Under consideration for publication in J. Plasma Phys.

Relative field-line helicity in bounded

arXiv:1811.02306v1 [astro-ph.SR] 6 Nov 2018

domains
Anthony R. Yeates†, Marcus H. Page
Department of Mathematical Sciences, Durham University, South Road, Durham DH1 3LE,
UK

(Received xx; revised xx; accepted xx)

Models for astrophysical plasmas often have magnetic field lines that leave the boundary
rather than closing within the computational domain. Thus, the relative magnetic helicity
is frequently used in place of the usual magnetic helicity, so as to restore gauge invariance.
We show how to decompose the relative helicity into a relative field-line helicity that
is an ideal-magnetohydrodynamic invariant for each individual magnetic field line, and
vanishes along any field line where the original field matches the reference field. Physically,
this relative field-line helicity is a magnetic flux, whose specific definition depends on
the gauge of the reference vector potential on the boundary. We propose a particular
“minimal” gauge that depends only on the reference field and minimises this boundary
contribution, so as to reveal topological information about the original magnetic field.
We illustrate the effect of different gauge choices using the Low-Lou and Titov-Démoulin
models of solar active regions. Our numerical code to compute appropriate vector
potentials and relative field-line helicity in Cartesian domains is open source and freely
available.

1. Introduction
Magnetic helicity, h(V ), has been known to be an integral invariant of ideal magne-
tohydrodynamics (MHD) since Woltjer (1958), and its elegant physical interpretation
as the average pairwise linking of magnetic flux tubes (or magnetic field lines) since
Moffatt (1969). Its practical importance is largely due to its robustness even in the
presence of finite resistivity (see, e.g., Berger 1984; Browning 1988). However, the linking
interpretation requires the field lines to be either closed curves, or to be ergodic within
the domain (Arnold 1986). In astrophysical situations, the magnetic field is typically not
confined to a finite volume, so that field lines will typically leave through the boundary
of any finite computational domain.
The seminal work of Berger & Field (1984) showed how to define a gauge-invariant
version of h(V ) even when the domain is not magnetically closed, by subtracting the
helicity of an appropriate reference magnetic field. This so-called relative magnetic
helicity, HR , effectively measures the average linking with respect to that of the reference
field. Typically one chooses a reference field with as little self-linking as possible, so that
the desired information about the structure of the original field itself is revealed. Usually,
this is done by fixing the reference field to the unique current-free/potential magnetic
field that minimises magnetic energy for the given boundary conditions. Like h(V ), the
relative helicity HR is invariant under an ideal-MHD evolution that vanishes on the
boundary. In practice, we are often interested in the injection of helicity due to ideal (or
nearly ideal) motions on the boundary.
† Email address for correspondence: [email protected]
2 A. R. Yeates and M. H. Page
The relative helicity, HR , was motivated by applications in the solar atmosphere. Here,
the magnetic field in three-dimensional space cannot generally be measured directly from
telescope observations, but it is routinely observed on the Sun’s photosphere (visible
surface). Following Berger & Field (1984), these observations have been used to estimate
the flux of relative helicity from the Sun’s interior into its atmosphere, both on a global
scale (Berger & Ruzmaikin 2000; Yang & Zhang 2012), and in individual solar active
regions (see Démoulin & Pariat 2009, for a review). Indeed, the need to shed relative
helicity into interplanetary space is understood to be the fundamental driver for coronal
mass ejections (Rust 1997). In the case of numerical models and simulations, it is possible
to perform the volume integral for HR directly. Accordingly, its build-up in the solar
atmosphere has been followed in a wide variety of simulations of solar active regions
(e.g., DeVore 2000; Cheung et al. 2005; Magara 2008; Mackay et al. 2011; Yang et al.
2013; Moraitis et al. 2014; Sturrock et al. 2015; Pariat et al. 2015, 2017; Yardley et al.
2018). Minimisation of magnetic energy subject to conservation of HR can also be used for
the computation of force-free equilibria (Finn & Antonsen Jr. 1985; Dixon et al. 1989).
However, HR , like h(V ), is only a single integral over the whole domain, so cannot
provide local information about the constraints on a particular magnetic field. For
example, observations of the solar atmosphere suggest a highly non-uniform distribution
of free energy, with the concentration of helicity in individual magnetic flux ropes believed
to play an important role in their eruptivity. Moreover, recent numerical simulations of
the resistive relaxation of braided magnetic flux tubes suggest the presence of additional
topological constraints, over and above conservation of the global HR (Del Sordo et al.
2010; Pontin et al. 2011).
To overcome this limitation, it is desirable to compute HR in smaller subvolumes of
the overall computational domain. This was addressed by Longcope & Malanushenko
(2008),
S who divide the overall computational volume V into a union of subvolumes V =
D
i i . The subvolumes Di must be defined in such a way that the boundaries between
them are magnetic surfaces (with vanishing normal magnetic field). In this way, the
only uncontained magnetic field is on their intersection with the global boundary ∂V .
(s)
Longcope & Malanushenko (2008) propose to define an “additive self-helicity”, Hi , for
each Di that is simply the relative helicity integrated only over Di , with respect to
a reference potential field defined locally within Di . In this way, we can measure the
helicity over-and-above any helicity which is forced to be present simply by the shape of
(s)
the domain Di . Malanushenko et al. (2009) were able to relate Hi to the kink-instability
of a particular magnetic flux rope.
(s)
However, a limitation with Hi as defined by Longcope & Malanushenko (2008) is
the need to identify a suitable union of subvolumes Di . In the absence of any further
information, it seems desirable to make the decomposition very fine (i.e., many Di ).
The finest possible decomposition is to take each Di to be an infinitesimal tube around
each magnetic field line, but in that limit the local reference potential field would simply
(s)
become the original magnetic field, so that Hi (when appropriately normalised) would
vanish. Therefore, in this paper, we will take a different approach that allows us to define
a limiting field-line helicity for each individual magnetic field line. We will call this a
relative field-line helicity because it will integrate to give HR . Note that this property is
(s)
not shared by the Hi ; although they add to give a relative helicity, it is relative to the
sum of the local reference fields rather than to the global reference field used in HR . A
(s)
further advantage of our proposed relative field-line helicity over Hi will be its ease of
computation, since it will not require computation of the potential reference field on an
irregularly shaped domain.
 Relative field-line helicity 3
The basic form of field-line helicity, which decomposes the original magnetic helicity
h(V ), was given explicitly by Berger (1988) and will be defined in Section 2. Like h(V ),
field-line helicity has been shown to be intimately related to the topological structure of
a magnetic field (Yeates & Hornig 2013, 2014; Prior & Yeates 2014), and is a valuable
tool for studying magnetic reconnection (Yeates & Hornig 2011; Russell et al. 2015).
In global models of non-potential magnetic fields in the solar corona, it has recently
been used successfully to identify local magnetic flux ropes (Yeates & Hornig 2016;
Lowder & Yeates 2017). The main objective of the present paper is to generalise the
definition of field-line helicity to give a relative field-line helicity that integrates to give
HR , rather than h(V ). Our proposed definition is given in Section 3. Numerical methods
for computing relative field-line helicity are summarised in Section 4, and used to explore
and illustrate its behaviour for particular examples in Section 5. Although these examples
by no means exhaust the possible situations where these ideas can be applied, they do
illustrate some of the more typical structures that are found in solar active regions.

2. Preliminaries
Throughout this paper, we will assume our magnetic field B (with ∇ · B = 0) to be
defined in a simply-connected closed domain V whose boundary ∂V is a single closed
surface. Our examples in Section 5 will use a simple Cartesian box, which is the typical
V we have in mind. However, it is only the topology of V (and ∂V ) that matters for the
theory in Section 3.

2.1. Relative magnetic helicity

Importantly, we do not require ∂V to be a magnetic surface, but permit non-zero
Bn := B · n̂ on ∂V . This is the typical situation in astrophysical applications. As is well
known, the magnetic helicity,
Z
h(V ) = A · B d3 x, (2.1)
V

is not gauge invariant in this situation, since under a gauge transformation from A to
A′ = A + ∇χ, the helicity becomes
Z I
′ 3
h (V ) = A·Bd x+ χBn d2 x, (2.2)
V ∂V

so that its value depends on the gauge function χ. A gauge-invariant “relative helicity”
may be defined as
Z
A + Ap · B − B p d3 x,



HR = (2.3)
V
where B p is some arbitrary reference field whose normal component Bpn matches Bn on
∂V (Berger & Field 1984; Finn & Antonsen Jr. 1985). Typically B p is chosen to be the
unique such field that is current-free (potential) within V , although this is not essential.
In (2.3), A is an arbitrary vector potential for B (so that B = ∇ × A) and Ap is an
arbitrary vector potential for B p (so that B p = ∇ × Ap ). It is easy to show from (2.3)
that HR is invariant under gauge transformation of either A or Ap , for a given B p .

2.2. Field-line helicity

The concept of field-line helicity arises if we consider the magnetic helicity in a
subvolume Di ∈ V . This clearly makes sense for any material (comoving) subvolume
4 A. R. Yeates and M. H. Page

Figure 1. Definition sketch for Equation (2.5), showing a thin flux tube Vε (x) of radius ε and
magnetic flux Φε (x) around the field line L(x) that is rooted at x ∈ ∂V .

Di provided that Bn = 0 on ∂Di , for then

Z
h(Di ) = A · B d3 x (2.4)
Di

would be both gauge invariant and conserved under ideal-MHD evolution. One can
envisage sub-dividing V into a finer and finer union of material subvolumes, in order to
gain more and more detailed information about the topological structure. As mentioned
in Section 1, the finest possible decomposition is where each Di is a thin magnetic flux
tube surrounding a single magnetic field line, because magnetic field lines are material
lines in an ideal evolution.
In this paper, however, we will consider only field lines that are rooted at both ends in
the boundary ∂V , so that their surrounding tubular domains have Bn 6= 0 on the ends,
violating this boundary condition. Figure 1 shows such a field line, L(x), along with a
surrounding flux tube Vǫ (x) of radius ε (in some cross-section). We label L(x) by one of
its endpoints x ∈ ∂V . Throughout this paper, x will exclusively denote a point on the
boundary ∂V .
Because Bn 6= 0 on ∂Vǫ (x), we expect h(Vε (x)) to change due to either motion of
the field line endpoints or gauge transformation of A. The latter is a more significant
problem, and will be addressed in detail in this paper. First, however, we take the limit
to an infinitesimally thin tube. Since h(Vε (x)) is a volume integral, we normalise by the
flux Φε (x) of the tube and define
h(Vε (x))
A(x) = lim . (2.5)
ε→0 Φε (x)

In the limit ε → 0, the tube collapses to the line L(x), and A(x) tends to a well-defined
limit – independent of the choice of cross-section – which Berger (1988) calls the field-line
helicity. Note that, although A(x) is defined by integrating only over the tubular volume
Vε (x), it nevertheless contains information about how this particular flux tube interacts
with the magnetic field outside, due to the fact that A is defined globally. In this paper,
we will assume that B has no ergodic field lines with infinite length, as we have not
shown that a well-defined limit exists in that case.
The definition (2.5) shows that A(x) has the dimensions of a magnetic flux. A more
convenient formula may be obtained by writing the volume integral h(Vε (x)) as an
 Relative field-line helicity 5
integral along L(x) of cross-sectional surface integrals. In the limit Vε (x) → L(x), the
vector potential A is constant on each cross-section, so that
R
V (x) Φε (x)A · dl
Z
A(x) = lim ε = A · dl. (2.6)
ε→0 Φε (x) L(x)

This formula shows clearly that, were L(x) to be a contractible closed loop, then A(x)
would be gauge independent, and by Stokes’ theorem it would simply be the magnetic
flux linked through that loop. Clearly A(x) would then be an ideal-MHD invariant,
representing the magnetic flux linked with L(x). In our case, L(x) is not a closed
loop. Nevertheless, A(x) remains an ideal-MHD invariant provided that there are no
boundary motions, and also that the gauge of A on the boundary is fixed. Indeed a
gauge transformation from A to A′ = A + ∇χ will change the value of A(x) to
A′ (x) = A(x) + χ(x+ ) − χ(x), (2.7)
where x+ is the other end of the field line, as in Figure 1. This is simply the limiting
version of the formula (2.2). In fact, since ∂V is a single closed surface, it follows that
A(x), like h(V ), depends only on n̂ × A. For if n̂ × A′ = n̂ × A on ∂V , then n̂ × ∇χ = 0
on ∂V , so that χ is constant over all of ∂V and A′ (x) = A(x). This would not be true
for a domain where ∂V is not a single connected surface, for example a spherical shell.
As we will see in this paper, the art of working with field-line helicity is to choose an
informative gauge for n̂ × A.
Finally, we note the relation between A(x) and the overall helicity h(V ). When the
field lines L(x) partition the whole volume (i.e., there are no closed or ergodic field lines),
integrating (2.5) over all field lines, weighted by magnetic flux, will give h(V ) (Berger
1988). In other words,
1
I
A|Bn | d2 x = h(V ), (2.8)
2 ∂V
where the factor half arises because each field line has two end-points on ∂V .

3. Definition of relative field-line helicity

The aim of this paper is to generalise the formula (2.8) to the relative helicity HR . In
other words, we would like to define a “relative field-line helicity” AR (x) for each field
line that is invariant under an ideal-MHD evolution and satisfies
1
I
AR |Bn | d2 x = HR , (3.1)
2 ∂V
for a fixed choice of reference field. We will firstly show that there are many different
ways to define such an AR , and will go on to make some specific choices that we believe
are physically reasonable.

3.1. Towards a definition

One option is to use A(x) directly, but restrict the gauge A. We know that, whatever
the gauge, A(x) is invariant under an ideal-MHD evolution for every field line provided
that n̂ × A on ∂V remains fixed in time. But in an arbitrary gauge, condition (3.1) will
not be satisfied, since h(V ) 6= HR in general. Nevertheless, there are a restricted family
of gauges of A where h(V ) = HR . To see this, write (2.3) in the form
Z I Z
HR = A · B d3 x + A × Ap · n̂ d2 x − Ap · B p d3 x. (3.2)
V ∂V V
6 A. R. Yeates and M. H. Page
It is well known that the boundary term vanishes if we restrict n̂ × A = n̂ × Ap on ∂V
(Barnes 1988; Berger 1988). This is always possible since Bpn = Bn on ∂V (Section 4.3
shows how to construct such an A, given an arbitrary Ap ). We then need only choose
Ap such that V Ap · B p d3 x = 0, and we will have HR = h(V ) and hence (3.1). In
R

general, this is not

H possible, since B p may have a closed-loop field line L within V whose
gauge-invariant L Ap · dl does not vanish. However, in the case where B p is a potential
field, it cannot have any closed-loop field lines, and we can always find a gauge for Ap
such that V Ap · B p d3 x = 0 (e.g., Barnes 1988; BergerR1988).
R

Even when B p is a potential field, the conditions that V Ap · B p d3 x = 0 and n̂× A =

n̂ × Ap on ∂V do not uniquely define A(x), because there is still freedom in the choice
of Ap . For example, if we have such a gauge Ap and change to A′p = Ap + ∇φ, then
Z I
A′p · B p d3 x = φBn d2 x. (3.3)
V ∂V

Provided Bn 6= 0, this integral may still vanish even if φ 6= 0 on ∂V .

3.2. General definition

In practice, it may beR inconvenient or (when B p is non-potential) impossible to find
a gauge Ap such that V Ap · B p d3 x = 0. Therefore, we propose instead an alternative
definition for relative field-line helicity that does not impose this requirement on Ap .
Namely,
AR (x) := A(x) − Ap (x), (3.4)
where A(x) is the usual field-line helicity (2.6) and
Z
Ap (x) = Ap · dl, (3.5)
Lp (x)

which is the field-line helicity of the reference field B p on its own field line rooted at the
same point x ∈ ∂V . Importantly, our definition assumes the gauge condition on A that
n̂ × A = n̂ × Ap on ∂V . (3.6)
It follows that AR (x) will satisfy (3.1), whatever the gauge of Ap . Note that, unlike
A(x), this AR (x) will not, in general, have the same value at both ends of a given field
line L(x), since the corresponding reference field line Lp (x) at each end will be different.
A nice property of AR (x) is that AR (x) = 0 whenever B p = B all along L(x) for
some x ∈ ∂V . For in that case, L(x) = Lp (x) and we must have A = Ap + ∇χ along
this line. Then Z Z
AR (x) = (A − Ap ) · dl = ∇χ · dl. (3.7)
L(x) L(x)
However, since ∂V is a connected surface, and n̂ × ∇χ = 0 on this surface thanks to our
gauge restriction, it follows that χ is constant on ∂V and hence AR (x) = 0. This property
makes AR useful for identifying non-potential regions within B, and R will not generally
be shared by the basic field-line helicity A, even in a gauge where V Ap · B p d3 x = 0
overall.

3.3. Gauge dependence and physical meaning

Unfortunately, our definition of AR in (3.4), even with the restriction that n̂ × A =
n̂ × Ap on ∂V , is still not uniquely defined, and depends on the gauge of Ap . To see this,
consider a gauge transformation from Ap to A′p = Ap + ∇φ. To preserve the boundary
 Relative field-line helicity 7

(a) (b)

Figure 2. Interpretation of AR (x) as a magnetic flux, for two different gauges in (a) and (b).
The field lines L(x) and Lp (x) are the same in each case, but the possible curves γ and γ ′
that complete the loop by linking x+ and xp+ on ∂V differ between the two gauges, so that
AR (x) 6= AR ′ (x).

restriction, we must also change the gauge of A to A′ = A + ∇χ, where n̂ × ∇(φ − χ) = 0

on ∂V . It follows that the two gauge functions can differ on ∂V only by a global constant
(since ∂V is connected). Using this fact, we find that AR (x) goes to
Z Z
AR ′ (x) = AR (x) + ∇χ · dl − ∇φ · dl, (3.8)
L(x) Lp (x)
= AR (x) + φ(x+ ) − φ(xp+ ). (3.9)

When L(x) 6= Lp (x), the other ends of the field lines (denoted x+ and xp+ ) will differ,
so the value of AR (x) will change.
To see the physical meaning of this gauge dependence, we can interpret AR (x) as a
magnetic flux (cf. Yeates & Hornig 2016). This is illustrated in Figure 2, where L(x) is
a field line of B and Lp (x) is a field line of B p rooted at the same point x. The curve
γ ⊂ ∂V in Figure 2(a) closes the loop R and defines a surface in V bounded by L(x), γ,
and −Lp (x). If γ is chosen so that γ Ap · dl = 0, then AR (x) will be precisely the
magnetic flux through this surface, by Stokes’ theorem. Provided the field-line endpoints
and n̂ × Ap on ∂V and remain fixed in time, this flux will be an ideal-MHD invariant. In
any gauge, it is possible to find such a curve γ along which the line integral vanishes, as
proven by Yeates & Hornig (2016). In fact, there are an infinite number of such curves
(whose corresponding surfaces all have the same flux). The possible γ depend on the
gauge of n̂ × Ap , as sketched in Figure 2(b) which shows a different curve γ ′ arising in a
different gauge A′ . This loop encloses a different amount of magnetic flux, corresponding
to the different value of AR ′ (x) in this new gauge. In summary, the gauge dependence
of AR may be viewed simply as a choice of how to close flux surfaces on the boundary.
Whichever choice is made, AR (x) is an ideal-MHD invariant.
At first sight, it may seem from (3.9) that we can choose any distribution of AR that
we like on field lines that intersect ∂V , simply by changing the gauge φ. Indeed, this is
true. However, choosing φ in this way requires you to first know AR for B in some initial
gauge. If we choose the gauge a priori, based only on B p , then the resulting AR can give
us meaningful information about B.
In Section 5, we will illustrate the effect on AR of some different choices of Ap in
particular examples. However, we will first propose a general candidate for the “best”
choice of gauge, that can be used to uniquely define AR .
8 A. R. Yeates and M. H. Page
3.4. Minimal gauge
Since AR depends only on the distribution of n̂ × Ap on ∂V , our idea is to choose this
gauge by minimising the integral ∂V |n̂ × Ap |2 d2 x over the whole boundary. Intuitively,
H
this will give us the smallest overall boundary contribution to AR , for a given reference
field B p (whether potential or not). In Appendix A, we prove that this integral is
minimised if
∇h · Ap = 0 on ∂V , (3.10)
where ∇h · denotes the two-dimensional divergence in the plane of ∂V . This is the so-
called “universal gauge” condition of Hornig (2006). Although this condition does not
specify Ap uniquely within V , it does specify n̂×Ap uniquely on ∂V , and hence uniquely
specifies AR . To see this, suppose both Ap and A′p = Ap + ∇φ satisfy (3.10). Then we
must have ∇2h φ = 0, but since ∂V is a closed surface, this implies that φ is constant on
∂V , so that n̂ × A′p = n̂ × Ap . Note that Ap does not necessarily vanish in this gauge
when B p is a potential field – we will see this in Section 5.
In the rest of this paper, we will denote vector potentials satisfying (3.10) by A∗ or
∗
Ap , the corresponding field-line helicities by A∗ and A∗p , and the corresponding relative
field-line helicity by A∗R . In Section 5, we will compute A∗R for some example magnetic
fields, and compare it to AR in some other gauges.

3.5. Summary of the proposed definition

To summarise, let A∗p be a vector potential for B p satisfying

∇h · A∗p = 0 on ∂V , (3.11)

and let A∗ be a vector potential for B satisfying

n̂ × A∗ = n̂ × A∗p on ∂V . (3.12)

Then our proposed definition of relative field line helicity at x ∈ ∂V is the difference

A∗R (x) := A∗ (x) − A∗p (x), (3.13)

where A∗ , A∗p are the field-line helicities of B and B p in these two gauges, defined along
their respective magnetic field lines.

4. Numerical methods
The numerical code used for this paper was written in Python, with a sup-
porting Fortran module for fast tracing of the magnetic field lines (taking ad-
vantage of OpenMP if available). This code is open source and available at
https://github.com/antyeates1983/flhtools. Although the definitions in Section 3
apply to more general domains, this code is specific to a Cartesian domain.
For field-line tracing, the code uses the second-order midpoint method with adaptive
step-size. The computations in this paper used simple linear interpolation of A and B
to trace field lines and compute A, demonstrating that sophisticated numerical methods
are not required to work with field-line helicity. However, to facilitate computation of
B p , A and Ap , the code uses a staggered mesh where each component of B (or B p ) is
located at the centre of the corresponding cell face and the components of A (or Ap ) are
located on the corresponding cell edges (Yee 1966).
 Relative field-line helicity 9
4.1. Reference potential field
In a Cartesian box, the potential reference field B p that matches Bn on all six boundary
faces is straightforward to compute. We write B p = ∇u leading to the three-dimensional
Laplace equation ∇2 u = 0. On the staggered mesh, the potential u is located at the centre
of each three-dimensional cell, so that each component of B p may be computed from u
by a central difference. We approximate the Laplacian operator ∇2 with a second-order
(7-point) finite-difference stencil, for which the solution is efficiently and simply obtained
using a fast-Poisson technique with Neumann boundary conditions (Press et al. 1992).
With this technique, the current density ∇ × B p vanishes to machine precision when
computed from B p with central differences.

4.2. Minimal-gauge vector potential

∗
To find A (or A∗p ),
we have developed a numerical routine that starts from A (or Ap )
in an arbitrary gauge, and performs the necessary gauge transformation. In particular,
for the case of A, we seek a scalar function χ(x, y, z) on V such that A∗ = A + ∇χ.
The condition ∇h · A∗ = 0 on ∂V requires that χ := χ|∂V satisfy the two-dimensional
Poisson equation
∇2h χ = −∇h · A (4.1)
over this whole boundary ∂V . Since ∂VH is a closed surface, the right-hand side necessarily
satisfies the compatibility condition ∂V ∇h · A d2 x = 0, so that the solution for χ exists
and is unique up to an additive constant χ0 .
Equation (4.1) is solved numerically by a finite-difference method, approximating the
Laplacian operator with the standard 5-point stencil. For our Cartesian domain, care
must be taken to couple values of χ on neighbouring faces so that the normal component
of ∇h χ is continuous. (Strictly speaking, we are approximating a weak solution to (4.1),
since the differential operators in (4.1) are not defined on the edges of the cube.) This
coupling between faces precludes the use of a fast-Poisson solver, but a direct solution of
the resulting (sparse) linear system remains practical for this two-dimensional problem.
Although A∗ does not depend on the additive constant χ0 , we ensure that the linear
system is invertible by fixing χ = 0 at one vertex of ∂V .
Given the solution for χ, we can then choose any arbitrary extension of χ into the
interior of V , since this choice does not affect A∗ . It is simplest to take ∇2 χ = 0 in
the volume, and solve this three-dimensional Laplace equation with Dirichlet boundary
conditions χ = χ on ∂V . This can be solved using a standard fast-Poisson method.

4.3. Matching the reference gauge

Although not required for computing A∗R , we have also implemented a numerical
routine for changing the gauge of A to A′ so that n̂ × A′ = n̂ × Ap on ∂V , given
some Ap . This will be used in Section 5 to illustrate the computation of AR in gauges
other than the minimal gauge.
Writing A′ = A+∇χ, we need to compute χ(x, y, z) ∈ V such that n̂×A′ = n̂×Ap on
∂V . This may be done by first solving for χ|∂V = χ on ∂V , then extending to a solution
on the interior of V . We need χ to satisfy ∇h χ = Aph − Ah , so take the divergence and
solve the resulting two-dimensional Poisson equation
∇2h χ = ∇h · (Ap − A) (4.2)
on ∂V . As in Section 4.2, a solution exists because the compatibility condition is satisfied.
Since the Neumann boundary condition n̂·∇χ = n̂·(Ap −A) is fixed on the edges of each
face, it is possible to solve (4.2) separately on each face. This means that a fast-Poisson
10 A. R. Yeates and M. H. Page
solver can readily be used, unlike in Section 4.2. This gives χ on each face Si up to a
constant χi , for i = 1, . . . , 6. One of these constants may be chosen arbitrarily, and the
other five are then easily determined by imposing continuity of χ at each edge. (Although
there are 12 edges, the corresponding jumps in χ are not independent, so the problem is
well-posed.)
Once the continuous solution for χ is obtained, this may be extended to χ in the
interior of V in the same way as Section 4.2, namely solving the three-dimensional Laplace
equation ∇2 χ = 0 in V with Dirichlet boundary conditions χ = χ on ∂V . Again, this
choice of interior distribution has no effect on the field-line helicity.

5. Examples
To investigate the behaviour of AR , we consider two well-known solar active region
models from the literature. These exhibit important magnetic structures found in MHD
simulations, namely sheared magnetic arcades and magnetic flux ropes. They are chosen
specifically to enable validation against the results of Valori et al. (2016), who made a
careful study of the numerical computation of HR in these particular examples (among
others).

5.1. Low-Lou equilibrium

First, we consider one of the class of nonlinear force-free equilibria derived by
Low & Lou (1990), which is given analytically except for the solution of a single
ordinary differential equation. Our calculations start from a three-dimensional datacube
of B, computed by Valori et al. (2016) and kindly shared by these authors. This
corresponds to the specific solution with parameters n = 1, l = 0.3, φ = π/4, in a
Cartesian box V = [−1, 1] × [−1, 1] × [0, 2]. We used four datacubes with uniform mesh
spacing ∆x = 2/32, 2/64, 2/128, and 2/256, respectively. As described in Section 4, our
numerical code uses a staggered mesh, so we first interpolated the original B components
from cell vertices to the centres of the cell faces, using trilinear interpolation. We also
tested tricubic interpolation here but found it unnecessary.
Figure 3 shows the magnetic field lines for this magnetic field B (panel a) and for
the corresponding reference potential field B p (panel b). The latter was computed
numerically to match Bn on all six boundaries, using the method described in Section
4.1. In Figure 3, each field line L(x) is coloured by its integrated parallel current density,
defined as Z
Jk (x) = ∇ × B · dl (5.1)
L(x)
and normalised by B0 , the mean value of |Bz (x, y, 0)|. Maps of Jk /B0 for all field lines
seeded from the z = 0 boundary are shown in panels (c) and (d). In the Low-Lou
equilibrium, the electric currents are smoothly distributed within the arcade of magnetic
loops surrounding the central polarity-inversion line. For B p , we have ∇ × B p ≡ 0 and
hence Jk vanishes (to machine precision).
Next, we have computed maps of the field-line helicities A, Ap and the relative field-
line helicity AR := A − Ap using four different gauges. These include the minimal gauge
A∗p , but also three others, denoted Aap , Abp , Acp and defined as follows. All three satisfy
the DeVore (2000) condition that Az = 0. The first is integrated upward in from the
lower z boundary,
Z z
Aap (x, y, z) = Aa0 (x, y) − ẑ × B(x, y, z ′ ) dz ′ , (5.2)
0
 Relative field-line helicity 11
(a) B (b) Bp

(c) Jk /B0 (d) Jk /B0 (potential)

1.00

0.75 20 20

0.50
10 10
0.25

0.00 0 0
y

−0.25
−10 −10
−0.50

−0.75 −20 −20

−1.00
−1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0
x x

Figure 3. Field-line structure of the Low-Lou example B (a) and its corresponding potential
reference field B p (b), for the datacube with ∆x = 2/128. The field lines are coloured
purple-orange according to their integrated parallel current, shown in the z = 0 plane in panels
(c) and (d). On the lower boundary, the normal field Bz (x, y, 0) = Bpz (x, y, 0) is shown by
greyscale contours (a and b) or dashed lines (c and d).

and uses the Coulomb-gauge form Aa0 = ∇ × (Ψ ẑ) for the boundary term, with Ψ com-
puted by solving the Poisson equation ∇2h Ψ = −Bz (x, y, 0) (we separate the monopole
component and solve for the remainder using periodic boundary conditions). The second
gauge is similar but integrated down from the upper z boundary, so that
Z 2
b b
Ap (x, y, z) = A0 (x, y) + ẑ × B(x, y, z ′ ) dz ′ (5.3)
z

and Ab0is now computed on z = 2 from Bz (x, y, 2). The third is an alternative DeVore
gauge that still has the form
Z z
c c
Ap = A0 (x, y) − ẑ × B(x, y, z ′ ) dz ′ , (5.4)
0
but uses the alternative boundary term
Z y  Z x 
1 1
Ac0 = − Bz (x, y ′ , 0) dy ′ x̂ + Bz (x′ , y, 0) dx′ ŷ. (5.5)
2 −1 2 −1

There is nothing special about these chosen gauges, other than their simplicity of
computation. The latter has led to the use of Aap and Abp for computation of HR (e.g.
12 A. R. Yeates and M. H. Page
Valori et al. 2016), where, of course, the choice of gauge does not affect the result, except
for numerical error.
Figure 4 shows maps of A, Ap , and AR for each of the four gauges, computed at
resolution 256 × 256 from the datacube with resolution ∆x = 2/128. These maps show
only the z = 0 boundary. For the first three gauges, n̂ × A has been matched to n̂ × Ap
using the gauge-matching procedure described in Section 4. For A∗p , this is not necessary
as both A and Ap are independently matched to the minimal gauge using the procedure
in Section 4.2.
The first important observation is that all three distributions depend on the gauge
choice n̂ × Ap . In fact, the results from Aap and especially Abp are quite close to A∗p in
this case. This is because, for this specific B, the majority of |Bn | is located on the lower
boundary, where Aa0 and Ab0 satisfy the solenoidal condition ∇h · A = 0. The gauge
Acp (Figures 4g–i) lacks this property, and gives more radically different results from A∗p .
Notice that Acp leads to larger maximum values of |AR |, and that these are more localised
than either the regions of strongest |A∗R | or of strongest Jk (Figure 3c).
Notice that the field-line helicity of the reference field, Ap , is weaker than A in both Aa
and Ab , and even more so in A∗ . This is a good indication that these are “sensible” gauge
choices, since B p is the minimum-energy field and cannot have significantly twisted field
lines (it can, however, still exhibit topological structure due to the non-uniform boundary
conditions Bpn , cf. Prior & Yeates 2014). Similarly, although A∗ is primarily located
within the main current-carrying region (cf. Figure 3c), it does not completely vanish
outside of this region. The boundary of this current-carrying region appears as a sharp
separatrix between field lines with both ends on the lower boundary z = 0 and field lines
with only one end on this boundary. Notice that the location of this separatrix differs
between B and B p , so that the relative field-line helicity map AR inherits discontinuous
jumps from both A and Ap .
As a validation exercise, we have verified that the all of the AR in Figure 4 integrate to
give the same HR , as they should by definition. Figure 5 shows the resulting HR for each
of the four gauges, as a function of the mesh spacing ∆x for the four datacubes. Solid
curves show volume integrals of (A+ Ap )·(B − B p ), using the composite trapezium rule,
whereas dashed curves show estimates of HR from integrating AR using equation (3.1).
For the latter computations, field lines were traced from a mesh with spacing ∆x × ∆x
on each of the six boundary faces, and the integral then estimated with the composite
trapezium rule. It is clear from Figure 5 that taking a finer mesh leads to convergence
to a common value of HR between the different gauges, whether computed by volume
integration or from AR . Moreover, this common value is consistent with the computations
of Valori et al. (2016, see their Figure 7a).

5.2. Titov-Démoulin equilibria

More realistic models of non-potential solar active regions are given by the Titov-
Démoulin family of equilibria (Titov & Démoulin 1999), in which the free magnetic
energy is localised to a toroidal current channel contained within a surrounding potential
field. We utilise datacubes of four such equilibria computed and kindly shared by
Valori et al. (2016). For details of the specific construction of these equilibria, see that
paper. Here we just note that the domain is the Cartesian volume V = [−3.18, 3.18] ×
[−5.10, 5.10]×[0, 4.56], and that the primary difference between the four cases is the twist
of magnetic field lines in the current channel. The cases are denoted N = 0.1, 0.5, 1, 3,
where N is the (approximate) number of twists made by a field line along the current
channel. Corresponding magnetic field line plots are shown in Figure 6 (a, d, g, j). Note
 Relative field-line helicity 13
(a) Aa /Φ0 (b) Aap /Φ0 (c) AaR /Φ0
1.0
0.2 0.2
0.2
0.5
0.1 0.1 0.1

0.0 0.0 0.0 0.0

y

−0.1 −0.1 −0.1

−0.5
−0.2
−0.2 −0.2
−1.0

(d) Ab /Φ0 (e) Abp /Φ0 (f) AbR /Φ0

1.0
0.20
0.15 0.15
0.15
0.5 0.10 0.10 0.10
0.05 0.05 0.05
0.0 0.00 0.00 0.00
y

−0.05 −0.05 −0.05

−0.5 −0.10 −0.10 −0.10

−0.15
−0.15 −0.15
−0.20
−1.0

(g) Ac /Φ0 (h) Acp /Φ0 (i) AcR /Φ0

1.0
0.4 0.4
0.4
0.5
0.2 0.2
0.2

0.0 0.0 0.0 0.0

y

−0.2
−0.2 −0.2
−0.5
−0.4
−0.4 −0.4
−1.0

(j) A∗ /Φ0 (k) A∗p /Φ0 (l) A∗R /Φ0

1.0
0.2 0.2 0.2

0.5 0.1 0.1 0.1

0.0 0.0 0.0 0.0

y

−0.5 −0.1 −0.1 −0.1

−0.2 −0.2 −0.2

−1.0
−1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0
x x x

Figure 4. Field-line helicity on z = 0 for the Low-Lou example in four different gauges (top to
bottom). The left and middle panels show A for the original and potential fields, respectively,
while the right panel shows their difference AR in each case, which is the relative field-line
helicity. The definitions of the vector potentials are given in the text, but note that panels (j)
to (l) show the minimal gauge A∗ . Dashed lines show contours of BRz = Bpz for context. All of
the field-line helicities are normalised by the boundary flux Φ0 = 21 z=0 |Bn | d2 x.

that the boundary conditions Bn and potential magnetic energy also differ between the
cases. The relative free energies (ratio of non-potential to total magnetic energy) are
≈ 0.002, 0.03, 0.20, and 0.15, respectively. In all computations here, we use datacubes
with ∆x = 0.06.
14 A. R. Yeates and M. H. Page
−0.1300
Aa
−0.1325
Ab
−0.1350 Ac
A∗
−0.1375
HR /Φ20

−0.1400

−0.1425

−0.1450

−0.1475

−0.1500
0.01 0.02 0.03 0.04 0.05 0.06
∆x

Figure 5. Numerically estimated relative helicity HR for the Low-Lou example, as a function
of the numerical mesh spacing, using four different gauges as shown in the legend. In each case,
solid lines/circles show direct volume-integrated HR using the composite trapezium rule, while
dashed lines/pluses are computed from by integrating AR |Bn |/2 over all six boundary faces.
The value of HR is normalised by the square of the boundary flux Φ0 .

We begin our analysis with the same computation of A, Ap , and AR in four different
gauges as for the Low-Lou equilibrium. These are shown in Figure 7 for one of the
datacubes: N = 0.5. Once again, the results are clearly dependent on the gauge. However,
in this case, the downward DeVore-Coulomb gauge (Abp ) gives very similar results to the
minimal gauge (A∗p ), whereas the upward DeVore-Coulomb gauge (Aap ) gives rather
different results, along with Acp . In all four cases, A agrees well with Ap outside of
the toroidal current channel (which appears as two tear-drop shaped footpoints on this
boundary). However, for Abp and A∗p , the current channel has a much higher strength
of |A| than outside, whereas for Aap and Acp , comparable strengths of |A| are found at
locations both inside and outside the channel.
In this example, the relative field-line helicity A∗R (and also AbR ) is much more
concentrated than in the Low-Lou equilibrium. This holds for all N , as shown in Figure
6. It is strongest in the tear-drop shaped footpoints of the current channel, although
diffuse non-zero values are found throughout the “closed field” region (i.e., the region
between the diagonal separatrices within which field lines have both ends on z = 0 and
cross the central polarity inversion line). There are also thin ridges in A∗R along the
separatrices; these arise because the separatrices are in slightly different locations in B
and B p . Outside of these separatrices, A∗R is close to zero, reflecting the fact that these
field lines are close to potential. Figure 8 shows a comparison of field lines of B and B p
for two of the datacubes, showing this effect.
If we compare the maximum magnitude of |A∗R | as N is increased (Figure 6), we see that
the maximum |A∗R | (relative to Φ0 ) increases from N = 0.1 to 0.5 and again from N = 0.5
to 1, but is comparable between N = 1 and 3, despite the fact that the maximum Jk /B0
increases by a factor 4. This reflects the fact that A∗R is a global topological measure
that does not solely measure the local twist (or “self helicity”) of magnetic field lines.
However, it does highlight that |A∗R | will likely not relate directly to the stability of a
given configuration, since the N = 3 case is expected to be kink-unstable while the N = 1
(and N = 0.1, 0.5) is expected to be stable (Valori et al. 2016). The similar peak values
of |A∗R | located in a smaller footprint for N = 3 (as seen in Figure 6k) is consistent with
this case having a lower |HR | than N = 1, as found by Valori et al. (2016).
 Relative field-line helicity 15
(b) Jk /B0 (c) A∗R /Φ0

(a) N = 0.1 4 0.3

40
0.2
2 20
0.1

0 0 0.0

y
−0.1
−2 −20
−0.2
−40
−4 −0.3

(e) Jk /B0 (f) A∗R /Φ0

(d) N = 0.5 4 0.3

40
0.2
2 20
0.1

0 0 0.0
y

−0.1
−2 −20
−0.2
−40
−4 −0.3

(h) Jk /B0 (i) A∗R /Φ0

(g) N = 1 4 0.3
40
0.2
2 20
0.1

0 0 0.0
y

−0.1
−2 −20
−0.2
−40
−4 −0.3

(k) Jk /B0 (l) A∗R /Φ0

(j) N = 3 4 0.3
40
0.2
2 20
0.1

0 0 0.0
y

−0.1
−2 −20
−0.2
−40
−4 −0.3

−2 0 2 −2 0 2
x x

Figure 6. The Titov-Démoulin solution for increasing (top to bottom) values of the twist
parameter N , showing the magnetic field lines, field-line integrated parallel current Jk /B0 , and
relative field-line helicity A∗R in each case. The field lines are coloured by Jk , with the same
colour scale as the middle column. The colour scales are fixed; actual maxima of |Jk /B0 | are
4.6, 21.2, 53.3, 199.4, and of |A∗R /Φ0 | are 0.03, 0.13, 0.31, 0.32, respectively.
16 A. R. Yeates and M. H. Page
(a) Aa /Φ0 (b) Aap /Φ0 (c) AaR /Φ0

4
0.10 0.10 0.10

2 0.05 0.05 0.05

0 0.00 0.00 0.00

y

−2 −0.05 −0.05 −0.05

−0.10 −0.10 −0.10

−4

(d) Ab /Φ0 (e) Abp /Φ0 (f) AbR /Φ0

4
0.10 0.10 0.10
2
0.05 0.05 0.05

0 0.00 0.00 0.00

y

−0.05 −0.05 −0.05

−2

−0.10 −0.10 −0.10

−4

(g) Ac /Φ0 (h) Acp /Φ0 (i) AcR /Φ0

4 0.20 0.20
0.15 0.15 0.2
2 0.10 0.10
0.1
0.05 0.05
0 0.00 0.00 0.0
y

−0.05 −0.05
−0.1
−2 −0.10 −0.10
−0.15 −0.15 −0.2
−4 −0.20 −0.20

(j) A∗ /Φ0 (k) A∗p /Φ0 (l) A∗R /Φ0

4
0.10 0.10 0.10

2 0.05 0.05 0.05

0 0.00 0.00 0.00

y

−0.05 −0.05 −0.05

−2

−0.10 −0.10 −0.10

−4

−2.5 0.0 2.5 −2.5 0.0 2.5 −2.5 0.0 2.5

x x x

Figure 7. Field-line helicity on z = 0 for the N = 0.5 Titov-Démoulin model in four different
gauges (top to bottom). The format is the same as Figure 4.
 Relative field-line helicity 17
(a) N = 0.1 (a) N = 1

Figure 8. Field lines of B (purple) and B p (green), traced from the same starting points in
y > 0 (to left in this view), for the Titov-Démoulin examples with (a) N = 0.1 and (b) N = 1.

Finally, we repeat the validation exercise of computing HR , as for the Low-Lou model.
Figure 9(a) shows the results in the same format as Figure 5, except that the horizontal
axis now labels the cases with different N (measured by the end-to-end twist, to facilitate
comparison with Figure 3a of Valori et al. 2016). As for the Low-Lou example, we find
consistent results when HR is computed by volume integration. When computed from
AR via (3.1), we obtain consistent values for HR for N = 0.1, 0.5 and 3, but |HR | is
underestimated by up to 5% for N = 1, depending on the gauge. Further investigation
shows that this is not due to the mesh resolution (of either the datacube or the mesh for
computing AR on the boundary), but is caused by the fact that the initial data do not
perfectly satisfy the solenoidal condition on our staggered mesh. To show this, Figure
9(b) shows the same computations of HR when we include an additional correction to
B after computing the (initial) A on the staggered mesh. This correction recomputes
B from the numerical curl of A, so that the solenoidal condition is satisfied to machine
precision, but at the expense of changing B. With this correction, Figure 9(b) shows that
the estimate of HR from AR then agrees with the volume integral, in all four gauges.
Notice however that |HR | is then overestimated in Ab , compared with the other gauges.
This arises because the solenoidal correction has a greater effect on the current-carrying
region in the downward DeVore gauge, since this region is located at lower z. Thus
this affects the estimated helicity more. This illustrates the importance of the solenoidal
condition in estimating helicity (Valori et al. 2016).

6. Conclusion
We have shown that it is possible to decompose the relative magnetic helicity HR into
a (flux-weighted) integral (3.1) over a quantity AR called the relative field-line helicity.
This quantity AR (x) is itself an ideal-MHD invariant for each magnetic field line L(x).
Moreover, AR (x) = 0 whenever B = B p all along the field line L(x). We propose that
AR is a useful measure to identify locations of significant topological structure within a
wider magnetic field. Though HR is gauge independent, our suggested definition of AR
requires computation of a vector potential satisfying certain conditions, as summarised
in Section 3.5. Both the vector potential and AR are straightforward to compute in a
Cartesian domain, and we have provided free numerical code for doing so.
18 A. R. Yeates and M. H. Page
(a) Original (b) With divergence correction
0.00
Aa Aa
Ab Ab
−0.02 Ac Ac
A∗ A∗
HR /Φ20

−0.04

−0.06

−0.08

0 1 2 3 4 5 6 0 1 2 3 4 5 6
End-to-end twist End-to-end twist

Figure 9. Numerically estimated relative helicity HR for the Titov-Démoulin example, as a

function of the end-to-end twist (in radians). As in Figure 5, colours denote different gauges,
and solid lines/circles show direct volume-integrated HR using the composite trapezium rule,
while dashed lines/pluses are computed by integrating AR |Bn |/2 AR over all six boundary faces.
In panel (b), an additional divergence-correction step is applied, as described in the text.

The principle difficulty faced in this endeavour has been the fact that AR is not uniquely
defined by the requirements of (i) ideal invariance for every field line and (ii) vanishing
when B = B p all along a field line. Rather, AR still depends on the gauge of the reference
vector potential Ap . Physically, we showed that this gauge dependence arises from a
fundamental choice of how to define a corresponding surface for each field line. The value
of AR measures the ideal-invariant magnetic flux through this surface, but the definition
of the surface depends on where it intersects the boundary, which in turn depends on the
gauge of n̂×Ap on the boundary. The gauge-dependence integrates out when integrating
over all field lines to give HR , but appears to be unavoidable when considering field-line
helicity alone. Nevertheless, this does not mean that AR is physically irrelevant or useless.
Rather, it means that one must choose the gauge so that the corresponding magnetic
fluxes measured by AR are informative (cf. Yeates & Hornig 2016).
Prior & Yeates (2014) discuss the physical meaning of the original h(V ) in different
gauges, for the particular case of a magnetic field between two planes with Bn = 0 on
the side boundaries. They show that it always measures the average pairwise winding
number between two curves, but with respect to a reference frame that varies depending
on the gauge. They proposed the best choice of gauge to be the so-called winding gauge,
AW , in which h(V ) measures the average winding number with respect to an untwisted
Cartesian reference frame (corresponding to physical space). Here, we allow for more
general magnetic fields which may enter or leave the side boundaries of the domain, so
the winding gauge of Prior & Yeates (2014) does not apply. Instead, we have proposed
to use a “minimal gauge” A∗ , also chosen by Hornig (2006). As discussed in Section
4.2, such a gauge gives the simplest possible boundary distribution of A, in a particular
mathematical sense. It behaves sensibly in the examples in Section 5 where it clearly
peaks within the main current-carrying regions. When we consider the geometry of
Prior & Yeates (2014), A∗ does not precisely reduce to the winding gauge AW , since
AW does not always satisfy condition (3.10). But preliminary computations suggest the
two to be rather similar except on the side boundaries of the domain. This requires
further investigation.
Finally, we have made some simplifying assumptions on the magnetic fields considered.
Firstly, we have assumed that all magnetic field lines have finite length. The work of
 Relative field-line helicity 19
Arnold (1986) suggests that the field-line helicity could be extended to an asymptotic
form for ergodic field lines, although the concept of field-line helicity would be useful only
if a single ergodic field line does not fill the whole volume. Secondly, we have assumed a
domain with simple topology. The theory could also be extended to toroidal or spherical
shell volumes, but in these cases additional restrictions would be needed on A so as to
define A and Ap uniquely. In fact, Yeates & Hornig (2016) have already considered the
evolution of field-line helicity A in a spherical shell representing the global solar corona,
using a specific DeVore-Coulomb gauge in spherical coordinates. Having established that
this gauge was suited to the identification of magnetic flux ropes in the low corona,
Lowder & Yeates (2017) went on to use this as a tool to identify magnetic flux ropes in
non-potential simulations of the global solar corona over a full solar cycle. However, no
attempt was made to optimise the gauge choice in the manner of Section 4.2. This should
also be addressed in future.

ARY was supported by STFC consortium grant ST/N000781/1 and Leverhulme Trust
grant RPG-2017-169. MHP thanks BP for a summer studentship. The authors are
indebted to G. Valori and E. Pariat for sharing the numerical data, made possible through
the ISSI International Team on Magnetic helicity estimations in models and observations
of the solar magnetic field. We thank Chris Prior for suggesting improvements to an
earlier draft, and the anonymous referees for further improving the paper.

Appendix A. Proof of minimal property

Here weH prove that a vector potential satisfying (3.10) on a closed boundary ∂V
minimises ∂V |n̂ × A|2 d2 x among all possible vector potentials.
To do this, suppose that A∗ satisfies the required condition ∇h · A∗ = 0 on ∂V , and
note that any other vector potential on ∂V may be written A = A∗ + ∇φ for some gauge
function φ. Then
I I h i
|n̂ × A|2 d2 x = |n̂ × A∗ |2 + 2(n̂ × A∗ ) · (n̂ × ∇h φ) + |∇h φ|2 d2 x. (A 1)
∂V ∂V

The cross term may be rewritten as

I I
(n̂ × A∗ ) · (n̂ × ∇h φ) d2 x = 2 n̂ · ∇h φ × (n̂ × A∗ ) d2 x,


2 (A 2)
∂V
I∂V
A∗ · ∇h φ d2 x,


=2 (A 3)
I∂V
=2 ∇h · (φA∗ ) d2 x = 0, (A 4)
∂V

which vanishes because ∂V is a closed surface. It follows that

I I I
|n̂ × A|2 d2 x = |n̂ × A∗ |2 d2 x + |∇h φ|2 d2 x, (A 5)
∂V ∂V V

so that the integral is minimised if ∇h φ = 0, i.e., if A ≡ A∗ .

REFERENCES
Arnold, V. I. 1986 The asymptotic Hopf invariant and its applications. Sel. Math. Sov. 5,
327–345.
20 A. R. Yeates and M. H. Page
Barnes, D. C. 1988 Mechanical injection of magnetic helicity. Phys. Fluids 31, 2214–2220.
Berger, M. A. 1984 Rigorous new limits on magnetic helicity dissipation in the solar corona.
Geophysical and Astrophysical Fluid Dynamics 30, 79–104.
Berger, M. A. 1988 An energy formula for nonlinear force-free magnetic fields. Astron.
Astrophys. 201, 355–361.
Berger, M. A. & Field, G. B. 1984 The topological properties of magnetic helicity. J. Fluid
Mech. 147, 133–148.
Berger, M. A. & Ruzmaikin, A. 2000 Rate of helicity production by solar rotation. J. Geophys.
Res. 105, 10481–10490.
Browning, P. K. 1988 Helicity injection and relaxation in a solar-coronal magnetic loop with
a free surface. J. Plasma Phys. 40, 263–280.
Cheung, M., Schüssler, M. & Moreno-Insertis, F. 2005 D Magneto-Convection and Flux
Emergence in the Photosphere. In Chromospheric and Coronal Magnetic Fields (ed. D. E.
Innes, A. Lagg & S. A. Solanki), ESA Special Publication, vol. 596, p. 54.1.
Del Sordo, F., Candelaresi, S. & Brandenburg, A. 2010 Magnetic-field decay of three
interlocked flux rings with zero linking number. Phys. Rev. E 81 (3), 036401, arXiv:
0910.3948.
Démoulin, P. & Pariat, E. 2009 Modelling and observations of photospheric magnetic helicity.
Adv. Space Res. 43, 1013–1031.
DeVore, C. R. 2000 Magnetic Helicity Generation by Solar Differential Rotation. Astrophys.
J. 539, 944–953.
Dixon, A. M., Berger, M. A., Priest, E. R. & Browning, P. K. 1989 A generalization of
the Woltjer minimum-energy principle. Astron. Astrophys. 225, 156–166.
Finn, J. M. & Antonsen Jr., T. M. 1985 Magnetic helicity: what is it and what is it good
for? Comments on Plasma Physics and Controlled Fusion 9 (3), 111–126.
Hornig, G. 2006 A Universal Magnetic Helicity Integral. ArXiv Astrophysics e-prints , arXiv:
astro-ph/0606694.
Longcope, D. W. & Malanushenko, A. 2008 Defining and Calculating Self-Helicity in
Coronal Magnetic Fields. Astrophys. J. 674, 1130–1143.
Low, B. C. & Lou, Y. Q. 1990 Modeling solar force-free magnetic fields. Astrophys. J. 352,
343–352.
Lowder, C. & Yeates, A. 2017 Magnetic Flux Rope Identification and Characterization from
Observationally Driven Solar Coronal Models. Astrophys. J. 846, 106, arXiv: 1708.04522.
Mackay, D. H., Green, L. M. & van Ballegooijen, A. 2011 Modeling the Dispersal of an
Active Region: Quantifying Energy Input into the Corona. Astrophys. J. 729, 97, arXiv:
1102.5296.
Magara, T. 2008 Emergence of a Partially Split Flux Tube into the Solar Atmosphere. Pub.
Astron. Soc. Japan 60, 809–826.
Malanushenko, A., Longcope, D. W., Fan, Y. & Gibson, S. E. 2009 Additive Self-helicity
as a Kink Mode Threshold. Astrophys. J. 702, 580–592, arXiv: 0909.4959.
Moffatt, H. K. 1969 The degree of knottedness of tangled vortex lines. J. Fluid Mech. 35,
117–129.
Moraitis, K., Tziotziou, K., Georgoulis, M. K. & Archontis, V. 2014 Validation and
Benchmarking of a Practical Free Magnetic Energy and Relative Magnetic Helicity Budget
Calculation in Solar Magnetic Structures. Solar Phys. 289, 4453–4480, arXiv: 1406.5381.
Pariat, E., Leake, J. E., Valori, G., Linton, M. G., Zuccarello, F. P. & Dalmasse,
K. 2017 Relative magnetic helicity as a diagnostic of solar eruptivity. Astron. Astrophys.
601, A125, arXiv: 1703.10562.
Pariat, E., Valori, G., Démoulin, P. & Dalmasse, K. 2015 Testing magnetic helicity
conservation in a solar-like active event. Astron. Astrophys. 580, A128, arXiv: 1506.09013.
Pontin, D. I., Wilmot-Smith, A. L., Hornig, G. & Galsgaard, K. 2011 Dynamics of
braided coronal loops. II. Cascade to multiple small-scale reconnection events. Astron.
Astrophys. 525, A57, arXiv: 1003.5784.
Press, W. H., Teukolsky, S. A., Vetterling, W. T. & Flannery, B. P. 1992 Numerical
recipes in FORTRAN. The art of scientific computing.
Prior, C. & Yeates, A. R. 2014 On the Helicity of Open Magnetic Fields. Astrophys. J. 787,
100, arXiv: 1404.3897.
 Relative field-line helicity 21
Russell, A. J. B., Yeates, A. R., Hornig, G. & Wilmot-Smith, A. L. 2015 Evolution
of field line helicity during magnetic reconnection. Phys. Plasmas 22 (3), 032106, arXiv:
1501.04856.
Rust, D. M. 1997 Helicity conservation. Washington DC American Geophysical Union
Geophysical Monograph Series 99, 119–125.
Sturrock, Z., Hood, A. W., Archontis, V. & McNeill, C. M. 2015 Sunspot rotation. I.
A consequence of flux emergence. Astron. Astrophys. 582, A76, arXiv: 1508.02437.
Titov, V. S. & Démoulin, P. 1999 Basic topology of twisted magnetic configurations in solar
flares. Astron. Astrophys. 351, 707–720.
Valori, G., Pariat, E., Anfinogentov, S., Chen, F., Georgoulis, M. K., Guo, Y., Liu,
Y., Moraitis, K., Thalmann, J. K. & Yang, S. 2016 Magnetic Helicity Estimations
in Models and Observations of the Solar Magnetic Field. Part I: Finite Volume Methods.
Space Sci. Rev. 201, 147–200, arXiv: 1610.02193.
Woltjer, L. 1958 A Theorem on Force-Free Magnetic Fields. Proceedings of the National
Academy of Science 44, 489–491.
Yang, S., Büchner, J., Santos, J. C. & Zhang, H. 2013 Evolution of Relative Magnetic
Helicity: Method of Computation and Its Application to a Simulated Solar Corona above
an Active Region. Solar Phys. 283, 369–382.
Yang, S. & Zhang, H. 2012 Large-scale Magnetic Helicity Fluxes Estimated from MDI
Magnetic Synoptic Charts over the Solar Cycle 23. Astrophys. J. 758, 61.
Yardley, S. L., Mackay, D. H. & Green, L. M. 2018 Simulating the Coronal Evolution of
AR 11437 Using SDO/HMI Magnetograms. Astrophys. J. 852, 82, arXiv: 1712.00396.
Yeates, A. R. & Hornig, G. 2011 A generalized flux function for three-dimensional magnetic
reconnection. Phys. Plasmas 18 (10), 102118–102118, arXiv: 1107.0594.
Yeates, A. R. & Hornig, G. 2013 Unique topological characterization of braided magnetic
fields. Phys. Plasmas 20 (1), 012102, arXiv: 1208.2286.
Yeates, A. R. & Hornig, G. 2014 A complete topological invariant for braided magnetic fields.
In Journal of Physics Conference Series, Journal of Physics Conference Series, vol. 544,
p. 012002, arXiv: 1304.8064.
Yeates, A. R. & Hornig, G. 2016 The global distribution of magnetic helicity in the solar
corona. Astron. Astrophys. 594, A98, arXiv: 1606.06863.
Yee, K. 1966 Numerical solution of inital boundary value problems involving maxwell’s
equations in isotropic media. IEEE Trans. Antennas and Propagation 14, 302–307.