MHD Wave Theory in the Magnetosphere
A Tutorial for Ground Magnetometer Analysts
This tutorial bridges MHD wave physics to the coherence regimes, correlation lengths, and anisotropy patterns you observe in IMAGE cross-spectral analysis. Each section presents the physics, shows an animated diagram, and connects directly to your data.
1. MHD Basics — Frozen-In Flux
In a highly conducting plasma (like the magnetosphere), magnetic field lines are "frozen in" to the plasma. Where the plasma goes, the field goes. This means disturbances propagate as waves on the magnetic field, not through the field. The magnetic field is the medium.
This frozen-in condition holds whenever the magnetic Reynolds number Rm = μ0σVL ≫ 1, which is satisfied everywhere in the magnetosphere except at reconnection sites (Section 7). It means that plasma elements that start on the same field line stay on that field line as the system evolves.
Three Key Speeds
The dynamics of a magnetized plasma are controlled by three characteristic wave speeds:
Speed
Formula
Physics
Magnetosphere
Alfvén speed VA
B / √(μ0ρ)
Transverse waves along field lines. Magnetic tension is the restoring force.
~1000 km/s at 4 RE, ~100 km/s at 10 RE
Sound speed Cs
√(γp / ρ)
Compressive waves in the plasma. Thermal pressure is the restoring force.
~200 km/s
Fast magnetosonic Vf
√(VA² + Cs²)
Compressive waves in all directions. Both magnetic and thermal pressure.
~1020 km/s at 4 RE
The key insight: VA varies enormously across the magnetosphere because B decreases as r−3 (dipole) while ρ increases with L-shell (more plasma at larger distances). This variation in VA is what creates the rich wave structure you observe.
Alfvén wave: a transverse perturbation propagates along the frozen-in field
Connection to Your Data
The frozen-in condition means that every current system you measure on the ground — electrojet, FACs, ring current — involves bulk plasma motion dragging field lines with it. When you see a ΔB perturbation, there is a corresponding plasma flow and field line displacement somewhere above you. The wave speeds determine how quickly perturbations propagate and therefore set the spatial coherence scales you observe.
2. The Three MHD Wave Modes
Three fundamental wave types exist in a magnetized plasma. Understanding these is essential because each mode creates different ground signatures with different spatial coherence properties:
1. Alfvén Wave (Shear / Transverse)
Field lines bend transversely but do not compress. Like shaking a rope sideways. Propagates only along B (the background magnetic field). Speed = VA. The wave carries a field-aligned current (FAC) that closes through the ionosphere — this is what your magnetometers detect as a magnetic perturbation in the N-S direction (the ΔBX component from the Hall current).
2. Fast Magnetosonic Wave (Compressive)
Field lines AND plasma compress together. Propagates in all directions (isotropic in the limit VA ≫ Cs). Speed = Vf ≈ VA. This is the "sound wave" of MHD. Because it propagates isotropically, it can carry energy across field lines — crucial for coupling different L-shells and creating global signatures.
3. Slow Magnetosonic Wave (Compressive)
A compressive wave guided along B at speed below VA. Less important for ground signatures in the Pc3–Pc5 bands because it is strongly damped in the magnetosphere. Included for completeness.
Three MHD wave modes: transverse bending, isotropic compression, and guided compression
Connection to Your Data
Alfvén waves produce localized signatures (they don't cross field lines) → SHORT correlation lengths. This maps to your Pc5 coherence dip.
Fast magnetosonic waves produce global signatures (they propagate isotropically) → LONG correlation lengths. This maps to your Pi2 coherence recovery and Pc3 high coherence.
The competition between these two modes at different frequencies is the fundamental reason your L(f) is not monotonic.
3. Alfvén Waves — The Plucked String
An Alfvén wave is a transverse oscillation of a magnetic field line, exactly like a vibrating guitar string. The physics is a direct analogy:
Guitar String
Magnetic Field Line
String tension T
Magnetic tension B²/μ0
Linear mass density μ
Plasma mass density ρ
Wave speed √(T/μ)
Alfvén speed B/√(μ0ρ)
Fixed ends (bridge & nut)
Ionospheric boundaries (north & south)
Standing wave harmonics
Field line eigenfrequencies
In the magnetosphere, field lines connect the northern and southern ionospheres. The ionospheres act as partial reflectors: the conductive ionosphere shorts out the wave's electric field, reflecting the Alfvén wave back along the field line. Standing waves form — just like standing waves on a guitar string with fixed ends.
The fundamental frequency of a field line is:
f1 ≈ VA / (2Lpath)
where Lpath is the total path length from ionosphere to ionosphere along the field line. Higher harmonics: fn = n · f1.
Standing Alfvén wave harmonics on a dipole field line: fundamental, 2nd, and 3rd harmonics
Connection to Your Data
The eigenfrequency of the field line directly above IMAGE's dense cluster (L ≈ 6, ~67° MLAT) falls in the Pc5 band (period ~300–600 s). This is not a coincidence — it is the fundamental physical reason why Pc5 pulsations are prominent in your data. The field lines overhead are resonating at their natural frequency, just like plucked strings. Higher harmonics of these same field lines produce oscillations in the Pi2 band.
4. Field Line Resonances (FLR) — Why Pc5 is Localized
Each magnetic field line has its own eigenfrequency, determined by its length and the Alfvén speed along it. Because both B and ρ vary along the field line, the eigenfrequency must be computed as an integral:
fn = n / (2 ∫ ds / VA(s))
But the key scaling is simple: longer field lines (larger L) have lower eigenfrequencies.
L-shell
MLAT (°)
Fundamental Period
ULF Band
L = 4
60°
~60–100 s
Pi2
L = 6
67°
~300–600 s
Pc5
L = 8
71°
~600–1000 s
Pc5/Pc6
L = 10
73°
~1000–1500 s
Pc6
Field Line Resonance (FLR) is what happens when a broadband fast-mode wave (Section 2) encounters a region where the local eigenfrequency matches the wave frequency. Energy transfers from the fast mode (compressive, global) to the Alfvén mode (transverse, localized). The field line resonates.
But here is the critical point: each field line resonates at its own frequency. At a given frequency f, only the L-shell where feigen(L) = f will resonate. Neighboring L-shells have different eigenfrequencies and do not resonate at that frequency. The result is a perturbation that is highly localized in latitude — extending only ΔL ≈ 0.5–1 RE around the resonant shell.
Field Line Resonance: a broadband fast-mode wave selectively excites the field line whose eigenfrequency matches
Connection to Your Data
Your L(f) minimum at Pc5 is a direct consequence of FLR. At Pc5 frequencies (~2–5 mHz), the resonant L-shell lies within or near your IMAGE network. The perturbation is localized to a narrow latitude band → short correlation length.
The anisotropy crossover at ~2.2 mHz maps to the FLR eigenfrequency at L ≈ 6. Below this frequency, perturbations are dominated by the electrojet (E-W extended). Above it, FAC filaments associated with FLR (N-S aligned) become dominant.
The FLR amplitude peaks at the resonant latitude and drops off rapidly on either side. In your variogram, this shows as a nugget + short-range structure in the N-S direction at Pc5 frequencies.
The magnetosphere also acts as a waveguide or cavity for fast-mode (compressive) waves. The boundaries are regions where VA changes sharply, creating partial reflections:
Outer boundary: The magnetopause, where the solar wind's low VA meets the magnetosphere's high VA.
Inner boundary: The plasmapause (L ≈ 4), where a sharp density increase causes VA to drop abruptly. Also, the upper ionosphere serves as a lower boundary.
Fast-mode waves bounce between these boundaries, setting up standing patterns — cavity modes. These are the MHD analog of acoustic resonances in a room. Their periods are determined by the cavity dimensions and the fast-mode speed:
Tcavity ≈ 2ΔR / Vf ≈ 2 × 6 RE / 1000 km/s ≈ 80 s (Pi2 band)
The essential difference from FLR: cavity modes are global. The entire magnetosphere between the boundaries participates in the same oscillation. All ground stations within the cavity see the same phase of compression and rarefaction. This is why your coherence recovers at Pi2 frequencies.
Your Pc5 coherence dip is therefore the transition zone: from global convection (high coherence at low-f) through localized FLR (low coherence at Pc5) back to global cavity modes (recovered coherence at Pi2).
Magnetospheric cavity mode: global compressive oscillation between the plasmapause and magnetopause
Connection to Your Data
Pi2 coherence recovery: Cavity modes are global → all IMAGE stations see the same oscillation → high coherence, long L.
Z component does NOT recover: Cavity modes produce horizontal (X, Y) perturbations, not vertical. The Z component coherence stays low because FAC (which drives ΔBZ) remains localized. This is a strong diagnostic for cavity modes vs. other mechanisms.
Your L ≈ 800 km at Pc5 is the FLR width. At Pi2, L increases again because cavity modes extend across the full network.
6. Magnetosphere Breathing — Solar Wind Pressure Pulses
The most basic MHD response of the magnetosphere to external forcing: when solar wind dynamic pressure increases, the magnetopause is compressed inward; when it decreases, the magnetopause expands outward. This is a fast-mode compression/rarefaction, and it creates the largest-scale perturbations observable on the ground.
Sudden commencements (SC): A sharp increase in solar wind pressure at storm onset drives rapid magnetopause compression. A fast-mode pulse propagates inward through the magnetosphere at Vf ≈ 1000 km/s. Since the IMAGE network spans ~1000 km and Vf is ~1000 km/s, the transit time across the network is ~1 second. This means SC signals arrive at all stations nearly simultaneously → extremely high coherence.
Continuous pressure fluctuations in the solar wind drive continuous fast-mode compressions at the magnetopause. These are the primary driver of the high coherence at frequencies below the FLR band (< 1.7 mHz in your data).
Magnetosphere breathing: the global response to solar wind dynamic pressure variations
Connection to Your Data
Regime 1 (>10 min): Solar wind pressure fluctuations on these timescales drive global compressions that are seen by all stations simultaneously. Vf ≈ 1000 km/s ≫ network size / period, so all stations are in phase → coherence ~0.9, L ≈ 2000–5000 km.
Sudden commencements are the extreme case: a step-function pressure increase propagates through the magnetosphere as a single fast-mode wavefront. Your cross-spectral analysis of SC events should show coherence ≈ 1.0 across ALL frequency bands, briefly.
The transition from high coherence (Regime 1) to the Pc5 dip (Regime 2) occurs when the wave frequency approaches the FLR eigenfrequency. Below FLR, the magnetosphere responds as a whole. At FLR, individual field lines decouple from the global response and resonate independently.
7. Field Line Decoupling During Reconnection
Reconnection is where the frozen-in condition (Section 1) breaks down. At an X-point (or X-line), the resistivity is locally enhanced, allowing field lines to change their topology. This is not merely a theoretical concept — it is a wave physics event that changes the entire character of MHD wave propagation.
Dayside Reconnection
When the IMF (interplanetary magnetic field) is southward, it is antiparallel to the Earth's northward field at the subsolar magnetopause. At the X-line:
A closed field line (both ends on Earth) splits into two open field lines (one end on Earth, one end in the solar wind)
These open lines get swept poleward and tailward by the solar wind flow
Open flux accumulates in the tail lobes
Nightside (Tail) Reconnection
In the magnetotail, the reverse process:
Two open field lines approach the tail X-line
They reconnect → one closed line (snaps earthward as a dipolarization front) + one disconnected IMF line (ejected tailward as a plasmoid)
Wave Physics Consequences
An open field line has no southern reflection point → no standing wave → no FLR eigenfrequency. When reconnection opens or closes field lines, it changes the Alfvén travel time discontinuously. The wave physics of the magnetosphere changes fundamentally. This is why FLR behavior changes dramatically between quiet and disturbed conditions.
Magnetic reconnection changes field line topology, fundamentally altering wave propagation
Connection to Your Data
During strongly southward IMF (active reconnection), the fraction of open flux over your stations increases. Open field lines cannot support FLR standing waves, so Pc5 FLR amplitude may decrease even as overall magnetic activity increases.
The Dungey cycle (dayside opening → tail closure) drives the large-scale convection that produces your Regime 1 high coherence. The two-cell convection pattern is the direct consequence of this flux circulation.
Tail reconnection launches earthward flows that create substorm signatures (Section 8) — another contributor to your ground perturbations, especially in the midnight sector.
8. Plasma Sheet Dynamics — Substorm Field Evolution
Growth Phase
During the substorm growth phase (typically 30–60 minutes):
Open flux accumulates in the tail lobes (from dayside reconnection, Section 7)
The lobes expand, squeezing the plasma sheet thinner
The cross-tail current intensifies to maintain pressure balance: J × B = ∇p
The magnetic field becomes increasingly stretched — tail-like rather than dipolar
Expansion Onset
When the tail becomes too thin and stretched, instability triggers:
A Near-Earth Neutral Line (NENL) forms at ~20 RE
Reconnection begins, and a dipolarization front surges earthward at ~200–400 km/s
Stretched field lines snap back to dipolar configuration (think of a rubber band released)
The cross-tail current is disrupted, and current is redirected as FACs into the ionosphere — the substorm current wedge (SCW)
The SCW produces the large ΔBX perturbations your magnetometers record during substorms
Pi2 Onset Signature
The formation of the SCW and the dipolarization front launch waves in the Pi2 band (40–150 s). These are the Pi2 onset pulsations used universally to time substorm onset. They are a mix of cavity modes (Section 5) and transient Alfvén oscillations of the newly dipolarized field lines.
Substorm cycle: growth phase stretching, expansion onset dipolarization, and recovery
Connection to Your Data
Pi2 pulsations at substorm onset are one of the clearest ULF signals in your data. They appear as a burst of ~40–150 s oscillations, strongest in ΔBX, beginning precisely at expansion onset. Their spatial coherence is high because the SCW is a large-scale current system.
Substorms primarily affect the midnight sector. If your IMAGE analysis includes stations in the midnight MLT sector, substorm signatures will dominate the perturbation statistics during active intervals.
The cross-tail current disruption changes the current geometry from a sheet (high spatial coherence, E-W extended) to a wedge with FACs (localized, N-S aligned). This topological transition in the current system may contribute to the anisotropy reversal you observe.
9. The 12-Hour View — How a Ground Station Sees the Magnetosphere
A ground magnetometer at fixed geographic coordinates rotates through magnetic local time (MLT) over 24 hours. At different MLT, it samples completely different parts of the magnetospheric current system:
MLT Sector
Overhead Currents
Dominant Signal
Character
00 (Midnight)
Tail current, SCW
Strong ΔBX perturbations
Substorm-dominated, bursty
06 (Dawn)
R2 FACs, transition
Mixed signatures
Transition zone
12 (Noon)
Magnetopause currents
Compression, Pc3
Continuous, driven
18 (Dusk)
R1 FACs, Harang
ΔBY perturbations
Shear-dominated
The station sees a completely different magnetosphere every 6 hours. The magnetic perturbation switches character: from compression-dominated (noon) to current-sheet-dominated (midnight) to shear-dominated (dusk). This is why diurnal variation in spatial statistics is not merely a nuisance — it reflects genuine physical changes in which current system is overhead.
IMAGE station rotating through MLT sectors, sampling different current systems over 24 hours
Connection to Your Data
Diurnal variation in variogram parameters is physically real, not an artifact. The spatial statistics of the magnetic field genuinely change as your network rotates through different current systems.
Noon sector: Dominated by magnetopause compression and Pc3 upstream waves. Expect high coherence (global source), E-W anisotropy (compression pattern).
Midnight sector: Dominated by substorm SCW and tail dynamics. Expect lower coherence (localized FACs), N-S anisotropy (FAC alignment), and bursty temporal behavior.
Implications for kriging: The optimal kriging parameters (L, anisotropy ratio, nugget) should ideally vary with MLT, because the underlying physics changes. A single variogram fit over 24 hours averages over fundamentally different physical regimes.
10. Your Four Regimes Explained — The Complete Physical Picture
Now we tie everything together. Your cross-spectral analysis found four coherence regimes. Here is the MHD wave physics behind each one:
Regime 1: >10 min (<1.7 mHz) — HIGH Coherence (0.85–0.95)
These are convection-driven perturbations from the Dungey cycle (Section 7). The electrojet is a single large-scale current sheet spanning the entire auroral zone. The current flows E-W for thousands of kilometers. All stations within the electrojet see essentially the same current → high coherence, long L (~2000–5000 km). The spatial pattern is strongly E-W extended, explaining your low-frequency E-W anisotropy.
Regime 2: 3–10 min (1.7–5.6 mHz) — DIP (0.55–0.75)
This is the Pc5 / FLR regime (Section 4). Each field line resonates at its own frequency, determined by its L-shell. Neighboring L-shells resonate at slightly different frequencies → perturbations are localized in latitude (ΔL ~ 0.5–1 RE) → low coherence, short L (~800–1500 km). The L(f) minimum is here. FLR drives FACs that are N-S aligned (localized filaments in the ionosphere), beginning the transition from E-W to N-S anisotropy.
Regime 3: 40s–3 min (5.6–25 mHz) — RECOVERY (0.65–0.76)
Pi2 cavity/waveguide modes (Section 5). Fast-mode waves bouncing between magnetopause and plasmapause create global standing patterns. All stations see the same cavity mode → coherence recovers, L increases. The Z component does NOT recover — these are horizontal (X, Y) oscillations. The cavity modes are the MHD "drumhead" of the magnetosphere.
Regime 4: <40s (>25 mHz) — HIGH (0.72–0.84)
Pc3 upstream waves. Ion cyclotron waves generated in the foreshock region, where reflected solar wind ions gyrate and interact with the incoming solar wind. The wave frequency is set by the ion cyclotron frequency in the solar wind: f ≈ 0.015 × BIMF(nT) Hz. These waves propagate through the magnetosheath and magnetopause into the dayside magnetosphere. They have a single, spatially extended source region → high spatial coherence across the ground network. The Pc3 signal is strongest at noon MLT (closest to the foreshock).
The four coherence regimes and their MHD wave physics origins
The crossover at ~2.2 mHz: below it, the E-W extended electrojet dominates; above it, N-S aligned FAC filaments dominate
Connection to Your Data — The Complete Map
Coherence vs. frequency: The non-monotonic coherence curve (high → dip → recovery → high) is NOT anomalous. It is the expected signature of four distinct MHD wave modes, each with different spatial coherence properties, dominating in different frequency bands.
L(f) minimum at Pc5: FLR creates the most spatially localized perturbations in the magnetosphere. Your kriging should use L ≈ 800 km (your finding) at Pc5, increasing to L ≈ 2000 km at lower and higher frequencies.
Anisotropy crossover at ~2.2 mHz: This is the FLR eigenfrequency at the L-shell directly overhead (L ≈ 6). Below it: the electrojet (E-W current sheet) dominates → E-W anisotropy. Above it: FLR-driven FAC filaments (N-S aligned) dominate → N-S anisotropy.
Z component coherence: Stays low at Pi2 because cavity modes produce horizontal (X, Y) perturbations. The Z component is driven by localized FACs, which remain spatially structured even when the X, Y components are globally coherent.
Frequency-dependent kriging: Your finding that optimal L varies with frequency is physically correct and deeply meaningful. Different frequencies sample different wave modes, which have fundamentally different spatial scales. A frequency-dependent variogram model is not merely empirically better — it is physically necessary.