planetMagFields

Routines to easily access information about magnetic fields of planets in our solar system and visualize them in both 2D and 3D. These require NumPy, Matplotlib and SciPy (pronounced "Sigh Pie"). Other than that, the following external libraries are used for a few different functions:

• 2D plotting for map projections other than Hammer : Cartopy library ( see more under Projections )
• Potential extrapolation: SHTns library
• Writing vts files for 3D visualisation: PyEVTK library

The planet class

This gives access to all the relevant properties of a planet and has methods to plot the field and write a vts file for 3D visualization. Usage:

from planetmagfields import *
p = planet(name='earth',datDir='planetmagfields/data/')

This displays the some information about the planet

Planet: Earth
l_max = 13
Dipole tilt (degrees) = -9.410531

and gives access to variables associated with the planet such as:

• p.lmax : maximum spherical harmonic degree till which data is available
• p.glm, p.hlm: the Gauss coefficients
• p.Br : computed radial magnetic field at surface
• p.dipTheta : dipole tilt with respect to the rotation axis
• p.dipPhi : dipole longitude ( in case zero longitude is known, applicable to Earth )
• p.idx : indices to get values of Gauss coefficients

Example:

In [1]: from planetmagfields import *

In [2]: p = planet(name='jupiter')
Planet: Jupiter
l_max = 10
Dipole tilt (degrees) = 10.307870

In [3]: p.glm[p.idx[2,0]]      # g20
Out[3]: 11670.4

In [4]: p.hlm[p.idx[4,2]]      # h42
Out[4]: 27811.2

as well as the functions:

planet.plot()

This function plots a 2D surface plot of the radial magnetic field at radius r given in terms of the surface radius. For example,

from planetmagfields import *
p = planet(name='earth',datDir='planetmagfields/data/')
p.plot(r=1,proj='Mollweide')

produces the info mentioned above first and then the following plot of Earth's magnetic field using a Mollweide projection

while

from planetmagfields import *
p = planet(name='jupiter',datDir='planetmagfields/data/')
p.plot(r=0.85,proj='Mollweide')

produces the following info about Jupiter and then plot that follows

Planet: Jupiter
l_max = 10
Dipole tilt (degrees) = 10.307870

This can be compared with Fig. 1 g from Moore et al. 2018

planet.spec()

This function computes the Lowes spectrum of a planet at a given radius. It adds an array p.emag_spec which contains the energy at different spherical harmonic degrees and two variables p.dipolarity and p.dip_tot which provide the fraction of energies in the axial dipole and the total dipole with respect to the total energy at all degrees. Usage example:

from planetmagfields import *
p = planet(name='jupiter')
p.spec()

will provide variables

In [8]: p.dipolarity
Out[8]: 0.7472047129875864

In [9]: p.dip_tot
Out[9]: 0.7719205112704368

In [10]: p.emag_spec
Out[10]:
array([0.00000000e+00, 3.47735422e+11, 2.36340423e+10, 2.12851278e+10,
       1.75661779e+10, 1.92219842e+10, 9.91200756e+09, 3.34535475e+09,
       3.95317971e+09, 2.59333412e+09, 1.23423769e+09])

and will produce Jupiter's surface spectrum:

The plotting can be suppressed setting the logical p.spec(iplot=False).

planet.writeVtsFile()

This function writes a vts file that can be used to produce 3D visualizations of field lines with Paraview/VisIt. Usage:

p.writeVtsFile(potExtra=True, ratio_out=2, nrout=32)

where,

• potExtra : bool, whether to use potential extrapolation
• ratio_out: float, radius till which the field would be extrapolated in terms of the surface radius
• nrout: radial resolution for extrapolation

Example of a 3D image produced using Paraview for Neptune's field, extrapolated till 5 times the surface radius.

Field filtering using planet.plot_filt

The planet class also provides a function for producing a filtered view of the radial magnetic field using the function plot_filt. This function can take in either an arbitrary array of spherical harmonic degrees and orders or cut-off values. This is illustrated below with examples, assuming the user is in the repository directory.

Saturn's small-scale magnetic field at a depth of 0.75 planetary radius ( degree > 3 )

from planetmagfields import *
p = planet(name='saturn')
p.plot_filt(r=0.75,lCutMin=4,proj='Mollweide')

Compare this with Fig. 20 B from Cao et al. 2020

Jupiter's surface field restricted to degrees 1,2,3 and order 3

from planetmagfields import *
p = planet(name='jupiter')
p.plot_filt(r=1,larr=[1,2,3],marr=[3],proj='Mollweide')

Earth's smaller scale surface field for degree > 4 and order > 3

from planetmagfields import *
p = planet(name='earth')
p.plot_filt(r=1,lCutMin=5,mmin=4,proj='Mollweide')

Potential extrapolation

The potextra module provides a method for potential extrapolation of a planet's magnetic field. This uses the SHTns library for spherical harmonic transforms. Usage example:

import numpy as np
from planetmagfields import *
p = planet('saturn')
ratio_out = 5 # Ratio (> 1) in terms of surface radius to which to extrapolate
nrout = 32 # Number of grid points in radius between 1 and ratio_out
rout = np.linspace(1,ratio_out,nrout)
brout, btout, bpout = potextra.extrapot(p.lmax,1.,p.Br,rout)

Quickplot using the magField.py script:

$ ./magField.py <planet> <radius> <projection>

This will plot the radial magnetic field of a planet (any of the names from the list below, case insensitive) at a radius given in terms of the surface radius. The default is the surface field. For example,

$ ./magField.py earth Mollweide

displays the same information as above about Earth's field and produces the surface field of Earth while

$ ./magField.py jupiter 0.85 Mollweide

produces the same plot of Jupiter's field as shown before.

$ ./magField.py all <radius> <projection>

would produce a table of information about dipole tilt for each planet and magnetic field maps of all different planets at the given radius in a single figure.

For example:

$ ./magField all 0.9 Mollweide

would give

|=========|======|=======|
|Planet   | Theta| Phi   |
|=========|======|=======|
|Mercury  | 0.0  | 0.0   |
|Earth    | -9.4 | -72.7 |
|Jupiter  | 10.3 | -16.6 |
|Saturn   | 0.0  | 0.0   |
|Uranus   | 58.6 | -53.6 |
|Neptune  | 46.9 | -72.0 |
|Ganymede | -4.2 | 25.5  |
|---------|------|-------|

followed by the following plot

Spherical harmonic normalization and Condon-Shortley phase

All the Gauss coefficients in the collected data are Schmidt semi-normalized. Only the data for Earth uses a Condon-Shortley phase, the others do not.

Projections

By default, the plot function tries to use the Mollweide projection. However, using the power of the cartopy library, any projection from this list is supported. In the absence of the cartopy library, the 2D plots fall back to the internally written Hammer projection. Examples of Jupiter's radial magnetic field at r=0.85 with different projections are shown below:

In [1]: from planetmagfields import *

In [2]: p = planet(name='jupiter')
Planet: Jupiter
l_max = 10
Dipole tilt (degrees) = 10.307870

In [3]: p.plot(r=0.85)

In [4]: projlist=['Mercator','Robinson','Stereographic','AzimuthalEquidistant']

In [5]: for k,proj in enumerate(projlist):
   ...:     p.plot(r=0.85,proj=proj)
   ...:     savefig('jup_r0_85'+proj+'.png',dpi=150,bbox_inches='tight')
   ...:     close()

In[3] produces the plot of Jupiter's field already shown above. In[5] produces the following figures with the Mercator, Robinson, Stereographic and azimuthal equidistant projections, respectively.

This also works with the magField.py script for quick plotting. Examples:

./magField.py earth 0.9 Robinson

or even with plots of all planets together

./magField.py all 0.9 Robinson

Note that I have chosen to keep the projection information out of the plot titles to prevent too much text.

Data sources

Mercury : Anderson et al. 2012

Earth : IGRF 13, Alken et al. 2021

Jupiter : JRM09, Connerney et al. 2018

Saturn : Cassini 11+, Cao et al. 2020

Uranus : Connerny et al. 1987

Neptune : Connerny et al. 1991

Ganymede: Kivelson et al. 2002

Acknowledgements

I would like to thank Regupathi Angappan for motivating me to create this package. I thank Jon Aurnou for testing it out and pointing out runtime errors. Lastly, I would like to thank Thomas Gastine for comparing the plots with real data and pointing out a normalization error which has been fixed.