Correspondence/Findings

Introduction

The International Geomagnetic Reference Field (IGRF) is a series of mathematical models describing the large-scale internal part of the Earth’s magnetic field between epochs 1900 A.D. and the present. The IGRF has been maintained and produced by an international team of scientists under the auspices of the International Association of Geomagnetism and Aeronomy (IAGA) since 1965 (Zmuda 1971). It results from a collaborative effort between magnetic field modelers and institutes involved in collecting and disseminating magnetic field data from magnetic observatories (see the Appendix for the list of World Data Centers), ground surveys, and low Earth orbiting (LEO) satellites. The IGRF is used by scientists in a wide variety of studies, for instance, concerning the dynamics of the Earth’s core field, space weather, or local magnetic anomalies imprinted in the Earth’s crust. It is also used by commercial organizations and individuals as a source of orientation information.

The IGRF model must be regularly revised in order to follow the continuous temporal changes of the geomagnetic field generated in the Earth’s outer core. The period between revisions is however sufficiently short to preserve its utility as a reference model in applications requiring a fixed reference standard. Table 1 provides the nomenclature and a summary of the history of previous generations of the IGRF. At present, each generation consists of three constituent models. One constituent is designated a Definitive Geomagnetic Reference Field (DGRF). The term ‘definitive’ is used because any further improvement of these retrospectively determined models is unlikely. The second constituent model, referred to as an IGRF model, is non-definitive - it will eventually be replaced by a definitive model in a future revision of the IGRF. The final constituent, referred to as the secular variation (SV), is provided to predict the time variation of the large-scale geomagnetic field for the 5 years following the latest revision of the IGRF. Readers interested in the history of the IGRF should consult Barton (1997), and users can find legacy versions of the IGRF at the online archive located at http://www.ngdc.noaa.gov/IAGA/vmod/igrf_old_models.html. These may prove useful for those wishing to recover data from which a previous generation of the IGRF has been subtracted or who wish to use the latest generation of the IGRF to carry out revised analyses. Here, attention will focus on the most recent 12th-generation IGRF, hereafter referred to as IGRF-12, that provides a DGRF model for epoch 2010.0, an IGRF model for epoch 2015.0, and a predictive SV model covering the epochs 2015.0-2020.0. IGRF-12 was agreed in December 2014 by a task force of the IAGA Working Group V-MOD. The purpose of this note is to document the release of IGRF-12, to act as a permanent published record of the IGRF-12 set of model coefficients, and to briefly describe some major features of the geomagnetic field at the Earth’s surface as revealed by the updated model.

Table 1 Summary of IGRF generations, their intervals of validity, and related references
Full size table

Mathematical formulation of the IGRF model

The IGRF is a series of mathematical models of the internal geomagnetic field \(\overrightarrow {B}(r,\theta,\phi,t)\) and its annual rate of change (secular variation). On and above the Earth’s surface, the magnetic field \( \overrightarrow {B}\) is defined in terms of a magnetic scalar potential V by \(\overrightarrow {B}=-\nabla V\) and where in spherical polar co-ordinates V is approximated by the finite series

$$ {\fontsize{9.6pt}{9.6pt}\selectfont{\begin{aligned} {\kern-16.5pt}V(r,\theta,\phi,t)&=a\sum_{n=1}^{N}\sum_{m=0}^{n}\left(\frac{a}{r}\right)^{n+1}\\ &\times\left[ {g_{n}^{m}}(t)\cos \left(m\phi \right) +{h_{n}^{m}}(t)\sin \left(m\phi \right) {P_{n}^{m}}(\cos \theta)\right] \!, \end{aligned}}} $$
((1))

with r denoting the radial distance from the center of the Earth, a=6,371.2 km being the geomagnetic conventional Earth’s mean reference spherical radius, θ denoting geocentric co-latitude, and ϕ denoting east longitude. The functions \({P_{n}^{m}}(\cos \theta)\) are the Schmidt quasi-normalized associated Legendre functions of degree n and order m. The Gauss coefficients \({g_{n}^{m}}\), \({h_{n}^{m}}\) are functions of time and are conventionally given in units of nanotesla (nT).

In the IGRF-12 model, the Gauss coefficients \({g_{n}^{m}}\) and \({h_{n}^{m}}\) are provided for the main field (MF) at epochs separated by 5 years between 1900.0 and 2015.0 A.D. The time dependence of the Gauss coefficients is assumed to be linear over 5-year intervals and is specified by the following expression

$$ \begin{aligned} {g_{n}^{m}}(t)& ={g_{n}^{m}}(T_{0})+\overset{.}{g}_{n}^{m}(T_{0}).(t-T_{0}), \end{aligned} $$
((2))
$$ \begin{aligned} {h_{n}^{m}}(t)& ={h_{n}^{m}}(T_{0})+\overset{.}{h}_{n}^{m}(T_{0}).(t-T_{0}), \end{aligned} $$
((3))

where \(\overset {.}{g}_{n}^{m} \left (\text {respectively}\; \overset {.}{h}_{n}^{m}\right)\) given in units of nT/year represent the 5-year average first time derivative (the linear secular variation) of the Gauss coefficients. t is the time of interest in units of year and T 0 is the epoch preceding t which is an exact multiple of 5 years, such that T 0≤t<(T 0+5.0). When MF models exist for both T 0 and T 0+5.0, then coefficients \(\dot { {g_{n}^{m}}}(T_{0})\) can be computed as \( [{g_{n}^{m}}(T_{0}+5.0)-{g_{n}^{m}}(T_{0})]/5.0\). For the final 5 years of the model validity (between 2015.0 and 2020.0 for IGRF-12), the coefficients \( \dot {{g_{n}^{m}}}(t)\) and \(\dot {{h_{n}^{m}}}(t)\) of the predictive average SV are explicitly provided. The geocentric components of the geomagnetic field in the northward, eastward, and radially inwards directions (X, Y and Z) are obtained from the model coefficients using Equation 1 and by taking the gradient of V in spherical polar co-ordinates

$$ X={\frac{1}{r}}{\frac{\partial V}{\partial \theta }},\quad Y=-{\frac{1}{ r\sin \theta }}{\frac{\partial V}{\partial \phi }},\quad Z={\frac{\partial V }{\partial r}}. $$
((4))

For some applications, the declination D, the inclination I, the horizontal intensity H, and the total intensity F are required. These components are calculated from X, Y, and Z using the relations,

$$ \begin{aligned} {}H&=\sqrt{X^{2}+Y^{2}},\quad F=\sqrt{X^{2}+Y^{2}+Z^{2}},\quad \\ D&=\arctan {(Y/X)},\quad I=\arctan {(Z/H)}. \end{aligned} $$
((5))

In Equation 1, the maximum spherical harmonic degree of the expansion N may vary from one epoch to another. The maximum degree N of the series is equal to 10 up to and including epoch 1995.0 and the coefficients are quoted to 1-nT precision. For epoch 2000, the coefficients are provided to degree and order 13 and quoted to 0.1-nT precision, and from epoch 2005 onwards they are quoted to 0.01-nT precision for the DGRF (and 0.1 nT for the latest non-definitive IGRF), to take advantage of the higher data quality and good coverage provided by the LEO satellite missions (Finlay et al. 2010a). The maximum truncation degree N=13 for epochs after 2000 is defined so as not to include the crustal magnetic field contributions that dominate at higher degrees (see e.g., Langel and Estes 1982).

The predictive SV coefficients \(\dot {{g_{n}^{m}}}(t)\) and \(\dot {{h_{n}^{m}}}(t)\) are given to degree and order 8 to 0.1-nT/year precision. Because of these changes in precision and nomenclature, it is recommended to always use the term ’IGRF-gg,’ where gg represents the generation, in order to keep track of the coefficients that were actually used in applications. This is a simple way to standardize studies carried out at different epochs that makes it apparent whether the results are ‘predictive’ and therefore less accurate or ’definitive’. For example, one cannot recover the original full-field measurement from an aeromagnetic anomaly map if one does not know which generation of the IGRF was used. This issue has important consequences when comparing magnetic surveys carried out at different epochs (e.g., Hamoudi et al. 2007; Hemant et al. 2007; Maus et al. 2007).

Equation 1 is expressed in the geocentric system of co-ordinates, but it is sometimes necessary to work in geodetic co-ordinates. When converting between geocentric and geodetic co-ordinates (see for instance Hulot et al. 2007), it is recommended to use the World Geodetic System 1984 (WGS84) datum as present-day satellite magnetic data are often positioned using it. The WGS84 spheroid is defined with major (equatorial) radius A = 6,378.137 km at the equator and a reciprocal flattening f = 1/298.257223563 (the polar semi-minor axis is therefore B = A(1-f) ≃6,356.752 km).

The 12th-generation IGRF

IGRF-12, the 12th generation of IGRF, is derived from candidate models prepared by international teams who answered a call issued by the IGRF-12 task force in May 2014. This call requested candidates for the Definitive Geomagnetic Reference Field (DGRF) for epoch 2010, for a provisional IGRF model for epoch 2015, and for a predictive SV model for the interval 2015.0-2020.0. The IGRF-12 model coefficients remain unchanged for epoch 2005 and earlier.

The number of institutions participating in IGRF-12 was larger than for any previous generation. This reflects the constructive effect of open and unconditional cooperation between scientists involved in modeling the magnetic field, the institutions archiving and disseminating the ground magnetic data, and the national and the European space agencies who actively worked to distribute their expertise, computer programs, and magnetic satellite data with documentation. This latter point was especially important for the MF for epoch 2015.0 given the short period that elapsed between the launch of the Swarm satellites (in November 2013) and the submission of IGRF candidate models by October 2014. The European Space Agency provided prompt access to the Swarm satellite measurements, including detailed documentation and information on the operational status of the instruments (https://earth.esa.int/web/guest/missions/esa-operational-eo-missions/swarm). This allowed the teams producing candidate models to rapidly use the Swarm data and helped IGRF-12 to be delivered on time. The collection of ground-based magnetic observatory measurements (see Table 2) and the availability of other satellite measurements, from the CHAMP (Reigber et al. 2002), Ørsted (Neubert et al. 2001) and SAC-C missions, were also crucial for IGRF-12.

Table 2 Magnetic observatories contributing data used in the construction of IGRF-12
Full size table

Seven candidate MF models for the DGRF epoch 2010.0 and ten candidate MF models for the IGRF epoch 2015.0 were submitted. In addition, nine SV models were submitted for the predictive part covering epochs 2015.0-2020.0. Team A was from BGS, UK (Hamilton et al. 2015); team B was from DTU Space, Denmark (Finlay et al. 2015); team C was led by ISTerre, France, with input from DTU Space (Gillet et al. 2015); team D was from IZMIRAN, Russia; team E was from NGDC/NOAA (Alken et al. 2015); team F was from GFZ, Germany (Lesur et al. 2015); team G was led by GSFC-NASA, USA, in collaboration with UMBC; team H was from IPGP (Fournier et al. 2015; Vigneron et al. 2015), France, in collaboration with the CEA-Léti (Léger et al. 2015) and with input from LPG Nantes and CNES, France; team I was led by LPG Nantes, France (Saturnino et al. 2015) with input from CNES; team J was from ETH Zurich, Switzerland. These teams contributed to all or parts of the three model constituents of IGRF. Following the IGRF specifications, the MF candidate models had a maximum spherical harmonic degree N=13 and the predictive SV model had a maximum spherical harmonic degree N=8.

The final IGRF-12 MF models for epochs 2010.0 and 2015.0 as well as the predictive SV model for 2015.0-2020.0 were calculated using a new weighting scheme of the candidate models. For the previous generation of IGRF, fixed weights were assigned to each candidate model based on information gleaned from the evaluations (see Finlay et al. 2010b, for instance) and most weight was given to those models showing the smallest scatter about the arithmetic mean of the candidate models. For IGRF-12, the evidence for significant systematic errors in one or more models was not thought to be sufficient to reject any of the models. A robust weighting scheme was instead applied to the candidate models in space, as agreed by a vote of the IGRF-12 task force. The specification of the candidate models and details of the evaluations and weighting scheme are described in a dedicated paper in this special issue (Thébault et al. 2015).

IGRF-12 model coefficients and maps

Table 3 lists the Schmidt semi-normalized spherical harmonic coefficients defining IGRF-12. In IGRF-12, only coefficients after epoch 2005.0 are modified, but all coefficients are included to serve as a complete record of the model since 1900. This should help to avoid any confusion with previous generations of IGRF, particularly with their provisional parts. The coefficients are given in units of nT for the MF models and of nT/year for the predictive SV model. The coefficients are also available at http://www.ngdc.noaa.gov/IAGA/vmod/igrf.html, together with software to compute the magnetic field components at times and locations of interest, in both geodetic and geocentric reference frames. IGRF-12 is also available from the World Data Centers listed at the end of this paper.

Table 3 12th Generation International Geomagnetic Reference Field
Full size table

We display in Figure 1 maps of the declination D, inclination I, and total intensity F in 2015.0 on the Earth’s reference sphere (r=a) in a Mercator projection that is well suited to navigation. The green lines are the zero contours; in the declination map, the line shows the agonic line where true geographic and magnetic north/south as predicted by the model coincide on the Earth’s surface. The general features shown by the maps in 2015 are well known (e.g., Finlay et al. 2010a) and have slowly evolved through the 115 years covered by IGRF-12. In particular, the minimum of magnetic intensity (see Figure 1 bottom), also known as the South Atlantic Anomaly, has continuously drifted westward and decreased since 1900. The point of minimum intensity at the Earth’s surface is currently over Southern Paraguay and is expected to cross the political boundary with Argentina during the second half of 2016. Maps of the predictive annual rate of change for D, I, and F between 2015 and 2020 at the Earth’s surface are shown in Figure 2. They are consistent with the continuation of the long-established westward drift and deepening of the South Atlantic Anomaly.

Figure 1
figure 1

Maps of the magnetic declination D (top, units are degrees), inclination I (middle, units are degrees), and total intensity F (bottom, units are nT) at the Earth’s mean radius r=a in 2015; the red dot indicates the minimum intensity. Projection is Mercator.

Full size image
Figure 2
figure 2

Maps of the predicted rate of change per year in the declination D (top, units are degrees/year), the inclination I (middle, units are degrees/year), and total intensity F (bottom, units are nT/year) at the Earth’s mean radius r=a for the interval 2015.0 to 2020.0. Projection is Mercator.

Full size image

The positions of the geomagnetic poles and the magnetic dip poles in the northern and southern hemispheres, tabulated in Table 4, are presented in Figure 3 on the Earth’s reference sphere. We recall that the geomagnetic poles are the points of intersection between the tilted axis of a central inclined magnetic dipole and the sphere of radius a=6,371.2 km. Their positions, expressed in the geocentric co-ordinate system, are antipodal and can be determined from only the three dipole (n=1) Gauss coefficients. The magnetic dip poles are defined as the points on the Earth’s surface where the magnetic field inclination, as determined from the entire field model to degree n=N, is vertical. They are referred to the north and south magnetic poles and are given in Table 4 for the field as observed in the geodetic WGS84 co-ordinate system. The comparison between the locations of the geomagnetic poles and the dip poles is of interest as, seen in the spherical frame, they would coincide if the Earth’s magnetic field was perfectly dipolar. However, this is not the case. The comparison also illustrates the comparatively slower drift in time of the Earth’s geomagnetic dipole compared to other contributions of the magnetic field. Interestingly, the movements of the north and south magnetic poles have not been erratic and have constantly moved northward since 1900. The tilt between the geomagnetic and the geographic axes is at present reducing with time; it is about 9.7 ∘ in 2015.0 and projected to be 9.4 ∘ in 2020. The north magnetic pole appeared to be accelerating rather smoothly over the last century (Figure 4) from about 5 to about 50 km/year with an increased acceleration around 1990 (Chulliat et al. 2010). The peculiar acceleration of the north and south magnetic poles between 1945 and 1955 as calculated by IGRF should be regarded with caution; see Xu (2000) for a discussion. Perhaps the most striking feature of IGRF-12 is that the north magnetic pole appears to have started a phase of deceleration with a velocity of about 53.2 km/year in 2015 and a projected velocity of 42.6 km/year in 2020. Note however that the later estimate relies on the predictive (SV) part of IGRF-12 for epoch 2015.0 to 2020.0 and that retrospective analysis has shown that errors could be significant (e.g., Finlay et al. 2010b). The locations computed from models are also intrinsically approximate due to the limited spatial resolution of the IGRF-12 models. For further details on the limitations of the IGRF for various applications and on difficulties in estimating its accuracy, readers should refer to Lowes (2000) or consult the IGRF ‘Health Warning’ found at http://www.ngdc.noaa.gov/IAGA/vmod/igrfhw.html.

Figure 3
figure 3

Motion of the magnetic dip pole (red) and geomagnetic pole (blue) since 1900 from IGRF-12 in the northern hemisphere (left) and the southern hemisphere (right). Stereographic projection is employed. The scale bar gives an indication of distance on the WGS84 ellipsoid that is correct along lines of constant longitude and also along the middle lines of latitude shown.

Full size image
Figure 4
figure 4

The northward velocity of the geomagnetic dip poles in the northern (purple dots) and southern (orange crosses) hemisphere as estimated by IGRF-12 on the WGS84 spheroid.

Full size image
Table 4 Magnetic pole positions since 1900 as determined from IGRF-12 in WGS84 geodetic latitude
Full size table

IGRF-12 online data products

Further general information about the IGRF: http://www.ngdc.noaa.gov/IAGA/vmod/igrf.html.

The coefficients of IGRF-12 in various file formats: http://www.ngdc.noaa.gov/IAGA/vmod/igrf12coeffs.txt

Fortran software for synthesizing the field from the coefficients: http://www.ngdc.noaa.gov/IAGA/vmod/igrf12.f

C software for synthesizing the field from the coefficients (Linux): http://www.ngdc.noaa.gov/IAGA/vmod/geomag70_linux.tar.gz

C software for synthesizing the field from the coefficients (Windows): http://www.ngdc.noaa.gov/IAGA/vmod/geomag70_windows.zip

Online computation of field components from the IGRF-12 model: http://www.ngdc.noaa.gov/geomag-web/?model=igrf http://www.geomag.bgs.ac.uk/data_service/models_compass/igrf_form.shtml http://wdc.kugi.kyoto-u.ac.jp/igrf/point/index.html

Archive of legacy versions of the IGRF model:http://www.ngdc.noaa.gov/IAGA/vmod/igrf_old_models.html

Appendix: World Data Centers

WORLD DATA SERVICE FOR GEOPHYSICS, BOULDERNOAA National Centers for Environmental Information, NOAA, 325 Broadway, E/GC, Boulder, CO 80305-3328UNITED STATES OF AMERICAINTERNET: http://www.ngdc.noaa.govWORLD DATA CENTRE FOR GEOMAGNETISM, COPENHAGENDTU Space, Diplomvej, Building 327, DK 2800, Kgs. Lynbgy, DENMARKTEL: +45 4525 9713FAX: +45 353 62475EMAIL: cfinlay@space.dtu.dkINTERNET: http://www.space.dtu.dk/English/Research/Scientific_data_and_models

WORLD DATA CENTRE FOR GEOMAGNETISM, EDINBURGHBritish Geological SurveyMurchison House, West Mains Road Edinburgh, EH9 3LAUNITED KINGDOM TEL: +44 131 650 0234FAX: +44 131 668 4368EMAIL: wdcgeomag@bgs.ac.ukINTERNET: http://www.wdc.bgs.ac.uk/

WORLD DATA CENTRE FOR GEOMAGNETISM, KYOTOData Analysis Center for Geomagnetism and SpaceMagnetism Graduate School of Science, Kyoto UniversityKitashirakawa-Oiwake Cho, Sakyo-kuKyoto, 606-8502, JAPANTEL: +81 75 753 3929FAX: +81 75 722 7884EMAIL: iyemori@kugi.kyoto-u.ac.jpINTERNET: http://wdc.kugi.kyoto-u.ac.jp

WORLD DATA CENTRE FOR GEOMAGNETISM, MUMBAIIndian Institute of GeomagnetismColaba, Mumbai, 400 005, INDIATEL: +91 22 215 0293FAX: +91 22 218 9568EMAIL: abh@iigs.iigm.res.inINTERNET: http://iigm.res.in