World Magnetic Model

The maintained US/UK spherical-harmonic model series that maps WGS-84 position, altitude, and date within a five-year epoch to Earth's large-scale main magnetic-field components, their secular change, and navigation quantities such as declination.

Version

v1 · 2026-08-30 · History

Domain-specific #

3127

Origin domain

geomagnetism

Subdomain

operational geomagnetic field modeling

Aliases

WMM

Core Idea

The World Magnetic Model (WMM) is the maintained US/UK operational model series used to estimate Earth's large-scale main magnetic field near the planet for navigation, attitude, and heading reference. A release combines a spherical-harmonic description of the field at a reference epoch with a spherical-harmonic estimate of its slow annual change. Standard software accepts geodetic latitude, longitude, altitude, and decimal year; evaluates the time-adjusted field in geocentric coordinates; rotates it into the local geodetic frame; and returns the north, east, and downward field components together with horizontal intensity, total intensity, inclination, and declination.^[1][2]

For the standard WMM, the magnetic scalar potential outside the field sources has the form

\[ V(r,\theta,\lambda,t)=a\sum_{n=1}^{12}\left(\frac{a}{r}\right)^{n+1} \sum_{m=0}^{n} \left[g_n^m(t)\cos(m\lambda)+h_n^m(t)\sin(m\lambda)\right] \bar P_n^m(\cos\theta), \]

where \(a\) is a reference radius, \((r,\theta,\lambda)\) are geocentric spherical coordinates, \(\bar P_n^m\) are normalized associated Legendre functions, and \(g_n^m,h_n^m\) are Gauss coefficients. The field is \(\mathbf B=-\nabla V\). Within one model epoch, each coefficient is advanced linearly:

\[ g_n^m(t)=g_n^m(t_0)+(t-t_0)\dot g_n^m(t_0),\qquad h_n^m(t)=h_n^m(t_0)+(t-t_0)\dot h_n^m(t_0). \]

The current standard release, WMM2025, uses degree and order 12 for both the main field and secular variation, is intended for the 2025.0–2030.0 cycle, and replaced WMM2020. It is one release of the abstraction, not the abstraction itself. The WMM identity persists through repeated coefficient replacement, common evaluation software, validation, uncertainty reporting, and a normally five-year maintenance cycle.^[1][3]

Structural Signature

The stable structure is:

filtered global geomagnetic observations + epoch-fitted main-field coefficients + secular-variation coefficients + bounded linear time advance + spherical-harmonic evaluation + WGS-84 coordinate conversion + standardized magnetic elements and uncertainty zones → a replaceable five-year navigation field model.

The mandatory roles are:

• Physical target — Earth's large-scale, predominantly core-generated main magnetic field and its slow temporal change.
• Reference epoch \(t_0\) — the date at which a release's main-field coefficients are fitted.
• Spectral representation — Gauss coefficients through degree and order 12 in the standard model. The 90 \((n,m)\) rows encode 168 independent \(g/h\) coefficient values because \(h_n^0\) is zero or undefined; an analogous set describes secular variation.^[1]
• Short-horizon forecast — coefficient rates \(\dot g_n^m,\dot h_n^m\) linearly advance the field only within the declared five-year epoch.
• Geodetic query — latitude, longitude, WGS-84-referenced altitude, and date locate a requested prediction.
• Evaluation pipeline — geodetic-to-geocentric conversion, potential evaluation, gradient calculation, and rotation to local north-east-down components.
• Derived elements — \(X,Y,Z,H,F,I,D\), their annual changes, uncertainty estimates, and warnings where heading is ill-conditioned.
• Maintenance covenant — annual performance review, bounded validity, and replacement or exceptional correction as the observed field departs from the forecast.

The invariant is not any particular coefficient table. It is the maintained mapping contract from position and in-epoch date to standardized large-scale geomagnetic elements under this model family.

What It Is Not

WMM is not a magnetometer measurement, a live feed, or the total instantaneous magnetic field at a device. Local crustal anomalies, nearby iron, electrical systems, and rapidly varying ionospheric and magnetospheric currents can make an observation differ materially from the prediction. NOAA explicitly separates omission error from fields not represented by WMM and commission error from model fitting and secular-variation forecasting.^[4][5]

It is not a magnetic-pole locator alone. Pole positions, blackout zones, maps, and declination values are outputs derived from the same field model. It is not a compass correction table; software evaluates the model for a query.

It is not WMM2025, WMM2020, or another dated coefficient file. Those are releases. It is not WMMHR2025, the separate high-resolution companion introduced in the 2025 generation: WMMHR extends the time-varying main field to degree 15 and includes a crustal field to degree 133, reaching roughly 300-km rather than standard WMM's roughly 3,330-km equatorial resolution.^[6] It is not the Enhanced Magnetic Model, High Definition Geomagnetic Model, or International Geomagnetic Reference Field, each of which has a different resolution, field-source, retrospective, update, or operational contract.

Scope of Application

The WMM is scoped to operational prediction of the large-scale geomagnetic field for navigation and orientation near Earth over a declared epoch. NGA identifies it as the standard for US and UK defense users and for organizations including NATO and the International Hydrographic Organization; the same model is widely used in civilian heading systems.^[7] It supports magnetic-to-true heading correction, aircraft and marine navigation, attitude reference, directional drilling, survey reduction, runway-number review, and software that needs field-vector or declination estimates at a location and date.^[1][8]

The scope has explicit temporal, spatial, and source limits. A current release is valid for five years because the core field's change is only approximated by constant secular-variation coefficients. Standard WMM resolves continental-scale structure, not local anomalies. Declination becomes unreliable where horizontal intensity approaches zero near the magnetic dip poles. Geomagnetic storms can temporarily produce large departures, especially at high magnetic latitudes.^[5]

This is an operational model series, not a general theory of geomagnetism. Research field models may use more coefficients, assimilate different observations, reconstruct historical epochs, or include crustal and external fields without becoming WMM.

Clarity

Recognize a WMM computation by asking six questions:

1. Is the coefficient set an identified WMM release and epoch, not an arbitrary spherical-harmonic field?
2. Does the standard model use degree/order 12 main-field and secular-variation coefficients?
3. Is the requested time within the release's five-year validity interval?
4. Are geodetic coordinates and height interpreted under the declared WGS-84 conventions and converted correctly?
5. Are the returned components derived from \(\mathbf B=-\nabla V\) and expressed in the documented local frame?
6. Are uncertainty, omitted sources, and blackout/caution zones honored rather than suppressed?

The output components supply an internal consistency check:

\[ H=\sqrt{X^2+Y^2},\qquad F=\sqrt{H^2+Z^2},\qquad D=\operatorname{atan2}(Y,X),\qquad I=\operatorname{atan2}(Z,H). \]

Here \(X\) is northward, \(Y\) eastward, and \(Z\) downward under the documented convention. A number labeled “declination” without model release, date, coordinate, height convention, and zone status is not a fully specified WMM result.

Manages Complexity

Earth's observed magnetic field varies over the globe, with altitude, and over time. Directly carrying satellite tracks, observatory series, selection rules, coordinate transformations, and every measurement into each navigation device would be impractical. WMM compresses those observations into a small replaceable coefficient file and a stable evaluation algorithm. NOAA's documented WMM.COF format is deliberately separable from the software so a new model can ordinarily be installed by replacing coefficients rather than rewriting the evaluation code.^[1]

The spherical-harmonic basis turns a global vector field into finitely many coefficients with controlled spatial scale. The secular-variation field turns a separate static model for every day into a linear in-epoch forecast. Derived elements convert a vector into the quantities different users need. Error models and blackout zones turn “the model may be wrong” into operational limits.

The maintenance cycle manages the complexity that the compression creates. A low-degree model is tractable and globally smooth but omits smaller-scale fields. Linear secular variation is easy to evaluate but eventually goes stale as core flow changes nonlinearly. Annual comparison with newer observations tests whether those simplifications remain within specification; scheduled replacement prevents indefinite extrapolation.

Abstract Reasoning

Predictive inference. Given a valid release, WGS-84 location, altitude, and in-epoch date, advance every coefficient from \(t_0\), evaluate \(V\), differentiate, and derive \(X,Y,Z,H,F,I,D\). The equations specify the order: time adjustment before field evaluation, geocentric computation before local-frame interpretation.

Diagnostic inference. If a device's compass differs from WMM, separate model failure from omitted signal. A stable local discrepancy can indicate crustal or man-made magnetic influence; a storm-time high-latitude discrepancy can indicate external disturbance; a growing global residual can indicate secular-variation forecast error. WMM is a baseline against which residual causes are classified, not proof that the measurement is wrong.

Validity inference. Never extrapolate a release indefinitely. A dated coefficient file carries an epoch, and linear secular variation is warranted only over its cycle. Using WMM2020 in 2026 is a version error even if software still produces a number.

Conditioning inference. Declination is \(\operatorname{atan2}(Y,X)\). When \(H=\sqrt{X^2+Y^2}\) becomes small, small component errors cause a large angular error. Blackout and caution zones follow from this mathematical ill-conditioning, not merely institutional conservatism. The 2025 performance report defines blackout regions by \(H<2000\) nT and caution regions by \(2000\le H<6000\) nT.^[5]

Intervention inference. If required spatial distinctions are smaller than standard WMM's scale, change models rather than over-interpreting its output. WMMHR, EMM, or a local survey may be appropriate, depending on whether the missing structure is large-scale core evolution, crustal field, external disturbance, or local magnetic noise.

Knowledge Transfer

The complete mechanism transfers literally across navigation, surveying, geodesy, mobile-device orientation, aerospace, and directional drilling. In each case, the same position/date query, coefficient advance, harmonic evaluation, field components, and limits are used; only the downstream decision changes.

WMM does not transfer as a prime across unrelated substrates. Spherical harmonics, Gauss coefficients, magnetic elements, WGS-84 geometry, secular variation, and compass blackout zones are constitutive. The portable structures are already represented by catalog primes: representation for the target-to-coefficient mapping and calibrated_rule_vs_moving_world for the need to monitor and replace a model fitted to an evolving field. The named WMM is one domain-specific realization with an official product lifecycle.

Examples

Canonical field query. A navigation system has geodetic latitude, longitude, ellipsoidal or converted altitude, and decimal date. It loads the coefficient file for the valid WMM epoch, linearly advances \(g_n^m,h_n^m\), computes the field components, and returns declination \(D\). The system converts a true heading to a magnetic reference—or the reverse—using that local predicted angle. The WMM result is an estimate of the main field; the magnetometer still needs calibration for platform and local disturbances.

Runway numbering. Runway numbers encode magnetic heading rounded to a tens-of-degrees convention. Because declination changes, the magnetic bearing associated with a physically fixed runway can cross a renumbering threshold. NCEI documents airport runway name changes as a practical consequence of the evolving field and WMM-supported declination estimates.^[8] The model does not decide the administrative change; it provides the standardized magnetic reference used in the review.

Polar boundary case. Near a dip pole, \(X\) and \(Y\) are both small. The model can still return a field vector, but declination—the azimuth of the horizontal projection—is unstable. Correct WMM use preserves the blackout or caution warning rather than treating the angular output as ordinary navigation guidance.

Release maintenance. WMM2015's secular-variation forecast became inadequate in part of the high northern region after unexpectedly rapid magnetic-pole motion, prompting the exceptional WMM2015v2 update. That event exemplifies the family invariant: observed performance governs correction and replacement; a release is not allowed to define the moving world by fiat. Regular WMM2020 and WMM2025 releases then continued the maintained series.^[9][5]

Structural Tensions

Compression versus local fidelity. Degree 12 makes a compact, robust global model but cannot resolve fine crustal or man-made anomalies. Diagnostic: compare the task's spatial scale and field sources with the model's stated resolution and omissions.

Prediction horizon versus nonlinear change. Linear secular variation makes five-year deployment feasible, while evolving core dynamics make longer extrapolation increasingly unreliable. Diagnostic: check the model epoch and current annual performance before trusting a computed date.

Stable software versus changing coefficients. A fixed evaluator simplifies implementation, but stale WMM.COF data silently produce obsolete results. Diagnostic: record and expose the loaded coefficient epoch independently of application version.

Global standard versus local observation. Standardization gives interoperable headings, yet the actual field at a device includes sources WMM intentionally omits. Diagnostic: decide whether the task needs the standardized main-field reference or the instantaneous local vector.

Numerical output versus physical conditioning. Software can emit declination everywhere, but an angle derived from a near-zero horizontal vector is not operationally reliable. Diagnostic: carry \(H\), uncertainty, and blackout/caution status with \(D\).

Structural–Framed Character

World Magnetic Model is mixed structural. Its computational core is formal: a field target, spectral medium, coordinate mapping, time forecast, differentiation, and derived quantities. Any competent implementation using the same release should reproduce test values.

Its full identity is partly framed by institutional maintenance. “WMM” denotes the joint NGA/DGC product developed and distributed through NOAA NCEI and BGS, governed by a military performance specification and operational update cycle.^[7][5] An independently fitted degree-12 geomagnetic model with identical equations would be WMM-like, but it would not thereby become the World Magnetic Model. The institutional frame names the authoritative coefficient lineage; it does not alter the underlying mathematics.

Structural Core vs. Domain Accent

The structural core is a versioned predictive representation: fit coefficients at an epoch, pair them with rates of change, map a coordinate-and-time query to derived values, declare an error model, monitor residuals, and replace the calibration before drift becomes unacceptable. That skeleton can recur in other maintained environmental models.

The domain accent is indispensable to this node: a geomagnetic scalar potential, spherical harmonics on Earth, WGS-84 coordinates, north-east-down components, nanotesla units, declination and inclination, core-field secular variation, pole conditioning, and US/UK operational stewardship. Removing those roles yields generic Representation or Calibrated Rule versus Moving World, not WMM.

Instantiates / Related Primes

World Magnetic Model strictly specializes representation. The target is Earth's main magnetic field; the medium is an epoch-labeled coefficient file plus evaluation software; the mapping is the time-adjusted harmonic expansion; and the faithfulness contract specifies represented sources, spatial scale, validity interval, uncertainty, and excluded signals. This is the proposed minimal DAG parent.

WMM also instantiates calibrated_rule_vs_moving_world: a coefficient set calibrated to observations loses accuracy as the core field evolves, so annual monitoring and five-year recalibration are integral. That is an important related-prime reading but not a taxonomic parent—WMM is a model, not a subtype of a degradation dynamic. frame_of_reference, approximation, spherical harmonics, and least-squares adjustment are ingredients or production relations rather than additional minimal parents.

Relationships to Other Abstractions

Current abstraction World Magnetic Model Domain-specific

Parents (1) — more general patterns this builds on

• World Magnetic Model is a kind of Representation Prime

  World Magnetic Model strictly specializes representation.

Hierarchy path (1) — routes to 1 parentless root

• World Magnetic Model → Representation → Abstraction

Neighborhood in Abstraction Space

World Magnetic Model sits in a sparse region of the domain-specific corpus (85^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

• Precise Point Positioning — 0.82
• Mesoscale Eddy — 0.81
• Primitive Equations — 0.80
• Karlsruhe Metric — 0.80
• Planetary Boundaries — 0.80

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

• WMM2025 or WMM2020. A dated release/instance, not the maintained model family.
• WMMHR2025. The 2025 high-resolution companion with degree-15 time-varying main field and degree-133 crustal field, not an alias for standard degree-12 WMM.
• International Geomagnetic Reference Field (IGRF). A comparable international scientific reference model with retrospective definitive updates; WMM is the predictive operational US/UK standard and does not retrospectively revise old epochs.^[4][6]
• Enhanced Magnetic Model or High Definition Geomagnetic Model. Higher-resolution products that include crustal contributions and follow different update contracts.
• A magnetometer or compass. Measures the local instantaneous field; WMM predicts a large-scale reference field.
• Magnetic declination. One derived angle, not the full vector model.
• World Geodetic System 1984. The geodetic reference used for coordinates and altitude handling, not the magnetic model itself.
• A live space-weather model. WMM does not forecast short-period external disturbance fields or geomagnetic storms.

References

[1] NOAA National Centers for Environmental Information, “World Magnetic Model (WMM),” current product specification, software, coefficient format, inputs, outputs, and release history. https://www.ncei.noaa.gov/products/world-magnetic-model registry ↩a ↩b ↩c ↩d ↩e

[2] Arnaud Chulliat et al., The US/UK World Magnetic Model for 2025–2030: Technical Report (NOAA National Centers for Environmental Information and British Geological Survey, 2025). https://doi.org/10.25923/prbc-s316 registry ↩

[3] NOAA NCEI Geomagnetic Modeling Team and British Geological Survey, World Magnetic Model 2025 dataset (2024). https://doi.org/10.25921/aqfd-sd83 registry ↩

[4] NOAA National Centers for Environmental Information, “Geomagnetism Frequently Asked Questions,” especially model comparison, accuracy, omitted sources, and validity sections. https://www.ncei.noaa.gov/products/geomagnetism-frequently-asked-questions registry ↩a ↩b

[5] NOAA NCEI and British Geological Survey, December 2025 State of the Geomagnetic Field (2025), DOI 10.25923/2gcf-s393. https://www.ncei.noaa.gov/sites/default/files/2025-12/WMM%20SoGF%20Dec2025%20508.pdf registry ↩a ↩b ↩c ↩d ↩e

[6] NOAA National Centers for Environmental Information, “World Magnetic Model High Resolution,” model comparison table. https://www.ncei.noaa.gov/products/world-magnetic-model-high-resolution registry ↩a ↩b

[7] National Geospatial-Intelligence Agency, “NGA, NOAA, BGS Release World Magnetic Model 2025,” 17 December 2024. https://www.nga.mil/news/NGA_NOAA_BGS_Publish_World_Magnetic_Model_2025.html registry ↩a ↩b

[8] NOAA National Centers for Environmental Information, “Airport Runway Names Shift with Magnetic Field,” 16 November 2017. https://www.ncei.noaa.gov/news/airport-runway-names-shift-magnetic-field registry ↩a ↩b

[9] NOAA National Geophysical Data Center, “World Magnetic Model — Old Reports,” including the WMM2015v2 technical note and earlier technical reports. https://www.ngdc.noaa.gov/geomag/WMM/WMM_old_reports.shtml registry ↩

[10] British Geological Survey, “World Magnetic Model (WMM),” current model-family and WMMHR boundary. https://geomag.bgs.ac.uk/research/modelling/WorldMagneticModel registry