Study on the Three-Dimensional Evolution of Ionospheric Disturbances in China During the Geomagnetic Storm on December 1, 2023

Abstract

On 1 December 2023, a strong geomagnetic storm was triggered by an interplanetary shock caused by a coronal mass ejection (CME). This study used data from 193 Global Navigation Satellite System (GNSS) observation stations in China to study the three-dimensional morphological total electron content (TEC) disturbances during this storm. By analyzing GNSS TEC data from 15 GNSS stations along the magnetic field lines, it was found that TEC disturbances spread from low to high latitudes, confirmed by ionosonde NmF2 data. The TEC disturbance first appeared at the LJHP station, (21.68° N) at 11:30 UT and propagated to the BJFS station (39.60° N) at 13:30 UT with a propagation speed of about 217 m/s and maximum amplitude of ±0.2 m. The TEC disturbance lasted the longest, approximately 4 h, between latitudes 25° N and 32° N. Additionally, this study investigated the ionosphere’s three-dimensional electron density distribution in the Guangxi region using an ionospheric tomography algorithm. Results showed that the TEC disturbances were mainly concentrated between 450 and 580 km in altitude. At 12:00 UT, the maximum change in electron density occurred at a 580 km height at 26° N, 112° E, increasing by 20.54 total electron content unit (TECU). During the main phase of the geomagnetic storm, the electron density expanded from higher to lower layers, while during the recovery phase, it recovered from the lower layers to the higher layers.

1. Introduction

When coronal mass ejections (CMEs) strike the Earth’s magnetosphere, they generate strong currents and electric fields. The interactions between these currents and fields trigger global geomagnetic storms. Storms inject large amounts of energy into the upper atmosphere, causing a rapid increase in the number of ionospherically charged particles. This causes significant changes in the structure and characteristics of the ionosphere, affecting wireless communications, satellite navigation, and power systems, all of which have significant impact on human society. Thus, studying the disturbance characteristics and three-dimensional morphological changes in the ionosphere during geomagnetic storms has important scientific significance and practical applications.

Traditional ionospheric detection methods such as ionospheric vertical sounding and incoherent scatter sounding suffer from limited spatial resolution owing to the scarcity of observation stations, which hinders a comprehensive understanding of subtle changes in the ionosphere. The advent of the Global Navigation Satellite System (GNSS) technology has enabled the acquisition of ionospheric structural characteristics and positioning with high spatial and temporal resolutions over large areas. GNSS-based ionospheric tomography (Computerized Ionospheric Tomography, CIT) has emerged as a focal point for obtaining the spatiotemporal distribution of the ionospheric electron density and its variations. CIT not only reflects horizontal changes in each ionospheric layer but also provides insights into the vertical coupling mechanisms between layers, making it a novel ionospheric detection approach.

In recent years, many scholars have used the GNSS to study ionospheric disturbances during geomagnetic storms. Ho et al. [1] observed large-scale traveling ionospheric disturbances (LSTIDs) during geomagnetic storms using global data from 60 GPS stations, revealing the conjugate relationship between total electron content (TEC) in the Northern and Southern Hemispheres; Hernández-Pajares et al. [2] analyzed the ionospheric response during two geomagnetic storms on 18–19 October 1995, and 10 January 1997, using data from more than 100 global GPS stations and GPS/MET low Earth orbit (LEO) satellites; Tsugawa et al. [3] studied ionospheric disturbances during the geomagnetic storm on 22 September 1999, using GEONET GPS data and presented TEC disturbance images; Wang et al. [4] conducted a global statistical analysis of the geomagnetic storm from 29 to 31 October 2013, based on observational data from more than 900 GPS stations, monitoring disturbances that propagated southward and moved westward; Yin et al. [5] calculated a TEC time series using observational data from GPS stations in North America during geomagnetic storms and analyzed the ionosphere on a plane; Borries et al. [6] studied the impact of geomagnetic storms on the European ionosphere from 2001 to 2007 using GPS station observational data and compared these disturbances with those in Japan; Habarulema et al. [7] obtained propagation parameters of traveling ionospheric disturbances (TIDs) using wavelet analysis and studied TIDs in South Africa during the magnetic storms of 2005 and 2011; Ding et al. [8] used data from GPS stations and ionospheric observation stations in China and Southeast Asia to study two LSTIDs that propagated poleward in China during a moderate storm from 27 May to 1 June 2011. Padokhin et al. [9] developed a global ionospheric model using a phase-difference-based total electron content (TEC) mapping method to monitor global ionospheric disturbances on 2 January 2017; Kuai et al. [10] studied the ionospheric storm response at low latitudes in the Americas and Austral-Asian regions during a geomagnetic storm in May 2021 using GNSS TEC and changes in TEC at the top of LEO satellites.

The aforementioned studies mainly relied on two-dimensional modeling techniques. Since Rius and Cucurull [11] implemented three-dimensional ionospheric tomography using the Kalman filter method, significant advancements in the theory and methods of ionospheric three-dimensional modeling have been made, leading to the development of various GNSS-based ionospheric tomography models. Many studies have used the GNSS and CIT to model ionospheric disturbances, revealing the spatiotemporal variations of the ionosphere during different geomagnetic activities. Yizengaw et al. [12] monitored the impact of a strong magnetic storm on the ionosphere in the Southern Hemisphere on 31 March 2001, using CIT technology based on multi-source data, revealing the characteristics of ionospheric structure; Wen et al. [13] used CIT technology, based on data from the Chinese Crustal Movement Observation Network (CMONOC), to study the impact of a geomagnetic storm on the Chinese ionosphere on 18 August 2003; Nesterov et al. [14] used high orbital radio tomography to perform high-resolution reconstruction of the evolution of the ionospheric trough over Europe on 17 April 2003; Seemala et al. [15] and Saito et al. [16] reconstructed three-dimensional ionospheric tomograms over Japan using CIT technology, monitoring the three-dimensional electron density distribution; Tang et al. [17] studied the spatial and temporal changes in ionospheric electron density (IED) and TEC in the Chinese region during geomagnetic storms, confirming that LSTIDs propagated southward at a horizontal speed of 400–500 m/s and moved southwestward at a horizontal speed of 250–300 m/s. Kong et al. [18] used CIT images to monitor two vertical electron movements during a magnetic storm on 17 March 2015 and discussed the dynamics of these movements. Prol et al. [19] constructed a global ionospheric electron density tomogram of a geomagnetic storm on 17 March 2015, using CIT technology based on data from more than 2700 GNSS stations and monitored plasma changes during geomagnetic disturbances, revealing characteristics such as enhancement and depletion of the equatorial ionization anomaly. Feng et al. [20] reconstructed ionospheric IED distributions using CIT technology based on data from the Crustal Movement Observation Network of China (CMONOC), studying the ionospheric storm effect in the Wuhan area during geomagnetic storms on 17 March and 22 June 2015. Cheng et al. [21] constructed a global-scale three-dimensional (3-D) ionospheric model using CIT technology combined with multi-GNSS and meteorological, ionospheric, and climate constellation observation system (COSMIC) radio occultation (RO) observation data to reveal the response characteristics and evolution of the ionosphere during storms on a global scale. Shan et al. [22] reconstructed the spatial distribution of the tongue of ionization (TOI) that appeared over Greenland during a moderate geomagnetic storm on 11 October 2010, using three-dimensional ionospheric tomography (3DCIT) technology with multi-source data integration. Wang et al. [23] analyzed three-dimensional ionospheric morphological changes and evolution characteristics in the Australian region during a moderate-intensity geomagnetic storm on 25 July 2020 by constructing a three-dimensional tomographic model with integrated GNSS and COSMIC-2 data.

Although numerous studies have revealed the spatiotemporal variations of ionospheric disturbances during geomagnetic storms, these studies have predominantly focused on high-latitude regions or global scales, with most employing two-dimensional modeling techniques. Research on the three-dimensional evolution characteristics of ionospheric disturbances in low-latitude regions remains relatively limited. To address this gap, this study used observational data from 193 GNSS stations to analyze the characteristics and propagation patterns of ionospheric disturbances in China during a geomagnetic storm on 1 December 2023; simultaneously, using a three-dimensional ionospheric tomography algorithm, the three-dimensional morphological changes of the ionosphere in China’s Guangxi region during a geomagnetic storm revealed the three-dimensional evolution characteristics of ionospheric disturbances in low-latitude areas. Specifically, this study addresses the following questions: How do ionospheric disturbances propagate across China? What are the three-dimensional evolution characteristics of ionospheric disturbances in low-latitude regions? By employing dense GNSS data and three-dimensional modeling techniques, this study provides novel insights into the spatial propagation and three-dimensional dynamic evolution of ionospheric disturbances during geomagnetic storms, offering new perspectives and data support for the modeling, monitoring, and prediction of low-latitude ionospheric disturbances.

2. Methods and Data

2.1. GNSS TEC Sequence Algorithm

The inter-epoch TEC variations derived from carrier data have very high precision (GNSS carrier observations are accurate to the millimeter level, approximately 0.03 total electron content unit (TECU), and second-order differencing is an effective means of extracting electron density anomalies. The ionospheric anomalies for the current epoch were detected based on TEC changes between the previous two epochs. The TEC time series at a GNSS station can be represented as

$$TE{C}_1,......,TE{C}_{i-1},TE{C}_i,TE{C}_{i+1},......,TE{C}_n$$

(1)

Equations (2) and (3) were used to determine the ionospheric disturbance at the ith epoch.

$$\begin{array}{l}\mathit{TEC\_TEST}=TECT=TE{C}_i-(TE{C}_{i-1}+TEC{V}_{EST})\\ TEC{V}_{EST}=TE{C}_{i-1}-TE{C}_{i-2}={\displaystyle \frac{f_1^2{f}_2^2}{40.28(f_2^2-f_1^2)}}[{\lambda }_1\Delta L_1(i-1)-{\lambda }_2\Delta L_2(i-1)]\end{array}$$

(2)

$$\begin{array}{l}TECT={\displaystyle \frac{f_1^2{f}_2^2}{40.28(f_2^2-f_1^2)}}[({\lambda }_1\Delta L_1(i)-{\lambda }_2\Delta L_2(i))-({\lambda }_1\Delta L_1(i-1)-{\lambda }_2\Delta L_2(i-1))]\\ {\lambda }_1\Delta L_1(i)={\lambda }_1{L}_1(i)-{\lambda }_1{L}_1(i-1)\\ {\lambda }_2\Delta L_2(i)={\lambda }_2{L}_2(i)-{\lambda }_2{L}_2(i-1)\end{array}$$

(3)

where $f_1$ and $f_2$ are the two carrier-phase frequencies of the signal, $L_1$ and $L_2$ are the carrier-phase observations, ${\lambda }_1$ and ${\lambda }_2$ are the wavelengths of the carrier signals, n is the total number of GNSS carrier observations per epoch, and i corresponds to the GNSS observation epoch. Using the single-layer model (SLM) with the model height set at 350 km and combining the ionospheric pierce point (IPP) coordinates of the dTEC sequence, the arrival time and position of disturbances in each observed arc segment were determined, thereby analyzing the propagation laws of ionospheric disturbances.

GNSS observation data from 15 stations in the Beijing, Hebei, Henan, Hubei, Hunan, and Guangxi regions of China with a data sampling interval of 30 s were selected along the magnetic field direction. The distribution range was 104° E to 120° E and 21–43° N. The estimated GNSS TEC sequences were obtained as shown in Figure 1.

2.2. Ionospheric Tomography Algorithm Based on Functional Basis and Projection Functions

The ionospheric delay is correlated with the electron density along the GNSS signal’s path from satellites to receivers via the ionospheres. The slant total electron content (STEC) can be expressed as the integral of the electron density along the path, as shown in Equation (4). Using projection functions, the STEC was transformed into the vertical total electron content (VTEC), as shown in Equation (5).

$$STEC(t)={\displaystyle {\int }_{s}Ne(s,t)ds}$$

(4)

$$VTEC={\displaystyle \frac{STEC(t)}{PF}}$$

(5)

where $s$ is the ray vector, $t$ is the observation time, $Ne\left(s,t\right)$ is the electron-density function along the path, and $PF$ is the projection function. Equation (4) is expressed in four segments, as shown in Equation (6).

$$\begin{array}{c}STEC\left(t\right)={\displaystyle {\int }_{h=50}^{h=275}Ne}(s,t)ds+{\displaystyle {\int }_{h=275}^{h=375}Ne(s,t)ds}+{\displaystyle {\int }_{h=375}^{h=580}Ne(s,t)ds}\hfill \\ +{\displaystyle {\int }_{h=580}^{h=1000}Ne(s,t)ds}\hfill \\ =VTE{C}_{200}\times P{F}_{200}+VTE{C}_{325}\times P{F}_{325}+VTE{C}_{450}\times P{F}_{450}\hfill \\ +VTE{C}_{580}\times P{F}_{580}\hfill \end{array}$$

(6)

In Equation (6), the electron density within each layer is represented by the height at the midpoint of each segment, and the STEC is converted to the VTEC using the projection function. Each layer was modeled separately, as shown in Equations (7) and (8):

$$\begin{array}{c}VTE{C}_{250}={\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m{A}_{ij}}}{(b-b_0)}^i{(s-s_0)}^j\hfill \\ =a_{11}+a_{12}b+a_{12}s+a_{13}bs+a_{14}{b}^2\hfill \\ s-s_0=(l-l_0)+(t-t_0)\hfill \end{array}$$

(7)

$$MF={\displaystyle \frac{1}{\cos (z^{\prime })}}={\displaystyle \frac{1}{\sqrt{1-{\sin }^2(z^{\prime })}}}$$

(8)

Equation (7) uses a second-order polynomial as an example, where $n$ and $m$ are the coefficients for the latitude and longitude terms, respectively. $S_0$ represents the solar hour angle at the central time of the observation period for the center point $(b_0,l_0)$ of the surveyed area. $b_0$ and $l_0$ are the latitude and longitude of the center point, while $b$ and $l_0$ are the latitude and longitude of the ionospheric pierce point. $a_{11}$ to $a_{14}$ are the parameters to be determined.

Equation (8) represents the SLM mapping function revised by the Center for Orbit Determination in Europe (CODE). $\sin (z^{\prime })=R/(R+H)\times \sin (\alpha z)$, $R=6371 \mathrm{km}$. Where $H$ is the ionospheric height and $a=0.9782$. $z$ is the zenith distance of the receiver and $z^{\prime }$ is the zenith distance of the ionospheric pierce point.

By substituting Equations (7) and (8) into (6), a tomographic algorithm function model [24] with an additional projection function can be derived, as shown in Equation (9).

$$\begin{array}{c}STEC=(a_{11}+a_{12}{b}_{200}+a_{13}{b}_{200}{s}_{200}+a_{14}{b}_{200}^2)\times P{F}_{200}\hfill \\ +(a_{21}+a_{22}{b}_{325}+a_{23}{b}_{325}{s}_{325}+a_{24}{b}_{325}^2)\times P{F}_{325}\hfill \\ +(a_{31}+a_{32}{b}_{450}+a_{33}{b}_{450}{s}_{450}+a_{34}{b}_{450}^2)\times P{F}_{450}\hfill \\ +(a_{41}+a_{42}{b}_{580}+a_{43}{b}_{580}{s}_{580}+a_{44}{b}_{580}^2)\times P{F}_{580}\hfill \end{array}$$

(9)

Using observation data from 182 GNSS stations in Guangxi, we performed ionospheric tomography using an ionospheric tomography algorithm based on the functional basis and projection functions.

We analyzed the three-dimensional morphological changes in ionospheric disturbances in the Guangxi region during geomagnetic storms. The distribution of the 182 GNSS stations is shown in Figure 2.

3. Geomagnetic Storm on December 1, 2023

On 1 December 2023, an interplanetary shock wave triggered by a coronal mass ejection (CME) reached the vicinity of the Earth, resulting in a significant geomagnetic storm. The evolution of this storm can be divided into three phases: an initial phase from 00:00 to 09:00 UT, a main phase from 10:00 to 13:00 UT, and a recovery phase after 14:00 UT. The variations in the solar wind velocity (Vp), the north–south component (Bz) of the interplanetary magnetic field (IMF), the geomagnetic ring current (Dst) index, and the geomagnetic disturbance (Kp) index from 29 November to 4 December 2023, are depicted in Figure 3. Between 23:00 and 24:00 UT on 30 November 2023, the first shock wave struck the magnetosphere, resulting in an immediate enhancement of the geomagnetic field. This enhancement is known as the “Storm Sudden Commencement” (SSC), as indicated by the red dashed line in Figure 3.

During the initial phase, the solar wind velocity surged, peaking at 439 km/s at 01:00 UT on December 1st. It remained relatively stable for the next 9 h, and then increased again, reaching the second peak of 532 km/s at 10:00 UT, about 1.5 times higher than the previous day. The sharp increase indicated heightened solar wind activity, setting the stage for geomagnetic storms. During this phase, the IMF Bz fluctuated in amplitude between north and south, the Kp index abruptly rose to 4, and the Dst index displayed noticeable undulations.

In the main phase, Vp continued at a high speed at around 500 km/s; the IMF Bz sharply turned southward at 10:00 UT, reaching a low of −22.3 nT at 11:00 UT, followed by a sudden northward turn, peaking at 21.6 nT at 14:00 UT. The Kp index remained above 6 throughout this phase, and the Dst index dropped steeply at 10:00 UT, reaching −87 nT at 11:00 UT, and then reaching its minimum of −107 nT at 13:00 UT.

During the recovery phase, Vp continued at approximately 500 km/s until 01:00 UT on 2 December, when it gradually decreased. The IMF Bz component shifted north to south, oscillated between north and south, and gradually stabilized. The Kp index decreased to approximately 5 and remained so until 06:00 UT on 2 December, eventually returning to normal. The Dst index still fluctuated but gradually recovered towards the quiet level of the geomagnetic field. Based on the classification criteria for geomagnetic storms and considering the changes in the Dst and Kp indices during this event, the geomagnetic storm was classified as a strong storm.

4. Experimental Analysis

4.1. Analysis of TEC Disturbance Propagation Patterns

Fifteen GNSS stations were chosen along the direction of the magnetic field in China. The differential total electron content (dTEC) sequences for satellites G8 and G31 were calculated using Equation (3). The characteristics and propagation patterns of ionospheric disturbances triggered by geomagnetic storms were analyzed. Figure 4 shows the dTEC sequence of the G8 satellite.

Prior to the disturbance initiation point, the ionosphere was in a stable state, with relatively flat dTEC sequences across all stations; following the initiation point, the dTEC sequences began to exhibit significant oscillations with amplitudes reaching ±0.2 TECU, from which the time corresponding to the initiation point can be determined as the first occurrence of ionospheric disturbance.

During the main phase at UT 11:00, the IMF Bz reached its lowest value of −22.3 nT, and the Dst index sharply declined to −87 nT. The southward magnetic field drives the injection of high-energy particles from the solar wind into the magnetosphere, enhancing magnetospheric convection and ring currents, leading to a steep decrease in the Dst index and triggering anomalies in the Earth’s space environment. At this time, the LJHP station, located at the lowest latitude (21.68° N), observed an ionospheric disturbance first at 11:30 UT. As time progressed, the disturbance gradually propagated to higher-latitude stations, with the highest latitude station, BJFS (latitude 39.61° N), observing the latest disturbance at 13:30 UT. The propagation time of the disturbance from the LJHP station to the BJFS station was approximately 2 h.

In the recovery phase, which began after 14:00 UT, stations in both high-latitude (32–40° N) and low-latitude (21–25° N) regions showed a return to stable dTEC sequences earlier than those in mid-latitude (25–32° N) regions. This indicates that the ionospheric disturbance persisted longer in areas between 25° N and 32° N, with the longest duration of approximately 4 h, whereas in the high-latitude (32° N–40° N) regions, the disturbance had a shorter duration, with the shortest time of approximately 30 min.

Figure 5 shows the dTEC sequence of the G31 satellite. Compared to Figure 4, the time at which the ionospheric disturbances are observed at each station in Figure 5 was slightly delayed, but the characteristics and propagation patterns of the disturbances are consistent with those in Figure 4. It can be concluded that ionospheric disturbances propagate from south to north along the latitude direction during geomagnetic storms.

To verify the propagation patterns of ionospheric disturbances during geomagnetic storms in China, this study compared the NmF2 values of the ionosonde at the Beijing (CPT), Wuhan (ZLT), and Hainan (FKT) stations on the storm day with the average NmF2 values for seven quiet days. Figure 6 depicts the comparison, showing a black curve representing the average NmF2 values on quiet days, and the gray shaded area indicating the range of two standard deviations (2σ) used to identify ionospheric anomalies. Anomalies were indicated when the measured NmF2 values exceeded this range. In selecting the threshold, this study used 2σ as the standard for anomaly detection. The 2σ threshold is widely used in ionospheric disturbance analysis as it strikes a balance between sensitivity and false-positive rates, making it particularly suitable for identifying significant ionospheric disturbances [17]. In contrast, the 1.34σ threshold, while more sensitive to smaller disturbances, may lead to a higher false-positive rate, especially during geomagnetic storms when the ionospheric background values fluctuate significantly [25]. Therefore, the choice of the 2σ threshold effectively captures the significant disturbances caused by geomagnetic storms while minimizing the interference from background noise induced by solar activity and other factors. Furthermore, although the IQR method [26] performs well in handling non-normal distributions, the 2σ threshold provides a more direct and reliable detection method when dealing with significant disturbances during geomagnetic storms.

It can be observed from the figure that before 11:00 UT, the NmF2 values remain relatively stable within the 2σ range. However, starting at UT 11:00, significant disturbance phenomena appeared successively at the three observation stations, as indicated by the red rectangle. At UT 11:00, the FKT station first exhibited a positive disturbance; at 11:30 UT, the ZLT station also showed a positive disturbance; and finally, at 14:00 UT, the CPT station recorded a significant positive disturbance. This indicates that ionospheric disturbances first appeared in low-latitude areas and then gradually propagated towards high-latitude areas. This propagation trend is essentially consistent with the disturbance propagation patterns shown in Figure 4 and Figure 5, further validating the spatiotemporal propagation characteristics of the ionospheric disturbances.

4.2. Analysis of TEC Disturbance Propagation Speed

Following the propagation pattern of the ionospheric disturbances from low to high latitudes, the propagation speed was estimated using the disturbance occurrence times and the corresponding coordinates of the ionospheric pierce points from various stations in Figure 4, as per Equation (10). The distances corresponding to the disturbance initiation points at each station were calculated using their coordinates, and the associated time difference was $t_1-t_2$.

$$V=\left|{\displaystyle \frac{D([B_1,L_1],[B_2,L_2],R_e)}{t_1-t_2}}\right|$$

(10)

In this equation, $D$ represents the distance between the disturbance initiation points of two stations and $R_e$ denotes the radius of the Earth. Using the disturbance arrival times and coordinates between each pair of stations, the horizontal propagation speed of the ionospheric disturbances caused by satellites G8 and G31 was estimated, as shown in Figure 7, with an average speed of approximately 217 m/s.

In this study, the propagation speed of ionospheric disturbances was estimated by calculating the disturbance onset times of ionospheric piercing points (IPPs) at 15 selected GNSS stations along the magnetic field direction and the distances between these stations. To estimate the propagation speed, it was assumed that the ionospheric disturbances propagate uniformly along the magnetic field direction, meaning the propagation speed is consistent across all stations. However, this assumption does not fully account for the various factors that could influence propagation speed in real-world conditions. First, latitude differences may lead to variations in propagation speed, as ionospheric density and conditions vary between low and high latitudes, with disturbances potentially propagating faster in low-latitude regions. Second, factors such as solar activity and geomagnetic storms may cause variations in the propagation characteristics of ionospheric disturbances over time. Therefore, the average propagation speed calculated in Figure 7 (217 m/s) serves as a rough estimate under specific conditions, and future studies should further consider the impact of these factors on propagation speed to improve the accuracy and applicability of the calculations.

4.3. Analysis of Two-Dimensional Variations of TEC Disturbances

To further analyze the two-dimensional variation characteristics of the ionospheric disturbances over time, we used Equation (3) to obtain the TEC disturbance sequences at the ionospheric pierce point (IPP) above 182 GNSS stations in the Guangxi region. Figure 8 shows the TEC sequence from UT 11:15 to 11:55 UT.

From Figure 8, it can be observed that between 11:25 and 11:55 UT, the TEC disturbances gradually expanded from low-to-high latitude areas. At 11:25 UT, significant TEC disturbances first appeared in the low-latitude region of Guangxi (approximately 21.3–22.7° N, 108.1–109.8° E), as indicated by the dashed red circle. At 11:35 UT, the area of TEC disturbance expanded, mainly concentrated in the region of 21.2–23.6° N and 107.8–110.2° E. By 11:45 UT, the TEC disturbance continued to extend northward and was distributed within the range of 21.0–24.5° N and 107.1–111.0° E. By 11:55 UT, the TEC disturbance covered the entire Guangxi region. At this time, the north–south component Bz of the geomagnetic field rapidly turned northward after reaching its minimum value, accompanied by a significant TEC disturbance, with the maximum disturbance amount reaching ±0.2 TECU. These results indicate that during the geomagnetic storm, the TEC disturbances in the Guangxi region showed a trend of propagating from low to high latitudes, which is consistent with the propagation patterns demonstrated by the dTEC sequences in Figure 4 and Figure 5.

4.4. Three-Dimensional Evolution Analysis of Ionospheric Disturbances

To reflect the three-dimensional evolution of the ionospheric electron density during geomagnetic storms in the Guangxi region, this study used the electron density of a quiet day (30 November 2023) as the background value. Using a tomographic ionospheric model based on the functional basis and projection functions, three-dimensional variations in electron density on the day of the storm compared to a quiet day were obtained, as shown in Figure 9.

Figure 9 shows that the TEC disturbances during the geomagnetic storm were primarily concentrated within the altitude range of 450–580 km. From 10:00 to 11:00 UT, there were no significant disturbances in the ionosphere; however, at 12:00 UT, notable TEC disturbances appeared at the 580 km altitude, with the maximum TEC value observed at 112° E, 26° N, increasing by 20.54 TECU, while TEC changes at other altitude layers were relatively stable. At 13:00 UT, the maximum TEC value at the 580 km altitude layer shifted to 110.9° E, 25° N, increasing by 14.00 TECU. Simultaneously, the maximum TEC values at the 450 km, 325 km, and 200 km altitude layers increased by 11.40 TECU, 10.44 TECU, and 7.12 TECU, respectively. This indicates that after reaching its peak at 12:00 UT, the TEC at the 580 km altitude layer began to extend to lower layers.

By 14:00 UT, the TEC at all altitude layers had decreased, with the maximum TEC disturbances at the 580 km and 450 km altitude layers weakening to 13.00 TECU and 12.32 TECU, respectively, while the TEC values at the 325 km and 200 km altitude layers were 3.77 TECU and 1.75 TECU, respectively. This suggests that the TEC disturbances gradually returned to a quiet state in the lower-altitudes first.

At 15:00 UT, the TEC disturbances in the 580 km and 450 km altitude layers continued to weaken, with maximum TEC values of 12.35 TECU and 8.30 TECU, respectively. From 16:00 to 17:00 UT, TEC disturbances returned to a quiet state, with no significant changes in the TEC observed across all altitude layers.

5. Discussion and Conclusions

5.1. Discussion

During geomagnetic storms, the interplanetary magnetic field turns southward, causing magnetic reconnection between the Earth’s magnetic field and the interplanetary magnetic field. High-energy particles from the solar wind can directly enter the Earth’s atmosphere, generating Joule heating and auroral effects in the auroral oval region. The lower atmosphere at an altitude of 100 km is heated and expands upwards and equatorward, thereby generating gravity waves. During their propagation, these gravity waves collide with the ions in the ionosphere, causing ion-neutral collisions, which in turn generate TIDs. As a result, large-scale TIDs during geomagnetic storms typically propagate from high latitudes to mid-low latitudes [27]. However, in this particular event, the direction of TID propagation was from low latitudes to mid-latitudes, which is a significant deviation from the traditional TID observation pattern. This geomagnetic storm occurred during the winter in the Northern Hemisphere, while the Southern Hemisphere was in summer. During the day, the thermospheric circulation flows from the Southern Hemisphere to the Northern Hemisphere. Therefore, gravity waves from high-latitude regions of the Southern Hemisphere could propagate into the Northern Hemisphere under the influence of the background wind field, causing the TIDs to propagate from low latitudes to mid-latitudes [28].

Furthermore, the three-dimensional tomography results indicate that ionospheric disturbances first appear in the altitude range of 450 to 580 km and then extend from the higher layers to lower layers. This phenomenon may be related to the prompt penetration of electric fields during geomagnetic storms and the propagation direction of TIDs [28]. The Guangxi region is located in the equatorial ionization anomaly (EIA) region, where the peak altitude is relatively high. During geomagnetic storms, high-latitude electric fields extend into mid- and low-latitude regions, further uplifting the plasma concentration in this area significantly higher than during quiet periods. Meanwhile, the plasma lifted by the fountain effect moves along magnetic field lines toward higher latitudes and gradually descends. Therefore, TIDs propagating from the Southern Hemisphere to the Northern Hemisphere first appear at higher altitudes and then gradually extend to lower altitudes.

5.2. Conclusions

This study used observational data from 193 GNSS stations in China to investigate the characteristics and three-dimensional morphological evolution of ionospheric disturbances during the geomagnetic storm of 1 December 2023. This study used GNSS TEC sequences and an ionospheric tomography algorithm built on a functional basis and projection functions. The findings were as follows:

(1)

During geomagnetic storms, TEC disturbances initially appeared in the low-latitude regions of China and then propagated towards the high-latitude regions. The propagation lasted from 11:30 to 13:30 UT, with an average speed of approximately 217 m/s and a maximum disturbance amplitude of ±0.2 m. The disturbances in the region between 25° N and 32° N lasted the longest, approximately 4 h, whereas in the high-latitude regions (32° N to 40° N), they lasted about 30 min. The propagation pattern was verified using NmF2 data from ionosondes in Beijing (CPT), Wuhan (ZLT), and Hainan (FKT).

(2)

In the three-dimensional ionospheric tomography study, ionospheric disturbances in the Guangxi region were primarily concentrated at altitudes of 450–580 km. At UT 12:00, the maximum TEC value was observed at 26° N, 112° E, at a 580 km altitude, reaching 20.54 TECU. The TEC disturbances at various altitude layers expanded from higher to lower layers between 11:00 and 13:00 UT, and then gradually recovered from lower to higher layers between 14:00 and 16:00 UT.

In summary, this study analyzed the characteristics and propagation patterns of ionospheric disturbances during geomagnetic storms in the Chinese region and reconstructed three-dimensional morphological changes of ionospheric disturbances in the Guangxi region. As a low-latitude region, the ionospheric disturbance response in Guangxi is more pronounced compared to higher-latitude areas. Additionally, the dense distribution of GNSS stations in this region provides ideal high spatial resolution data support for the study of low-latitude ionospheric disturbances. The results reveal the interaction mechanisms between the ionosphere and Earth’s magnetic field, which is of significant importance for the prediction, monitoring, and mitigation of the impacts of geomagnetic storms on the space weather system.

Limitations and potential improvements of this study:

This study has made significant progress in analyzing the propagation characteristics and three-dimensional morphological changes of ionospheric disturbances, but there are still certain limitations. Firstly, although data from 193 GNSS stations were used, they may not evenly cover all regions, and some areas may have observation blind spots that affect the detection of ionospheric disturbances. Future studies could enhance spatial resolution by increasing the number of GNSS stations and incorporating other types of ionospheric data, such as COSMIC radio occultation observation data. Secondly, this study lacks a comparative analysis with other major geomagnetic storm events. Cross-event comparison studies would help verify the generalizability and reliability of the findings. Therefore, future research could expand to include multiple geomagnetic storms for comprehensive analysis. Additionally, although the Guangxi region is representative of low-latitude areas, its results may not fully represent the ionospheric characteristics of low-latitude regions worldwide. Therefore, future studies should consider conducting comparative analyses in other low-latitude regions to enhance the broader applicability of the findings.