The European Heliospheric Forecasting Information Asset (EUHFORIA) is a three-dimensional magnetohydrodynamic (MHD) model of the inner heliosphere that simulates the propagation of solar wind structures from the Sun to 2 AU . The model domain extends from a spherical inner boundary at 0.1 AU to an outer boundary at 2.0 AU, with spatial resolution increasing with heliocentric distance. EUHFORIA solves the ideal MHD equations for a fully ionised plasma under the influence of solar gravity and the interplanetary magnetic field.

Coronal Mass Ejections are injected at the inner boundary as time-dependent, flux-rope structures characterised by their launch time, location (longitude/latitude), angular width, and initial speed. The model self-consistently propagates these transients through the ambient solar wind, capturing CME–CME and CME–solar wind interactions, shock formation, and magnetic field evolution en route to Earth.

At the L1 Lagrange point (approximately 1.5 million km sunward of Earth), EUHFORIA provides time series of key solar wind parameters: proton number density np (cm⁻³), bulk velocity V (km s⁻¹), thermal pressure Pth (nPa), and the three components of the interplanetary magnetic field B (nT) in GSE coordinates. These parameters serve as the direct upstream boundary conditions for global magnetospheric models such as Gorgon-Space.

EUHFORIA's ambient solar wind predictions have been extensively validated against OMNI observations in previous studies , and the present study focuses on the subsequent magnetospheric-ionospheric response using these EUHFORIA outputs as input to Gorgon-Space.

Gorgon is a 3D magnetohydrodynamic code, originally developed to study high-energy, collisional plasma interactions, such as Z-pinches , laser-plasma interactions , and magnetic tower jets . It has recently been adapted to simulate planetary magnetospheres and their interaction with the solar wind . The code solves the resistive MHD equations using a semiconservative formulation. In this approach, the equations for mass, momentum, and energy are solved in a fully conservative form, which ensures accurate capture of shocks and discontinuities. The magnetic field, however, is evolved using the induction equation (a non-conservative form), typically coupled with a divergence-cleaning technique to maintain the solenoidal constraint (∇⋅B=0). Most of its primary features are designed for intricate collisional and resistive Magnetohydrodynamics (MHD) phenomena .

Gorgon-Space has been created specifically to tackle the challenge of conducting global magnetospheric simulations for space weather and space physics. Therefore, it currently represents a distinct area of enhancement and has been optimised and evaluated for particular use cases. A key feature of Gorgon-Space is that it evolves the magnetic vector potential (A) instead of the magnetic field (B) directly. This approach, also used in other models (e.g., ), inherently guarantees that the magnetic field remains divergence-free (i.e., ∇⋅B=0) by construction, which is a significant numerical advantage .

Gorgon-Space utilises a definite-sized 3D Cartesian grid to solve the MHD equations. Moreover, Gorgon possesses a highly advanced capability to scale linearly and in parallel across numerous cores, using fully explicit numerical methods. It found the impact of the solar wind on several important parameters in the tail, including the magnetic field, field-aligned current, plasma velocity, and plasma pressure. These global quantities depend on the magnetosphere's large-scale equilibrium, which is believed to be accurately described by magnetohydrodynamics (MHD).

Gorgon uses a fully explicit Eulerian formulation of the resistive semiconservative MHD equations for a fully ionised plasma as 1∂ρ∂t+∇⋅(ρV)=0,2ρ∂V∂t+(V⋅∇)V=1μ∇×B×B-∇(Pe+Pp),3∂B∂t=∇×(V×B)+η∇2B,4∂ϵp∂t+∇⋅(ϵpV)=-Pp∇⋅(V)-Δpe,5∂ϵe∂t+∇⋅(ϵeV)=-Pe∇⋅(V)+Δpe-Λ+η|J2|, The plasma density, plasma velocity, magnetic field, proton internal energy density, electron internal energy density and thermal pressure are represented by the variables ρ, V, B ϵp, ϵe, and P respectively. The values μ, η, and γ=5/3 represent the magnetic permeability, the magnetic diffusivity, and the ratio of specific heats at constant pressure and volume, respectively. The permeability of vacuum is considered i.e. μ=4π×10-7 Hm⁻¹. The deBar correction, known as Von Neumann artificial viscosity, is utilised to accurately detect and enhance energy conservation in the presence of shocks . Gorgon makes use of second-order Van Leer advection to solve the advection terms in equations. It also employs a variable time step, which automatically satisfies the relevant Courant conditions.

The MHD formulation in Gorgon has its roots in high-energy-density plasma physics, leading to several features that are atypical of global magnetospheric simulations. A key aspect is that it solves separate internal energy equations for electrons and protons, allowing them to be out of thermodynamic equilibrium. The electron energy equation includes physically important terms.

Ohmic Heating (η|J2|), where the resistivity η is treated as a constant value or a Spitzer-like model dependent on local electron temperature . Electron-Proton Energy Exchange (Δpe), which is modelled as a collisional relaxation term of the form Δpe=νepnk(Tp-Te), driving the two populations towards thermal equilibrium . Given the tenuous plasma conditions in the magnetosphere, the optically thin radiation loss term (Λ) is neglected in the present study (Λ=0) because radiative cooling is negligible relative to other energy transport mechanisms.

Furthermore, the code evolves internal energy rather than total energy. While this means total energy is not strictly conserved, it is a deliberate design choice that significantly enhances numerical robustness. This approach effectively mitigates the occurrence of non-physical negative pressures in regions of high magnetic energy or strong shocks, which is a common numerical challenge in conservative total-energy schemes.

The Gorgon-Space code solves separate internal energy equations for electrons and protons. Therefore, the total thermal pressure provided by EUHFORIA must be partitioned at the boundary. For this study, we assume the proton and electron populations are in thermal equilibrium (Tp=Te) upon entry, resulting in an equal split of the pressure (Pp=Pe=Ptot/2). The upstream (sunward) boundary of the simulation is driven by time-dependent solar wind parameters obtained from the EUHFORIA heliospheric model.

The treatment of the inner boundary of the magnetospheric domain in Gorgon is similar to that used in other codes, which are coupled to a thin-shell ionosphere model. The approach is to define the plasma density, temperature, and velocity throughout the spherical region enclosed by the inner boundary and enforce them at every time step. The cells for which this is performed are flagged as inner boundary cells if the cell centre lies within the specified inner boundary radius RIB, typically set to 4 RE (Earth radius). Whilst the MHD equations are not solved in this region, the cells just outside of the boundary are updated based on the inner boundary properties. Thus, the cold, dense plasma expands to form a plasmasphere, and the inner boundary flow feeds into the global magnetosphere.

The grid namelist specifies the domain size and resolution, as well as the dimensional splitting for parallelisation. The default box has dimensions 90×80×80 RE and spans (X,Y,Z)=([-24,66],[-40,40],[-40,40]) RE, and should be kept as close as possible to these dimensions to balance runtime and stability. The simulation utilizes a fixed computational grid with a resolution of 0.5 RE to resolve the global structure of the magnetosphere. This resolution was chosen to accurately capture large-scale boundaries and dynamics, such as the positions of the bow shock and magnetopause, and the overall evolution of the magnetotail, while remaining computationally tractable for the extended simulation runs required for this study. The grid dimensions are (nx,ny,nz)=(180,160,160) cells, which defines the 0.5 RE grid spacing throughout the domain.

The simulation utilises a fixed computational grid within a Cartesian domain spanning from (X,Y,Z)=([-24,66],[-40,40],[-40,40]) RE. This domain size encompasses the critical regions from the dayside bow shock to a significant portion of the magnetotail. The grid consists of (nx,ny,nz)=(180,160,160) cells, resulting in a uniform resolution of 0.5 RE. This resolution was chosen to accurately capture large-scale boundaries and dynamics, such as the positions of the bow shock and magnetopause and the general evolution of the magnetotail, while maintaining computational tractability for the parameter study.

The magnetosphere is influenced by interactions with the solar wind, which can be either viscous or pressure-driven, as described by . The physical mechanisms that allow plasma and energy to flow from the solar wind into the magnetosphere are reconnection and viscous interaction . Dayside magnetic reconnection opens Earth’s closed field lines, enabling direct entry of solar wind plasma and momentum into the magnetosphere via magnetic flux transport. The solar wind conditions driving this simulation are the direct output from the EUHFORIA heliospheric model at the Earth's location (L1 point). This provides a physically consistent, time-dependent input boundary condition for the Gorgon-Space model, linking the CME's interplanetary propagation to its geoeffective impact.

The pronounced tailward stretching and distortion of the magnetotail observed in the XY plane during the CME1 impact (Fig. c–d) are consistent with the onset of magnetic reconnection in the magnetotail. Under enhanced solar wind dynamic pressure and southward IMF conditions (often associated with strong CMEs), reconnection in the near-Earth plasma sheet can lead to the formation of X-lines and the ejection of plasmoids, resulting in the observed elongation and structural deformation of the tail (e.g., ). This process efficiently converts magnetic energy into kinetic and thermal energy, contributing to the global restructuring of the magnetosphere during geomagnetic storms. While reconnection also occurs at the dayside magnetopause, the tail configuration shown here is particularly indicative of nightside reconnection dynamics triggered by the intense solar wind forcing.

Figure  illustrates the evolution of thermal pressure and magnetic field topology during the arrival of CME1, as simulated by our global MHD model. The black lines represent magnetic field lines traced from seed points on a spherical surface at r=1RE, covering latitudes from -80 to +80° and longitudes spaced 15° apart. Field lines are numerically integrated along the magnetic field vector until they either close within the magnetosphere (connecting back to the opposite hemisphere) or extend beyond the magnetopause into the solar wind (defined as open if they reach x>10RE in the tail). The use of both x–z (noon–midnight meridian) and x–y (equatorial) planes allows us to reconstruct the magnetosphere's three-dimensional response to CME1, revealing dayside compression, tail stretching, and topological signatures of magnetic reconnection.

Panel (a) depicts the magnetospheric state approximately 2 d before CME1 arrival, corresponding to the time of the CME's initiation at the Sun (21 June 2015, 08:04 UT). While this snapshot provides context for the initial solar event, it does not represent the magnetospheric configuration immediately preceding the CME's impact at Earth. Panel (b), captured at 00:03 UT on 23 June 2015, depicts the magnetosphere at the time of CME1 arrival, showing significant dayside compression and elevated thermal pressure throughout the magnetosphere, particularly in the equatorial plane (x–y). This pressure enhancement coincides with a dramatic reconfiguration of the magnetic topology, most notably, the pronounced tailward stretching and distortion of the magnetotail observed in the x–y plane (panels c–d).

In Fig. , reconnection sites are inferred from topological changes in field line connectivity and localised enhancements in thermal pressure gradients. During this event, the solar wind reconnection electric field Ey=-VxBz reached +18 mV m⁻¹ (see Fig. ), well above the nominal 1–2 mV m⁻¹ threshold for strong magnetospheric driving. This large positive Ey implies a substantial southward (Bz<0) IMF component. Dayside reconnection is therefore inferred near the subsolar magnetopause, around (x≈10–12 RE, z≈0) in the x–z plane, where closed field lines abruptly transition to open ones. Nightside reconnection is inferred in the near-Earth plasma sheet of the magnetotail, approximately at (x≈-20 to -25RE, y≈0) in the x–y plane, where the stretched tail field lines exhibit a sharp reversal in direction and enhanced pressure, consistent with the formation of a neutral X-line and plasmoid ejection.

This tail deformation is consistent with the onset of magnetic reconnection in the near-Earth plasma sheet. Under the enhanced solar wind dynamic pressure and southward IMF conditions indicated by the large positive Ey (see Fig. ), reconnection efficiently converts magnetic energy into kinetic and thermal energy, leading to the formation of X-lines and the ejection of plasmoids, resulting in the observed elongation and structural disruption of the tail . This process contributes to the global restructuring of the magnetosphere during geomagnetic storms. While reconnection also occurs at the dayside magnetopause, the topological changes in the tail, including increased open flux and the development of a stretched, asymmetric configuration, are particularly indicative of nightside reconnection dynamics triggered by the intense solar wind forcing.

This enhanced energy input for CME1 directly explains the observed magnetospheric compression, tail stretching, and nightside plasmoid formation (Fig. ), as well as the intense field-aligned currents exceeding 23 MA (Fig. ) and the elevated cross-polar cap potential of ∼160 kV (Fig. ). For CME2, the minimal reconnection activity results in negligible energy transfer, consistent with the absence of significant magnetospheric distortion (Fig. ), weak FACs (Fig. ), and low ionospheric convection potential (Fig. ).

Accurate prediction of CME arrival times must account for magnetospheric shielding and internal reconfiguration effects; neglecting them can lead to overestimation of direct solar-wind coupling and mischaracterisation of the geomagnetic response. The dynamic pressure and magnetic topology during the CME2 impact are shown in Fig. . A comparison with the pre-CME state in Fig.  reveals the extreme compression caused by CME1. The magnetopause and bow shock are pushed inward to approximately X/RE∼10 for CME1 (i.e., 10 Earth radii from the centre), whereas they are located around X/RE∼20 for the weaker CME2. This dramatic difference in standoff distance reflects the much stronger dynamic pressure of the solar wind during CME1. The magnetosphere continues to deflect most of the solar wind, thereby shielding the inner region, albeit within a significantly reduced and deformed volume.

In other words, the MHD simulations reveal a stark contrast in magnetospheric response between the strong CME1 and the weak CME2. The arrival of CME1 (Fig. ) induces severe compression of the dayside magnetopause and a significant enhancement of solar wind pressure throughout the magnetosphere, resulting in a highly distorted, elongated tail structure. In contrast, the weaker CME2 (Fig. ) produces only a moderate compression of the magnetopause and a much less pronounced increase in pressure, resulting in a magnetospheric configuration that remains closer to its nominal state, with a less perturbed tail. This demonstrates a direct correlation between CME intensity and the degree of global magnetospheric deformation and pressure loading.

Figure  depicts the changes in velocity caused by the CME1 at the magnetosphere. The first panel of Fig.  shows that the speed at the time of CME1 happening on the Sun is not noteworthy. Additionally, the magnetospheric magnetic field has a consistent shape and can fulfil its primary role as a protective barrier. Panels (b) to (d) depict the wind speed and magnetic configuration of the magnetosphere upon the arrival of CME1 at Earth. Upon initial observation, these three panels represent a notable degree of speed and an irregular magnetic structure resulting from the interaction between the solar wind and the magnetosphere. One can infer that energy from the solar wind is transferred to the Earth's magnetosphere via magnetic reconnection. This occurs when the interplanetary magnetic field connects to the Earth's magnetic field at the magnetopause on the planet's dayside.

Figure  depicts the flow velocity and magnetic field configuration upon CME2's arrival at Earth. CME2 does not lead to high plasma velocity in the magnetosphere. Besides, in Fig. , in the x–z plane, the closed-field regions are larger on the nightside, which could be due to a reduction in open-flux content in the magnetosphere. This might lead to decreased geomagnetic activity and smaller polar caps in the ionosphere.

The rate of magnetic reconnection and the amount of energy transferred are influenced by the solar wind speed, magnetic field strength, and orientation. Subsequently, it was discovered that CME1 triggers a significant tempest on Earth and is accompanied by changes in the magnetic arrangement, increased speed and pressure of the solar wind enveloping the Earth, as previously anticipated in Fig. .

Figure  showcases the changes in pressure and magnetic field in the vicinity of the Earth upon the arrival of CME2. As shown in Fig. , there was no indication of a significant magnetic storm when CME2 arrived. It can shed light on the impact of the background solar wind, which is the same in both CMEs, and thus it is not the primary factor in causing magnetic storms. Hence, it can be deduced that the nature of the CME and its interactions with other CMEs are crucial factors in determining the geomagnetic impact of CMEs on Earth.

This enhanced energy input for CME1 directly explains the observed magnetospheric compression, tail stretching, and nightside plasmoid formation (Fig. ), as well as the intense field-aligned currents exceeding 23 MA (Fig. ) and the elevated cross-polar cap potential of ∼160 kV (Fig. ). For CME2, the minimal reconnection activity results in negligible energy transfer, consistent with the absence of significant magnetospheric distortion (Fig. ), weak FACs (Fig. ), and low ionospheric convection potential (Fig. ).

To model magnetosphere–ionosphere (M–I) coupling, Gorgon-Space employs a separate, height-integrated ionospheric module. The 3D MHD domain does not resolve the ionosphere; instead, its inner boundary is placed at RIB=4RE, where field-aligned currents (FACs) are computed from the magnetospheric solution via Ampère's law, J∥=(J⋅b)b, with b=B/|B|. These FACs are mapped along dipole magnetic field lines onto a 2D spherical ionospheric grid at an effective altitude of ∼110 km – corresponding to the E region (90–150 km), where Hall and Pedersen conductivities peak.

In the Gorgon-Space model, the ionospheric electrical conductivity is represented by height-integrated Pedersen (ΣP) and Hall (ΣH) conductances. These are calculated using a combination of solar EUV-driven background and auroral precipitation-driven enhancements: 6ΣP=ΣP,EUV+ΣP,aurora7ΣH=ΣH,EUV+ΣH,aurora

EUV-driven background. The solar EUV contribution is parameterized as a function of solar zenith angle (χ) and the F10.7 solar flux index (in sfu, 1sfu=10-22Wm-2Hz-1): 8ΣP,EUV=5.0×F10.71000.5×cos⁡(χ)0.59ΣH,EUV=0.8×ΣP,EUV These empirical relations follow the standard formulation used in many global MHD models (e.g., ), where the cos⁡0.5(χ) dependence approximates the Chapman production function for ionospheric ionization.

Auroral component. The auroral contribution is derived from the mapped magnetospheric electron energy flux (ε, in mWm-2) and characteristic energy (E0, in keV) using the empirical relations of : 10ΣP,aurora=40εE016+E0211ΣH,aurora=40εE00.8516+E02 These relations are derived from fits to auroral observations and have been widely validated for storm-time conditions . The factor 40 has units of mW-1m2keV-1 to produce conductances in siemens.

During the simulation, the electron energy flux and characteristic energy at the inner boundary (R=4RE) are mapped along dipole field lines to the ionospheric altitude (∼110 km). The conductances are updated at each time step (Δt≈10 s) to reflect changes in magnetospheric precipitation.

The total ionospheric conductance (ΣP, ΣH) is then used in Ohm's law for the height-integrated ionospheric current: 12J⟂=ΣPE+ΣH(B×E) where E is the ionospheric electric field, b is the unit vector of the magnetic field, and J⟂ is the height-integrated perpendicular current density. This dynamic conductance model is critical for accurately reproducing the observed FAC intensities and cross-polar cap potentials during the storm main phase. The Gorgon-Space code has been previously validated for ionospheric coupling and FAC simulations under storm-time conditions . In a recent study using the same EUHFORIA-Gorgon-Space modelling chain to investigate the May 2024 solar storm, further demonstrated that the dynamic conductance model successfully reproduced observed FAC patterns and cross-polar-cap potentials measured by SuperDARN, providing additional validation of this implementation.

The electrostatic potential ϕ is obtained by solving the 2D current continuity equation: 13∇⟂⋅J⟂=-∇⟂⋅(ΣP∇⟂ϕ)=J∥ ensuring current closure. The resulting potential distribution is mapped back along the dipole field lines to the MHD inner boundary, thereby providing a self-consistent time-dependent boundary condition for the plasma flow.

Within the thin-shell approximation, the height-integrated ionospheric current is governed by Ohm's law: J⟂=ΣPE+ΣH(b×E), where ΣP and ΣH are the Pedersen and Hall conductances. In this study, these conductances are calculated using empirical relations: a solar EUV-driven background (with square-root dependence on solar flux) and an auroral component derived from magnetospheric electron energy flux mapped along field lines. The electrostatic potential ϕ is obtained by solving the 2D current continuity equation ∇⟂⋅J⟂=-∇⟂⋅(ΣP∇⟂ϕ)=J∥, ensuring current closure. The resulting potential distribution is mapped back along the dipole field lines to the MHD inner boundary, thereby providing a self-consistent time-dependent boundary condition for the plasma flow.

This coupling enables energy and momentum transfer across the system: FACs generated in the magnetosphere (where ∇×B≠0) flow into the ionosphere, where they are closed by perpendicular Pedersen and Hall currents. The resulting Lorentz force, J⟂×B, drives ionospheric plasma convection and, through ion–neutral collisions, neutral winds. This convection pattern, anchored by FAC stress, transports magnetic flux tubes globally, thereby mediating the large-scale M–I energy exchange that underpins space-weather impacts.

The accumulation of current density aligned with the magnetic field can be attributed to the interaction between the solar wind and the magnetosphere. The predominant energy transfer and current alignment in the ionosphere were driven primarily by the boundary conditions. To accurately predict space weather, it is necessary to have a solid understanding of the physics underlying these coupling processes. Determining the accuracy of MHD codes in reproducing plasma sheet electric fields during solar wind-magnetosphere interactions is of utmost significance for Space Weather applications.

Figure  presents the evolution of field-aligned current (FAC) density in the Northern Hemisphere ionosphere in response to the sudden commencement (SC) driven by the arrival of CME1 at Earth, capturing the onset, peak, and initial recovery of the disturbance. Panel (a) shows the quiet-time pre-impact baseline at 08:02 UT on 21 June. Following the SC, panels (b) and (c) depict the development and peak of the intense electrodynamic response. Panel (b), immediately after the SC at 00:03 UT on 23 June, reveals a localized, dayside-enhanced FAC pattern. This initial enhancement is consistent with the characteristic Preliminary Impulse (PI) observed during shock-induced compressions, often attributed to the compression of the dayside magnetosphere and the associated Chapman-Ferraro (CF) currents (e.g., ).

Panel (c) shows the peak intensity, indicating a substantial increase in the total current magnitude to approximately 23 MA and a poleward shift of the primary current regions. This stronger, poleward-expanding structure corresponds to the Main Impulse (MI). The MI signifies the full engagement of the global current system (R1 FACs), which is directly driven by the large-scale magnetospheric convection electric field. This convection electric field (Econv=-v×B) is itself established by enhanced dayside magnetic reconnection under the prolonged southward IMF conditions in the CME sheath.

Panel (d) shows the onset of the post-MI decay phase in the Northern Hemisphere, where the current density magnitude decreases, and the pattern shifts poleward from 0 to 10–20° colatitude. The clear temporal sequence from quiet conditions (a) through the distinct PI and MI phases (b, c) and into post-MI decay (d) demonstrates that the Gorgon-Space code captures the fundamental two-phase ionospheric electrodynamic response to a CME-driven shock, a key benchmark for validating space-weather models. The intense MI-phase FACs are a primary conduit for energy transfer from the magnetosphere, underscoring their critical role in driving space-weather effects such as geomagnetically induced currents .

To validate the magnetosphere-ionosphere coupling in our simulations, we compare the Gorgon-Space-derived field-aligned current (FAC) distributions with observational data from the Active Magnetosphere and Planetary Electrodynamics Response Experiment (AMPERE). Figure  presents a side-by-side comparison of modelled (left) and AMPERE-observed (right) FAC densities in the Northern Hemisphere at four key stages of the CME1 event.

Panel (a) shows the quiet-time pre-impact baseline (21 June, 08:04 UT), where both datasets exhibit weak, patchy currents with total integrated FAC <5 MA. Following the sudden commencement at 00:03 UT (panel b), both the model and AMPERE show a rapid intensification of the Region 1/Region 2 FAC system, with peak densities exceeding ±1 µAm-2 and total integrated currents reaching approximately 23 MA in the simulation.

During the main-phase peak at 00:48 UT (panel c), the classic two-cell convection pattern is well established in both datasets, with strong dawnside downward (blue) and duskside upward (red) currents. By 01:00 UT (panel d), the system begins to decay, though residual currents remain elevated at approximately 15–18 MA in the model.

The close agreement in morphology, timing, and integrated current magnitude validates the dynamic conductance model and the EUHFORIA-derived upstream boundary conditions. Minor discrepancies in fine-scale structures are expected: AMPERE's ∼10–20 min temporal resolution and spatial smoothing differ from the model's grid-scale output, and current wedges are not fully resolved in the global MHD framework. Nevertheless, the large-scale FAC evolution is robustly captured, confirming the model's fidelity for space weather applications.

For completeness, we also compare the Gorgon-Space-simulated FACs with AMPERE observations for the CME2 event (28 June 2015). Figure  presents this comparison at three key times during the CME2 passage.

Panel (a) shows the model prediction at 13:00 UT, shortly after the CME2 arrival. The FACs are weak and disorganized, with a total integrated current of only 3.7 MA and peak densities ranging from -0.72 to +0.59 µAm-2. The corresponding AMPERE observations (panel c) confirm this weak activity, showing no clear Region-1/Region-2 structure.

At 16:40 UT (panels b and d), both the model and AMPERE show a slight intensification, with total FAC reaching 7.8 MA in the model and 5.0 MA in AMPERE. However, the currents remain significantly weaker than those observed during CME1 (in which the total FAC exceeded 23 MA). The pattern remains disorganized, lacking the classic two-cell convection structure.

By 19:10–20:00 UT (panels e and f), the FACs begin to decay further. The model predicts a total FAC of 7.5 MA, while AMPERE shows 5.0 MA. The close agreement between model and observations in capturing the weak, disorganized nature of the FACs during CME2 validates the model's ability to correctly simulate both geoeffective (CME1) and non-geoeffective (CME2) events. This demonstrates that the EUHFORIA-Gorgon-Space chain correctly reproduces the observed contrast between the two CMEs as a function of IMF orientation.

Minor discrepancies in the exact timing of peak currents are expected due to AMPERE's ∼10–20 min temporal resolution and the inherent variability in CME-solar wind interactions. Nevertheless, the overall agreement confirms that the model successfully captures the muted ionospheric response to CME2, consistent with the absence of sustained southward IMF.

Table  summarizes the key differences between the two CME events. CME1 produced a strong geomagnetic storm (G4 level) with intense FACs (∼23 MA), high CPCP (∼160 kV), and severe magnetospheric compression. In contrast, CME2 produced only weak FACs (<8 MA) and minimal geoeffectiveness, despite having a higher peak speed. This contrast confirms that IMF orientation, not CME speed, governs storm intensity.

Figure  shows the electric potential response in the ionosphere. The illustration shows the cross-polar cap electric potential (CPCP) in the ionosphere, which is widely used to assess the worldwide convection intensity. This shows the disparities between the highest and lowest electric potential values in the ionosphere. CPCP was initially derived from the integral rate of dayside-merging and can be used as an estimate of the key parameter of dayside-merging, as in .

The first panel in Fig.  shows the level of the electric potential in the ionosphere at the onset of the initial coronal mass ejection (CME1). Panel (b) of Fig.  displays the ionospheric electric potential reaction during the initial arrival of the first coronal mass ejection (CME1) at the Earth. Comparing the two panels helps determine the impact of the solar wind on Earth, yielding an ionospheric electric potential of approximately 160 kV during the CME1 impact. This elevated cross-polar cap potential (∼160 kV) is a direct measure of the strength of the magnetospheric convection electric field. This same convection electric field, when mapped back along magnetic field lines into the magnetosphere, drives the bulk plasma motion via the E×B drift. Consequently, the weak ionospheric potential during CME2 (Fig. ) is physically consistent with the low plasma velocities observed in the magnetosphere for that event (Figs. , ). Figure  illustrates the electric potential in the ionosphere when CME2 reaches the Earth. As predicted, the arrival of CME2 at Earth does not cause any notable electric potential in the ionosphere. The elevated ionospheric potential (∼160 kV) is a signature of a strong magnetospheric convection electric field. This large-scale field governs the global transport and energisation of plasma. While the quasi-static ionospheric field itself does not directly accelerate particles to high energies, the associated magnetospheric convection can inject plasma-sheet particles into the inner magnetosphere, where they may subsequently be accelerated to ring-current and radiation-belt energies by processes such as radial diffusion and wave-particle interactions.