latitude to produce scalograms. Estimated periods are approximately 60 and 30 min on May 14, 2019. In summary, this study provides comprehensive insights into the ionospheric response to geomagnetic storms in the Northern Arabian Peninsula. It highlights the limitations of the IRI-2016 model and sheds light on the complex dynamics of the ionosphere, including positive ionospheric storms, post-sunset enhancement, and TIDs during stormy conditions.

Figure 7

Latitude-time plot of DTEC within longitude sector of 3 5 ∘ – 4 5 ∘ E and latitude range 2 2 ∘ S– 3 6 ∘ N. Black dots show the signature propagation of the TIDs the Northern Arabian Peninsula sector on 14 May 2019.

Figure 8

30 min moving averaged Δ TEC from selected latitudes of + 2 5 ∘ and + 3 5 ∘ within the longitude sector of 3 5 ∘ – 5 0 ∘ E on the 14 May 2019. The period of TIDs is ≈ 30  min to ≈ 60  min.

4 Discussion

In this study, we conducted a comprehensive investigation into the ionospheric response to a moderate geomagnetic storm that occurred on May 14, 2019. We utilized GPS observations from three ground stations (RFHA, ARAR, and TRIF) in the northern Arabian Peninsula region. During periods of low geomagnetic activity, we observed a daily variation in VTEC, with peak values occurring around noon. This pattern is consistent with expected trends, driven by increased ionization rates due to maximum solar radiation. However, during the geomagnetic storm, significant increases in VTEC were noted during the main phase. This underscores the transient and dynamic nature of ionospheric conditions during storm events. The increase in VTEC levels at noon during the storm conditions over Saudi Arabia can be attributed to the combined influence of the prompt penetration electric field (PPEF) and the disturbance dynamo electric field (DDEF), as highlighted by Sharma et al. (2020a). The PPEF, which originates from magnetospheric interactions with the solar wind, is triggered by the southward turning of the interplanetary magnetic field (IMF-Bz) and propagates rapidly from high latitudes to low latitudes. This electric field penetrates the low-latitude ionosphere due to an undershielding condition and causes an uplift of the ionosphere through the eastward E × B drift. This process raises the F2 layer, leading to an increase in parameters like the hmF2 and total electron content (TEC) (Kikuchi and Araki 1979; Nava et al. 2016). On the other hand, the DDEF operates over a longer duration and is driven by thermospheric wind disturbances caused by Joule heating at high latitudes during the storm (Blanc and Richmond 1980). These winds flow equatorward, further influencing the ionospheric dynamics by uplifting the F2 layer. The combined effects of PPEF and DDEF substantially enhance the O/N2 ratio, reducing recombination rates and contributing to the observed increase in VTEC (Burns et al. 1995). In summary, the simultaneous presence of both PPEF and DDEF during geomagnetic storms plays a crucial role in elevating the F2 layer and enhancing VTEC, as observed in our study during the May 14, 2019, storm. Our study conducted an evaluation of the IRI-2016 model, focusing on its performance during geomagnetic storms. The comparisons revealed significant limitations in the model’s ability to capture the transient and dynamic nature of the ionosphere during such disturbed periods. Specifically, the IRI-2016 model consistently underestimated VTEC at the start and end of the day and overestimated VTEC between 04:00 and 16:00 UT. These discrepancies align closely with findings from at a nearby station (HALY, 2 9 ∘ N) and (Sharma et al. 2018) at station (BHR3, 2 6 ∘ N), reinforcing concerns about the model’s performance during storm conditions. One major limitation of the IRI-2016 model is its inability to fully account for the effects of geomagnetic storms, particularly the rapid changes caused by PPEF and DDEF. These electrodynamic forces can cause significant, short-term variations in ionospheric parameters, such as VTEC, that the model, with its smoother statistical approach, fails to accurately represent. The IRI2016 model relies heavily on climatological averages, which are not sufficient to capture the sudden and sharp changes in ionospheric electron content driven by storms (Bilitza et al. 2017). Furthermore, the model struggles to reflect the spatial variability in VTEC, where storm-time dynamics are highly influenced by thermospheric winds and electric field disturbances (Lu et al. 2008). The overestimation of VTEC during daytime hours and underestimation at dawn and dusk highlight the IRI-2016’s limited ability to simulate the complex interactions between solar radiation, geomagnetic activity, and ionospheric behavior, which can vary significantly with time and location. In summary, while the IRI-2016 model provides a useful baseline for quiet ionospheric conditions, its limitations in accurately forecasting ionospheric responses during geomagnetic storms suggest a need for more dynamic, storm-specific modeling approaches that can better capture the real-time impacts of geomagnetic disturbances. Our investigation provided valuable insights into the occurrence of poleward TIDs, which are most probably medium-scale TIDs (MSTIDs). However, further investigation is needed to confirm this assumption. Moreover, our findings revealed the complex origins of these disturbances, emphasizing the intricate nature of ionospheric responses during geomagnetic storms. These disturbances, persistently observed throughout the day, highlighted the multifaceted mechanisms involved in their generation. Poleward MSTIDs during geomagnetic storms are primarily driven by alterations in ionospheric electrodynamics, particularly changes in the vertical E → × B → drift following the storm’s onset (Habarulema et al. 2015). The initiation of these MSTIDs is closely linked to the enhancement of equatorial electrojet (EEJ) (MacDougall et al. 2011). However, it is essential to consider other potential physical mechanisms, such as atmospheric gravity waves and magnetosphere–ionosphere–atmosphere coupling, which may also contribute to the launching of TIDs at the geomagnetic equator during disturbed conditions (Hocke et al. 1996; Jonah et al. 2020).

5 Conclusion

This study provides a detailed investigation into the ionospheric response to the geomagnetic storm of May 14, 2019. A significant increase in VTEC was observed during the main phase of the storm, driven by a combination of PPEF and DDEF. These mechanisms led to significant uplift of the ionosphere, particularly around noon, resulting in a substantial increase in VTEC. The analysis highlighted the limitations of the IRI-2016 model, which consistently overestimated VTEC during both geomagnetically quiet and disturbed periods. This underperformance points to the need for further long-term statistical studies to comprehensively assess the model’s predictive accuracy for the ionosphere in this region. Additionally, continuous poleward TIDs were detected throughout the day, indicating the complex nature of ionospheric dynamics during geomagnetic storms. These disturbances, driven primarily by changes in the E → × B → drift, emphasize the importance of conducting further studies on TIDs to better understand the underlying mechanisms that influence ionospheric electrodynamics during such events. Post-sunset VTEC enhancements were also observed during both quiet and storm conditions, attributed to PRE of the east–west electric field in the equatorial F-region and the disturbance dynamo mechanism. These findings align with previous studies conducted in neighboring regions. In conclusion, this study demonstrates the effectiveness of the new installed GPS stations (RFHA, ARAR, TRIF) in detecting ionospheric variations triggered by geomagnetic storms. It also underscores the need for more comprehensive research to better understand the complex ionospheric dynamics over the Arabian Peninsula.