Palaeomagnetism and Polar Wandering
CSIR NET Part C Geophysics: rock magnetism, NRM, demagnetisation, Fisher statistics, field tests, VGP, palaeointensity, APW and true polar wander.
- 20 min
On this page · 15 sections
Prerequisites
- The Geomagnetic Field and Its Variations: the dipole field, \( \tan I = 2\tan\lambda \) and secular variation.
- The Part B picture of remanence, apparent polar wander and sea-floor stripes in Paleomagnetism, Sea-Floor Spreading and Plate Tectonics.
- Spherical trigonometry (the cosine rule on a sphere) and basic statistics.
The Hook: Two Continents, Two Wandering Poles
Rock Magnetism
Kinds of magnetic behaviour
| Behaviour | Spins | Susceptibility \( \chi \) (SI) | Examples |
|---|---|---|---|
| Diamagnetic | no permanent moments | small, negative (about \( -10^{-5} \)) | quartz, calcite, water |
| Paramagnetic | moments randomly oriented without a field | small, positive | Fe-bearing silicates: olivine, pyroxene, biotite |
| Ferromagnetic (strict sense) | parallel, coupled | large | native iron (rare in rocks) |
| Ferrimagnetic | antiparallel sublattices of unequal moment | large | magnetite, titanomagnetite, maghemite, monoclinic pyrrhotite, greigite |
| Antiferromagnetic (canted or defect) | antiparallel sublattices, weak net moment from spin canting (hematite) or defects (goethite) | weak but stable remanence | hematite; goethite |
Only the last three (ferromagnetic in the broad sense, sensu lato) carry remanence.
| Mineral | Composition | Curie / Néel temperature (approx.) |
|---|---|---|
| Magnetite | Fe\(_3\)O\(_4\) | 580 °C |
| Hematite | α-Fe\(_2\)O\(_3\) | 675 °C |
| Titanomagnetite | Fe\(_{3-x}\)Ti\(_x\)O\(_4\) | decreases as Ti increases (well below 580 °C for Ti-rich basaltic compositions) |
| Pyrrhotite (monoclinic) | Fe\(_7\)S\(_8\) | about 320 °C |
| Goethite | α-FeOOH | about 120 °C |
Heated above its Curie temperature, a mineral becomes paramagnetic and loses all remanence; on cooling it again becomes ordered.
Grain size and domain state
| State | Size (magnetite, order of magnitude) | Behaviour |
|---|---|---|
| Superparamagnetic (SP) | below a few tens of nm | moments flip by thermal energy; no stable remanence at room temperature |
| Single domain (SD) | tens of nm to about 0.1 μm | uniformly magnetised; most stable carriers of ancient remanence |
| Pseudo-single domain (PSD) | about 0.1 μm to roughly 10–20 μm (sources differ) | intermediate |
| Multidomain (MD) | larger | domain walls move easily; low stability |
Néel relaxation and blocking
For an assembly of SD grains, the time for remanence to decay is
$$\boxed{\tau = \tau_0\exp\left(\frac{Kv}{k_BT}\right)}$$where \( \tau_0 \approx 10^{-9} \) s is an atomic attempt time, \( K \) the anisotropy energy density, \( v \) the grain volume, \( k_B \) Boltzmann’s constant and \( T \) temperature. The exponential makes the transition very sharp:
- \( Kv/k_BT = 25 \) gives \( \tau \approx 72 \) s;
- doubling the volume (\( Kv/k_BT = 50 \)) gives \( \tau \approx 5\times 10^{12} \) s \( \approx 1.6\times 10^5 \) years.
The blocking temperature \( T_B \) is where \( \tau \) passes the time scale of interest. Above \( T_B \) the grain follows the field; below it, its moment is “frozen”. This single equation explains TRM (blocking on cooling), VRM (slow unblocking at room temperature over long times) and why thermal demagnetisation works.
Remanence types
| Remanence | Acquired | Character |
|---|---|---|
| TRM (thermoremanent) | cooling through blocking temperatures in a field | strong, stable, parallel to and proportional (for weak fields) to the field; lavas and intrusions |
| DRM / pDRM (detrital, post-depositional) | alignment of magnetic grains as sediment settles or while still wet | weaker; may record a slightly shallowed inclination |
| CRM (chemical) | magnetic minerals grow through their blocking volume below \( T_B \) | records the field at the time of growth, which may be long after the rock formed |
| VRM (viscous) | gradual acquisition in the present field | soft secondary component; removed early in demagnetisation |
| IRM (isothermal) | brief strong field at constant temperature (lightning, laboratory) | can overprint outcrops; used to identify minerals |
The natural remanent magnetisation (NRM) of a rock is the vector sum of a primary component (acquired at formation) and secondary components acquired later.
The Koenigsberger ratio compares remanent with induced magnetisation:
$$Q = \frac{M_{NRM}}{\chi F/\mu_0}$$Basalts commonly have \( Q > 1 \) (remanence dominates, which is why oceanic crust records reversals so clearly); many crustal rocks have \( Q < 1 \).
From Sample to Direction
Laboratory demagnetisation
- Alternating-field (AF) demagnetisation: a decaying alternating field of increasing peak value randomises grains with coercivity below the peak. Works well for magnetite; poorly for hematite and goethite.
- Thermal demagnetisation: heating in steps to higher temperatures in zero field unblocks grains with \( T_B \) below each step. Works for all carriers; temperatures above 580 °C that still leave remanence point to hematite.
- Chemical demagnetisation (acid leaching) removes secondary hematite pigment in red beds.
Zijderveld (orthogonal vector) diagrams
The remaining NRM vector after each step is projected on to a horizontal plane (north vs east) and a vertical plane (north or east vs down) on the same axes.
- A straight segment heading to the origin is a single component being removed: the characteristic remanent magnetisation (ChRM), usually taken as the primary direction.
- A segment that does not head to the origin means another component is still present; the early steps usually remove a secondary component (VRM, a later CRM).
- Curved paths mean overlapping blocking-temperature or coercivity spectra.
Directions are fitted by principal component analysis; the maximum angular deviation (MAD) measures how straight the segment is.
Fisher statistics
Unit vectors of \( N \) directions are summed; the resultant length \( R \le N \). The Fisher distribution, \( P(\theta) \propto e^{k\cos\theta} \), describes scatter about the true mean.
$$\boxed{k \approx \frac{N - 1}{N - R}} \qquad \boxed{\cos\alpha_{95} = 1 - \frac{N - R}{R}\left[\left(\frac{1}{0.05}\right)^{1/(N-1)} - 1\right]} \qquad \alpha_{95} \approx \frac{140°}{\sqrt{kN}}$$- \( k \) is the precision parameter: large \( k \) means tight clustering.
- \( \alpha_{95} \) is the radius of the 95% confidence cone about the mean.
- The angular standard deviation is roughly \( 81°/\sqrt{k} \).
For \( N = 9 \), \( R = 8.82 \): \( k = 8/0.18 = 44.4 \); the exact \( \alpha_{95} = 7.8° \); the approximation gives 7.0° (it underestimates for small \( N \)).
Field tests of stability and age
| Test | Positive result | Meaning |
|---|---|---|
| Fold test | directions cluster better after tilt correction (\( k \) increases) | remanence predates folding |
| Conglomerate test | directions in clasts are random | clasts’ remanence predates deposition and has not been reset |
| Baked-contact test | baked country rock near a dyke has the dyke’s direction; far away it differs | dyke remanence is primary (TRM at intrusion) |
| Reversal test | normal and reversed means are antiparallel within error | secondary overprints have been removed |
A negative fold test (better clustering before tilt correction) means the magnetisation was acquired after folding: a remagnetisation.
From Direction to Pole
The geocentric axial dipole (GAD) hypothesis
Averaged over \( 10^4 \)–\( 10^5 \) years, secular variation cancels and the field looks like a dipole at Earth’s centre aligned with the rotation axis. Then:
- the mean declination points to the geographic (palaeo) pole: \( D = 0° \) for normal polarity;
- the mean inclination gives the palaeolatitude:
| \( I \) | 0° | 20° | 30° | 40° | 50° | 60° | 70° | 80° |
|---|---|---|---|---|---|---|---|---|
| \( \lambda \) | 0° | 10.3° | 16.1° | 22.8° | 30.8° | 40.9° | 54.0° | 70.6° |
Virtual geomagnetic pole (VGP)
Given a site at latitude \( \lambda_s \), longitude \( \phi_s \), and a direction \( (D, I) \):
- Palaeocolatitude \( p \) (angular distance from site to pole): \( \cot p = \tfrac{1}{2}\tan I \).
- Pole latitude: \( \sin\lambda_p = \sin\lambda_s\cos p + \cos\lambda_s\sin p\cos D \).
- Longitude difference: \( \sin\beta = \dfrac{\sin p\sin D}{\cos\lambda_p} \).
- If \( \cos p \ge \sin\lambda_s\sin\lambda_p \): \( \phi_p = \phi_s + \beta \); otherwise \( \phi_p = \phi_s + 180° - \beta \).
A pole from a single spot reading is a VGP; the average of many VGPs spanning enough time to average out secular variation is a palaeomagnetic pole.
Confidence about a pole
A circular cone \( \alpha_{95} \) about the mean direction maps to an oval about the pole, with semi-axes
$$dp = \tfrac{1}{2}\alpha_{95}\left(1 + 3\cos^2 p\right) \ \ \text{(along the site–pole great circle)} \qquad dm = \alpha_{95}\,\frac{\sin p}{\cos I}\ \ \text{(perpendicular)}$$What palaeomagnetism can and cannot tell you
- Palaeolatitude: yes, from inclination.
- Rotation about a vertical axis: yes, from a declination that differs from the expected one for the same age and plate.
- Palaeolongitude: no. A GAD field is symmetric about the rotation axis, so moving a continent east or west along a line of latitude changes nothing in its record.
- Hemisphere: only if polarity is known, which needs a polarity sequence (magnetostratigraphy) or continuity of the path.
Inclination shallowing
Flattened, compacting sediments often record inclinations shallower than the field: \( \tan I_{obs} = f\tan I_{field} \) with \( f < 1 \). Ignoring this underestimates palaeolatitude. Lavas do not suffer from it.
Palaeointensity
Because TRM is proportional to the field (for weak fields), the ancient field strength can be found by replacing the NRM with a laboratory TRM acquired in a known field. In the Thellier method, heating steps alternately remove NRM and add partial TRM; on an Arai plot (NRM remaining against pTRM gained) the slope magnitude gives
$$B_{anc} = |\text{slope}|\times B_{lab}$$The result is often expressed as a virtual axial dipole moment (VADM), using the palaeolatitude:
$$\text{VADM} = \frac{4\pi R^3}{\mu_0}\,\frac{B_{anc}}{\sqrt{1 + 3\sin^2\lambda}}$$Alteration of minerals during heating is the main failure mode, checked by repeating earlier pTRM steps.
Reversals and Magnetostratigraphy
- Rocks of the same age worldwide show the same polarity, and reversals appear in lavas, sediments and ocean floor alike. Self-reversal in a few unusual minerals exists but cannot explain the global, synchronous record.
- A polarity transition takes a few thousand years; during it the intensity drops and VGPs wander far from the poles.
- Excursions are brief large departures that do not complete a reversal.
- The geomagnetic polarity time scale (GPTS) dates the sequence of chrons and subchrons from radiometric ages on lavas, astronomical tuning of sediments and marine anomaly profiles.
| Boundary or event | Approximate age |
|---|---|
| Brunhes (normal) / Matuyama (reversed) | about 0.78 Ma |
| Jaramillo normal subchron | about 0.99–1.07 Ma |
| Olduvai normal subchron | about 1.78–1.95 Ma |
| Matuyama / Gauss (normal) | about 2.58 Ma (base of the Quaternary) |
| Gauss / Gilbert (reversed) | about 3.6 Ma |
| Cretaceous Normal Superchron | roughly 121–84 Ma, no reversals |
Magnetostratigraphy matches the polarity sequence in a section to the GPTS, giving ages and sedimentation rates. See Quaternary Stratigraphy for its use alongside oxygen-isotope stages. The Deccan Traps illustrate it on a large scale: much of the lava pile carries reversed polarity assigned to chron C29r, the chron that contains the Cretaceous–Palaeogene boundary.
Polar Wandering
Apparent polar wander (APW)
Palaeomagnetic poles of successively younger ages from one continent, plotted in that continent’s present coordinates, form an apparent polar wander path (APWP). “Apparent” because it is mostly the continent that moved, not the pole.
- A segment along a small circle means rotation of the plate about a fixed Euler pole; cusps (sharp bends) mark changes in plate motion.
- Divergent APWPs from two continents that converge when the continents are restored to a former fit prove relative motion, as in the hook (Europe and North America).
- The Indian APWP records rapid northward motion through the Cretaceous and early Cenozoic, from southern-hemisphere latitudes to the present position, and a slowdown associated with collision with Asia. See Himalayan Orogeny and Supercontinents for the tectonic story.
True polar wander (TPW)
True polar wander is the rotation of the whole solid Earth (or of the mantle and crust together) relative to the rotation axis, driven by changes in the moment-of-inertia tensor as mass is redistributed (subducted slabs, mantle upwellings, ice sheets). The spin axis stays fixed in space; the Earth re-orients beneath it so that its axis of maximum moment of inertia lines up with the spin axis.
| Apparent polar wander | True polar wander | |
|---|---|---|
| What moves | individual plates relative to the pole | the whole solid Earth relative to the spin axis |
| Paths from different plates | differ | share a common component |
| Reference frame needed | plate’s own coordinates | a frame fixed to the deep mantle (for example, hotspots) |
| Cause | plate motion | redistribution of mass changing the inertia tensor |
To separate them, plate motions are expressed in a mantle-fixed (hotspot) frame; the motion of the palaeomagnetic pole common to all plates in that frame is TPW.
graph TD
A["Collect oriented samples
from several sites"] --> B["Demagnetise
(AF, thermal)"]
B --> C["Zijderveld plot:
isolate ChRM"]
C --> D["Fisher mean:
k, alpha95"]
D --> E["Field tests:
fold, baked contact,
conglomerate, reversal"]
E --> F["GAD: palaeolatitude
tan I = 2 tan lambda"]
F --> G["VGP, then palaeomagnetic
pole with dp, dm"]
G --> H["Order poles by age:
APW path"]
H --> I["Compare plates or a
hotspot frame:
drift vs true polar wander"]Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| Curie temperature (mineral property; ordering lost) | Blocking temperature (grain property; remanence frozen); always below Curie | |
| VGP (one spot reading) | Palaeomagnetic pole (average over secular variation) | |
| Primary remanence (formation) | Secondary (VRM, CRM, IRM overprints) | |
| Positive fold test (pre-folding) | Negative fold test (post-folding remagnetisation) | |
| Excursion (brief, incomplete) | Reversal (complete, lasting) | |
| Apparent polar wander (plate moves) | True polar wander (whole Earth reorients) | |
| VADM (uses palaeolatitude from GAD) | VDM (uses the observed inclination) | |
| \( \alpha_{95} \) (cone about the mean direction) | \( dp, dm \) (oval about the pole) |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Sites and data are hypothetical.
Level 1: Foundation
- What is the Curie temperature of magnetite?
- State the condition for a positive fold test.
- Why can palaeomagnetism not give palaeolongitude?
Answers
- About 580 °C.
- Directions cluster more tightly after correction for bedding tilt (higher \( k \)).
- The time-averaged field is a geocentric axial dipole, symmetric about the rotation axis, so it is the same at all longitudes for a given latitude.
Level 2: Part B Style
Practice Q1. Nine site-mean directions give a resultant length \( R = 8.82 \). The precision parameter and the exact \( \alpha_{95} \) are closest to:
(a) \( k = 44 \), \( \alpha_{95} = 7.8° \) (b) \( k = 50 \), \( \alpha_{95} = 6.6° \) (c) \( k = 44 \), \( \alpha_{95} = 2.1° \) (d) \( k = 5.6 \), \( \alpha_{95} = 19° \)
Answer: (a) k = 44, alpha95 = 7.8°
\( k = (N - 1)/(N - R) = 8/0.18 = 44.4 \).
\( \cos\alpha_{95} = 1 - \frac{0.18}{8.82}\left[20^{1/8} - 1\right] = 1 - 0.020408\times 0.4542 = 0.99073 \Rightarrow \alpha_{95} = 7.8° \).
(b) uses \( N/(N - R) \) for \( k \) and then the approximate \( 140°/\sqrt{kN} \); with the correct \( k \) the approximation gives 7.0°, still below the exact value. (c) and (d) do not follow from these data.
Practice Q2. A lava flow has an NRM of 2 A m\(^{-1}\) and a susceptibility of 0.01 (SI) in a present field of 45,000 nT. Its Koenigsberger ratio is about:
(a) 0.18 (b) 1.0 (c) 5.6 (d) 44
Answer: (c) 5.6
Induced magnetisation \( = \chi F/\mu_0 = 0.01\times 45000\times 10^{-9}/(4\pi\times 10^{-7}) = 0.358 \) A m\(^{-1}\).
\( Q = 2/0.358 = 5.6 \). Remanence dominates, as is typical for basalts. (a) is the inverse.
Practice Q3. In a deep-sea core, the Brunhes–Matuyama boundary (about 0.78 Ma) lies at 15.6 m below the sea floor. Assuming a constant sedimentation rate, at what depth would you expect the top of the Jaramillo subchron (about 0.99 Ma)?
(a) 12.3 m (b) 17.8 m (c) 19.8 m (d) 24.6 m
Answer: (c) 19.8 m
Rate \( = 15.6\ \text{m}/0.78\ \text{Myr} = 20 \) m Myr\(^{-1}\) (2 cm kyr\(^{-1}\)). Depth \( = 20\times 0.99 = 19.8 \) m.
Practice Q4. Which observation is the strongest evidence that a dyke’s remanence is primary?
(a) The dyke’s remanence has \( Q > 1 \). (b) Baked host rock next to the dyke carries the dyke’s direction, while unbaked host rock farther away carries a different direction. (c) Dyke samples cluster tightly (\( k = 200 \)). (d) The dyke’s direction lies close to the present field.
Answer: (b)
A positive baked-contact test shows the dyke’s heat reset the host rock at intrusion and that neither has been remagnetised since. (d) is a warning sign of a present-field (VRM) overprint; (c) shows precision, not age.
Level 3: Part C Style
Practice Q5. Rocks of age 70 Ma from a hypothetical plate show a mean inclination of −50° (normal polarity confirmed); the same plate today lies where the GAD inclination is +30°. Which statements are correct?
- The palaeolatitude at 70 Ma was about 30.8° S.
- The present latitude is about 16.1° N.
- The minimum northward displacement is about 5,200 km, averaging about 7.4 cm per year.
- The data also show that the plate moved eastward.
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 3 only
- 1 true: \( \tan\lambda = \tan(-50°)/2 = -0.596 \Rightarrow \lambda = -30.8° \).
- 2 true: \( \tan\lambda = \tan 30°/2 = 0.289 \Rightarrow \lambda = 16.1° \).
- 3 true: \( \Delta\lambda = 46.9° \times 111.2 \) km per degree \( \approx 5{,}214 \) km; over 70 Myr this is \( 5.214\times 10^8 \) cm \( / 7\times 10^7 \) yr \( \approx 7.4 \) cm yr\(^{-1}\). It is a minimum because any east–west motion is invisible.
- 4 false: palaeomagnetism cannot resolve longitude.
Practice Q6. Ten sites in folded sandstones give: before tilt correction \( R = 8.90 \); after tilt correction \( R = 9.85 \). Which conclusion is correct?
(a) \( k \) falls from 60 to 8 on unfolding: the remanence was acquired after folding. (b) \( k \) rises from about 8 to 60 and \( \alpha_{95} \) shrinks from about 18° to about 6°: the remanence predates folding. (c) \( k \) rises from about 8 to 60, so the remanence is a recent VRM. (d) The test is inconclusive because both \( R \) values are less than \( N \).
Answer: (b)
Before: \( k = 9/1.10 = 8.2 \); \( \cos\alpha_{95} = 1 - \frac{1.10}{8.90}(20^{1/9} - 1) = 1 - 0.1236\times 0.3950 = 0.9512 \Rightarrow \alpha_{95} \approx 18.0° \).
After: \( k = 9/0.15 = 60 \); \( \cos\alpha_{95} = 1 - \frac{0.15}{9.85}(0.3950) = 0.99398 \Rightarrow \alpha_{95} \approx 6.3° \).
Clustering improves markedly on unfolding: a positive fold test, so the ChRM was acquired before folding. (In a real study the increase would be checked for significance with a precision-ratio test.) (d) is wrong: \( R < N \) always unless every direction is identical.
Practice Q7. A Thellier experiment on a lava from palaeolatitude 30° gives an Arai plot slope of −0.80 with a laboratory field of 40 μT. Taking \( R = 6{,}371 \) km, the palaeointensity and VADM are closest to:
(a) 32 μT, \( 6.3\times 10^{22} \) A m\(^2\) (b) 50 μT, \( 9.8\times 10^{22} \) A m\(^2\) (c) 32 μT, \( 8.3\times 10^{22} \) A m\(^2\) (d) 32 μT, \( 4.1\times 10^{22} \) A m\(^2\)
Answer: (a) 32 μT, 6.3 x 10^22 A m^2
\( B_{anc} = 0.80\times 40 = 32 \) μT.
\( \text{VADM} = \frac{4\pi R^3}{\mu_0}\frac{B}{\sqrt{1 + 3\sin^2 30°}} = \frac{(6.371\times 10^6)^3\times 32\times 10^{-6}}{10^{-7}\times\sqrt{1.75}} = \frac{8.275\times 10^{22}}{1.323} = 6.26\times 10^{22} \) A m\(^2\).
(c) omits the latitude factor (treats the site as equatorial); (b) divides by the slope instead of multiplying.
Quick Revision Table
| Idea | Fact or formula |
|---|---|
| Remanence carriers | magnetite (580 °C), hematite (675 °C), titanomagnetite, pyrrhotite (about 320 °C) |
| Best recorders | single-domain grains |
| Néel relaxation | \( \tau = \tau_0\exp(Kv/k_BT) \), \( \tau_0 \approx 10^{-9} \) s |
| Remanence types | TRM, DRM/pDRM, CRM, VRM, IRM; NRM = primary + secondary |
| Koenigsberger | \( Q = \text{NRM}/(\chi F/\mu_0) \); basalts \( Q > 1 \) |
| Demagnetisation | AF (magnetite), thermal (all; hematite above 600 °C), chemical |
| Zijderveld | straight line to origin = ChRM |
| Fisher | \( k \approx (N - 1)/(N - R) \); \( \alpha_{95} \approx 140°/\sqrt{kN} \) |
| Field tests | fold, conglomerate (random = positive), baked contact, reversal |
| GAD | \( \tan I = 2\tan\lambda \); \( D = 0 \) |
| VGP | \( \cot p = \tfrac{1}{2}\tan I \); \( \sin\lambda_p = \sin\lambda_s\cos p + \cos\lambda_s\sin p\cos D \) |
| Pole oval | \( dp = \tfrac{1}{2}\alpha_{95}(1 + 3\cos^2 p) \); \( dm = \alpha_{95}\sin p/\cos I \) |
| Palaeointensity | Thellier; \( B = \lvert\text{slope}\rvert B_{lab} \); VADM with \( \sqrt{1 + 3\sin^2\lambda} \) |
| GPTS anchors | B/M 0.78 Ma; Jaramillo 0.99–1.07; Olduvai 1.78–1.95; M/G 2.58; G/G 3.6 Ma |
| Superchron | Cretaceous Normal, about 121–84 Ma |
| APW vs TPW | plate motion vs whole-Earth reorientation; TPW from a hotspot frame |
| Not resolvable | palaeolongitude |
Key Takeaways
What’s Next
- Gravity and Magnetic Fields of the Earth: Quick Revision: the whole unit on one page.
- Quaternary Dating Methods: where magnetostratigraphy sits among the dating tools.
- Normal Gravity, Clairaut’s Theorem and the Shape of the Earth: start the unit again from the gravity side.